Citation
Comparative Analysis of Functional Assessment Methodologies and Functional Analysis for Young Children with or at Risk for Emotional or Behavioral Disorders

Material Information

Title:
Comparative Analysis of Functional Assessment Methodologies and Functional Analysis for Young Children with or at Risk for Emotional or Behavioral Disorders
Creator:
ALTER JR, PETER JAY
Copyright Date:
2008

Subjects

Subjects / Keywords:
Childhood ( jstor )
Classrooms ( jstor )
Disabilities ( jstor )
Functional analysis ( jstor )
Observational research ( jstor )
Questionnaires ( jstor )
Research methods ( jstor )
Research studies ( jstor )
Syntactical antecedents ( jstor )
Teachers ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Peter Jay Alter, Jr. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2006
Resource Identifier:
649815497 ( OCLC )

Downloads

This item has the following downloads:

alter_p ( .pdf )

alter_p_Page_049.txt

alter_p_Page_017.txt

alter_p_Page_025.txt

alter_p_Page_030.txt

alter_p_Page_091.txt

alter_p_Page_089.txt

alter_p_Page_031.txt

alter_p_Page_128.txt

alter_p_Page_046.txt

alter_p_Page_039.txt

alter_p_Page_145.txt

alter_p_Page_084.txt

alter_p_Page_158.txt

alter_p_Page_115.txt

alter_p_Page_175.txt

alter_p_Page_058.txt

alter_p_Page_054.txt

alter_p_Page_038.txt

alter_p_Page_061.txt

alter_p_Page_033.txt

alter_p_Page_162.txt

alter_p_Page_006.txt

alter_p_Page_130.txt

alter_p_Page_078.txt

alter_p_Page_195.txt

alter_p_Page_163.txt

alter_p_Page_007.txt

alter_p_Page_048.txt

alter_p_Page_041.txt

alter_p_Page_138.txt

alter_p_Page_159.txt

alter_p_Page_109.txt

alter_p_Page_069.txt

alter_p_Page_191.txt

alter_p_Page_101.txt

alter_p_Page_166.txt

alter_p_Page_118.txt

alter_p_Page_010.txt

alter_p_Page_201.txt

alter_p_Page_161.txt

alter_p_Page_044.txt

alter_p_Page_105.txt

alter_p_Page_086.txt

alter_p_Page_027.txt

alter_p_Page_020.txt

alter_p_Page_051.txt

alter_p_Page_187.txt

alter_p_Page_171.txt

alter_p_Page_190.txt

alter_p_Page_062.txt

alter_p_Page_081.txt

alter_p_Page_149.txt

alter_p_Page_021.txt

alter_p_Page_066.txt

alter_p_Page_009.txt

alter_p_Page_157.txt

alter_p_Page_004.txt

alter_p_Page_042.txt

alter_p_Page_129.txt

alter_p_Page_040.txt

alter_p_Page_073.txt

alter_p_Page_122.txt

alter_p_Page_150.txt

alter_p_Page_090.txt

alter_p_Page_107.txt

alter_p_Page_135.txt

alter_p_Page_082.txt

alter_p_Page_153.txt

alter_p_Page_178.txt

alter_p_Page_111.txt

alter_p_Page_002.txt

alter_p_Page_164.txt

alter_p_Page_019.txt

alter_p_Page_026.txt

alter_p_Page_096.txt

alter_p_Page_106.txt

alter_p_Page_016.txt

alter_p_Page_065.txt

alter_p_Page_169.txt

alter_p_Page_052.txt

alter_p_Page_142.txt

alter_p_Page_194.txt

alter_p_Page_155.txt

alter_p_Page_176.txt

alter_p_Page_170.txt

alter_p_Page_137.txt

alter_p_Page_147.txt

alter_p_Page_120.txt

alter_p_Page_008.txt

alter_p_Page_011.txt

alter_p_Page_167.txt

alter_p_Page_005.txt

alter_p_Page_117.txt

alter_p_Page_100.txt

alter_p_Page_079.txt

alter_p_Page_032.txt

alter_p_Page_022.txt

alter_p_Page_067.txt

alter_p_Page_189.txt

alter_p_Page_193.txt

alter_p_Page_057.txt

alter_p_Page_113.txt

alter_p_Page_102.txt

alter_p_Page_095.txt

alter_p_Page_121.txt

alter_p_Page_136.txt

alter_p_Page_182.txt

alter_p_Page_098.txt

alter_p_Page_179.txt

alter_p_Page_148.txt

alter_p_Page_146.txt

alter_p_Page_112.txt

alter_p_Page_127.txt

alter_p_Page_151.txt

alter_p_Page_072.txt

alter_p_Page_177.txt

alter_p_Page_141.txt

alter_p_Page_183.txt

alter_p_Page_140.txt

alter_p_Page_087.txt

alter_p_Page_064.txt

alter_p_Page_050.txt

alter_p_Page_192.txt

alter_p_Page_012.txt

alter_p_Page_080.txt

alter_p_Page_172.txt

alter_p_Page_143.txt

alter_p_Page_037.txt

alter_p_Page_197.txt

alter_p_Page_168.txt

alter_p_Page_199.txt

alter_p_Page_152.txt

alter_p_Page_018.txt

alter_p_Page_036.txt

alter_p_Page_124.txt

alter_p_Page_029.txt

alter_p_Page_188.txt

alter_p_Page_123.txt

alter_p_Page_116.txt

alter_p_Page_185.txt

alter_p_Page_173.txt

alter_p_Page_075.txt

alter_p_Page_132.txt

alter_p_Page_125.txt

alter_p_Page_059.txt

alter_p_pdf.txt

alter_p_Page_200.txt

alter_p_Page_024.txt

alter_p_Page_181.txt

alter_p_Page_013.txt

alter_p_Page_053.txt

alter_p_Page_043.txt

alter_p_Page_014.txt

alter_p_Page_070.txt

alter_p_Page_071.txt

alter_p_Page_003.txt

alter_p_Page_198.txt

alter_p_Page_103.txt

alter_p_Page_055.txt

alter_p_Page_156.txt

alter_p_Page_094.txt

alter_p_Page_047.txt

alter_p_Page_034.txt

alter_p_Page_099.txt

alter_p_Page_085.txt

alter_p_Page_083.txt

alter_p_Page_131.txt

alter_p_Page_134.txt

alter_p_Page_074.txt

alter_p_Page_076.txt

alter_p_Page_015.txt

alter_p_Page_108.txt

alter_p_Page_068.txt

alter_p_Page_144.txt

alter_p_Page_174.txt

alter_p_Page_088.txt

alter_p_Page_001.txt

alter_p_Page_092.txt

alter_p_Page_184.txt

alter_p_Page_023.txt

alter_p_Page_104.txt

alter_p_Page_093.txt

alter_p_Page_077.txt

alter_p_Page_154.txt

alter_p_Page_063.txt

alter_p_Page_060.txt

alter_p_Page_196.txt

alter_p_Page_186.txt

alter_p_Page_028.txt

alter_p_Page_139.txt

alter_p_Page_180.txt

alter_p_Page_045.txt

alter_p_Page_114.txt

alter_p_Page_160.txt

alter_p_Page_126.txt

alter_p_Page_056.txt

alter_p_Page_119.txt

alter_p_Page_035.txt

alter_p_Page_165.txt

alter_p_Page_097.txt

alter_p_Page_133.txt

alter_p_Page_110.txt


Full Text












COMPARATIVE ANALYSIS OF FUNCTIONAL ASSESSMENT
METHODOLOGIES AND FUNCTIONAL ANALYSIS FOR YOUNG CHILDREN
WITH OR AT RISK FOR EMOTIONAL OR BEHAVIORAL DISORDERS















By

PETER JAY ALTER JR.


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2006



























Copyright 2006

by

Peter Jay Alter Jr.

































With love to Erin and Betsy.















ACKNOWLEDGMENTS

I am extremely grateful to all for the help and support that I received throughout

this program. It has been fulfilling, challenging and enriching. Most of all, I need to thank

my wife, Erin. I can truly say that without her I would not have gotten to experience this,

and I am eternally grateful.
















TABLE OF CONTENTS

page

ACKNOW LEDGM ENTS ........................................ iv

LIST OF TABLES ...................... ............................... viii

LIST OF FIGURES ...................................... ix

ABSTRAC T ............. ............ ................. ........ ........x

CHAPTER

1 INTRODUCTION ................... ...................................... ......... .......

Overview .................... ...................................................... ...............
State ent of the Problem ...........................................................................8
Significance of the Study ................. ............................................... ..8
Purpose of the Study ................... ............................ ................ .. ......... ..9

2 L ITE R A TU R E R E V IE W ...................................................................................... 10

Prevalence and Impact of Young Children with Challenging Behaviors................ 10
Defining Functional Behavior Assessment ....................................... ......12
Theoretical Framework and Historical Perspective Supporting FBA........................17
R eview of the C current L literature ........................................................................... 21
M ethod to Select Reviewed Studies ............................... ...............22
Results of the Literature Review .......................................................... ... ...... 26
Descriptive Assessment Methodologies.........................................27
Participant and Setting Characteristics..............................................................28
Im plem enter Characteristics ............................................................. ......... 33
Components of the Descriptive Approach ...................................................33
Functions Identified ................ ....... ............... 37
Interventions Im plem ented ............................................................................. 39
Sum m ary ................. ............... ... ..... ... ........... ............. 41
Descriptive Assessments and Functional and Structural Analyses within FBAs.......42
Participant and Setting Characteristics.................................. ....... .............45
Implementer Characteristics ......... ............. ...............47
Components of the Descriptive Approach and Analog Conditions ....................48
Functions Identified ........... ..... ............... 50
Interventions Implemented ..................................................... 52










Su m m ary ................ ...... ... .. ............. ........................ 54
Direct Evaluation of Descriptive Methodologies .............. ...............56
Comprehensive Summary of Research Literature............... .... .................80
Research Questions................ ...... ................. 84

3 METHODS ........................... .............................85

Participants .......................................................85
Settings ........................................ .............................. .........89
M materials ....................................................... .........90
M easurem ent P rocedures....................................................................................... 92
D dependent M measures ................................................. ............... 92
In dep en dent V ariab les .................................................................................... 92
D ata Collection Procedures ........................................... ............... 95
Interobserver Agreement .................... ............................97
Experimental Procedures ............... ......... ................99
Recruitment and Training....................... ......... ....................99
Phase I--Pre-experim ental A ssessm ents............................................................100
Phase II--Descriptive Assessments............................. ..... ........ 101
Phase III--Functional Analysis .............. ...... ...................... 101
Social Validity ......................................... ......... 105
Procedural Integrity ................................................ ........ 105
D esign and D ata A nalysis................................................... 106
Design ............... ...... .................................. 106
Data Analysis. ........... ........... .............. .......... ..........106

4 RESULTS ............................ ............. ............... 110

Individual Child Participant Results............................ ......... ...............111
Ramon.................. .................................112
Anthony ........................ ............ ............. ........113
Jimmy .......................... ............. ............. .........114
Greg ............... ....... ............. ................116
Comparison of Outcomes of Descriptive Assessments..............................118
Social Validity Questionnaire and Interview Results................ ......... .........121
Social Validity Questionnaire................................ ........121
Social Validity Interview..................................... ........123
Summary ..................... ......................... ............124

5 DISCUSSION ............ ........... ................. 126

Limitations.................. .......... ........... ........128
Interpretations of Results.............. ..... .. .................. .........135
Comparison of Descriptive Assessments .............. ................... ................... 135
Comparison of Descriptive Assessments with Functional Analyses ..............138
Social Validity .............................................................. .......142
Future Research ......................................................... ............................. ........143









Im plication s for P practice ...................................................................................... 14 7
Summary ...................................... ................... .................... 150

APPENDIX

A APPROVED IRB FORMS AND PARTICIPANT INFORMED CONSENT
FORMS ......................... .. ...... .... ......... ..52

B FLOW CHART OF TEACHER AND CHILD PARTICIPANT RECRUITING
PROCESS ......... .......................................... 162

C STUDENT PARTICIPANT CHILD BEHAVIOR CHECKLIST (CBCL)
SCORES .......................... ..................... ......... 163

D IN S T R U M E N T S ................................................................................................. 164

E SOCIAL VALIDITY MEASURES ............................ ...............165

F TEACHER TRAINING OUTLINE ........................................... 169

G FUNCTIONAL ANALYSIS PROCEDURES.........................................................173

H FINAL SCORING SUMMARY CHART..............................................................178

L IST O F R E F E R E N C E S ............................................................................................ 179

BIOGRAPHICAL SKETCH .............................................................................189
















LIST OF TABLES


Table page

2-1 Descriptive Methods for Young Children...............................................................29

2-2 Descriptive Methods and Functional Analysis for Young Children........................44

2-3 Direct Evaluations of Descriptive Assessments....................................................59

3-1 Child Participant Inform ation................................................ 88

3-2 Teacher Participant Inform ation............................................ 88

3-3 Summary of Assessments and Their Settings For Each Phase ................................90

3-4 Summary of Child Participants' Preferred Items, Neutral Activities and Difficult
T asks ............... ..................................................................... 9 1

4-1 Summary of the Functions of Challenging Behaviors Across Assessment
M methods .................................................................. .................... 112

4-2 Summary of Agreements Between Descriptive Assessments Compared With
Each Other and w ith a Functional Analysis..........................................................119

4-3 Results of the Social Validity Questionnaire ................. ................. .........121
















LIST OF FIGURES


Figure page

4-1 Challenging Behaviors per Minute During Functional Analysis for Ramon.........113

4-2 Challenging Behaviors per Minute During Functional Analysis for Anthony. .....114

4-3 Challenging Behaviors per Minute During Functional Analysis for Jimmy. ........116

4-4 Challenging Behaviors per Minute During Functional Analysis for Greg. ...........117
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

COMPARATIVE ANALYSIS OF FUNCTIONAL ASSESSMENT
METHODOLOGIES AND FUNCTIONAL ANALYSIS FOR YOUNG CHILDREN
WITH OR AT RISK FOR EMOTIONAL OR BEHAVIORAL DISORDERS

By

Peter Jay Alter Jr.

August 2006

Chair: Vivian I. Correa
Cochair: Maureen A. Conroy
Major Department: Special Education

The prevalence of functional behavior assessments (FBAs) in schools continues to

increase as a result of federal mandates and widespread school policy reform. As the use

of FBAs in schools increases, it is necessary to improve their current condition by

evaluating the concurrent and construct validity of the process components. Failure to

establish FBA as a process with valid outcomes will adversely impact students who

demonstrate challenging behaviors because a method for designing effective

interventions will either be unused or the interventions that result from the process will be

ineffective. There are a large number of descriptive methodologies that can be used

within the FBA process. However, the outcomes of many of these methodologies have

not been compared with each other or validated through functional analysis. Thus, there

is a continuing need to empirically compare and validate the outcomes of these

descriptive assessment methods. The two purposes of this investigation were: (1) to









compare the results of three commercially produced descriptive assessments with each

other to determine concurrent validity, and (2) to compare the descriptive assessment

results with the results of a functional analysis to determine construct validity.

Descriptive assessments including one checklist, an interview, and a direct observation

were completed for four young children (ages 4-6) with or at risk for emotional and

behavioral disorders. Teachers and researchers then completed functional analyses within

the classroom setting of the same participants. Results of the assessments were kept

independent from each other until completion. Teachers also completed a social validity

questionnaire and interview. This study indicated that descriptive assessment results had

low consistency with each other, 8/24 (33%) agreements between assessments, and that

the outcomes of two of the three descriptive assessments (i.e., the interview and

checklist) also had low agreement with the functional analyses with 3/8 (38%)

agreements. However, the direct observations agreed with the functional analyses

conclusions for all 4/4 (100%) of the participants. These results do not support the

concurrent validity of the three descriptive assessments or the construct validity of the

checklist or the interview, but do support the construct validity of direct observations.














CHAPTER 1
INTRODUCTION

This chapter provides a brief overview of the literature and presents the need to

examine the validity of various assessments in the functional behavior assessment (FBA)

process. This brief overview outlines the use of FBA with preschool-aged children who

demonstrate challenging behavior and also highlights articles that have directly evaluated

the results of these various assessments. The articles that have directly evaluated various

assessments compare the reported results of the descriptive assessments to each other and

to results of functional analyses. The literature provides a synthesis of the evolution of

the FBA process, the need for distinct comparisons of the various assessments and a

rationale for the current study. In addition, the contributions this study makes to improve

the effectiveness of the FBA methodology, in both research and practice, are identified.

Finally, this chapter presents the specific research questions that were addressed in this

study.

Overview

As the utilization of functional behavior assessments (FBAs) increases, it is

incumbent on researchers to improve FBAs' current practices by evaluating the

effectiveness and the efficiency of this process. The increasing prevalence of FBAs is a

direct result of two interrelated factors. The first reason for more widespread application

of FBAs is the increased number of children who are engaging in challenging behaviors

(Chandler & Dahlquist, 2002). The number of young children exhibiting challenging

behaviors is rising. Many of these children may ultimately meet the criteria for









emotional/behavioral disorders (Fox, Dunlap, & Powell, 2002). According to Webster-

Stratton (2000), 7% to 25% of preschool-aged children currently meet criteria that place

them at risk for future development of oppositional defiant disorder or a conduct disorder.

For example, Kaiser, Kai, Hancock and Foster (2002) tested 332 three-year old children

enrolled in Head Start and found that 39% of boys and 22% of girls demonstrated

behavioral deficits that placed them in the clinical range for both internalizing and

externalizing problem behaviors on the Child Behavior Checklist (Achenbach, 1992).

FBAs are one strategy that can be used to address children's needs; however, many of the

FBA procedures used are problematic (Johnston & O'Neill, 2001). Increased

effectiveness of FBAs are needed to appropriately meet the increasing needs of young

children with challenging behaviors.

Second, the reauthorization of the Individuals with Disabilities Education

Amendments (IDEA, 2004) has mandated the use of functional behavior assessments for

all students who have been suspended for over ten days, who demonstrate behaviors that

warrant consideration of a change in educational placement, or disciplinary measures

such as placement in a more restrictive setting (i.e., an interim alternative educational

setting) or expulsion. As a result, school districts now view the completion of an FBA as

the first step in any significant intervention. This IDEA mandate has also impacted early

childhood personnel working with young children ages 3 to 6 served under Section 619

of Part B of the act. The prevalence of FBAs in early childhood settings has been

stimulated both by IDEA legislation (Nielsen & McEvoy, 2004) and recognition that the

process of FBA represents an important methodology in addressing challenging

behaviors demonstrated by young children (Fox, Dunlap & Powell, 2002).









Functional behavioral assessment is a process in which information is gathered

regarding a specific behavior or set of behaviors in order to determine whether that

behavior has any predictable relationship with the environment (O'Neill et al., 1997).

Relationships exist when these environmental factors both predict the behavior and

explain why it is maintained. Identification of these relationships increases understanding

of why a particular behavior occurs or how behavior is reinforced and maintained by the

specific environmental consequences (i.e., functions) they elicit. The FBA must identify

the environmental factors surrounding challenging behaviors. Once environmental factors

are identified and understood through an effective FBA, individualized interventions may

be designed to replace undesirable behavior with appropriate alternatives that efficiently

obtain the same functions of the challenging behavior. For example, if a child engages in

challenging behaviors to receive teacher attention, an intervention is designed that gives

the child increased access to teacher attention by demonstrating desired behaviors and

decreased access to teacher attention for challenging behaviors. In most educational

settings, information for FBAs is gathered using a combination of both indirect and direct

descriptive assessment measures (Johnston & O'Neill, 2001). Indirect measures are those

that do not require direct observation of the student. Indirect measures include interviews

with parents, teachers, and children via the use of questionnaires, rating scales and record

reviews. In contrast, direct measures typically consist of scatter plots (Touchette,

MacDonald, & Langer, 1985) and observation worksheets that reflect a running narrative

of antecedents, behaviors and consequences (ABCs) (Bijou, Peterson, & Ault, 1968) that

are recorded as they are observed in natural settings.









An alternative method for gathering information to determine the function of a

challenging behavior is a functional analysis (Sasso, Conroy, Stichter, & Fox, 2000).

Functional analysis refers to the use of analog conditions, manipulating specific

environmental conditions that follow challenging behaviors, to systematically identify

consequences as they relate to the function of a behavior (Sugai, Horner, & Sprague,

1999). Following the manipulation of each of these conditions, the consequence that

produces the highest level of behavior is identified as the function. The difference

between the former two descriptive methods (direct and indirect) and the third method of

functional analysis is that the descriptive methods generate correlational results while a

functional analysis identifies causal relationships between behavior and the maintaining

consequences (Lerman & Iwata, 1993). No combination of these three methods has been

accepted as the standard protocol either by research literature or federal mandates (Scott

et al., 2004).

A literature review of the research of the use of FBAs with young children

demonstrating challenging behaviors revealed three distinct groupings: (a) studies that

only employed indirect and direct descriptive methodologies to determine hypothetical

functions; (b) studies that combined descriptive methodologies with functional analysis

methodologies to determine functions; and (c) studies that directly compared the results

of different assessments with each other and/or a functional analysis. Although some

FBA methodologies have been available for use by teachers for over 20 years (Ervin,

Ehrhardt & Poling, 2001), FBA methods are still in a fledgling state of development.

Studies that only used descriptive assessment methods in the FBA process remain

the most prevalent for young children demonstrating challenging behavior. By using only









descriptive assessments, and not using functional analysis, the function identified in this

body of research remains hypothetical. Interventions based on the hypothetical functions

that are not experimentally manipulated (i.e., a reversal design) only provide correlational

support (Kazdin, 1982). Additionally, in studies that only used descriptive assessments,

many of the interventions implemented within the studies were either not directly related

to the hypothetical function or included additional interventions that muddled whether or

not the hypothetical function identification was accurate. For example, Grandy and Peck

(1997) hypothesized that the function of the challenging behavior was teacher attention.

However, rather than creating an intervention that limited teacher attention for

challenging behavior and increased teacher attention for desired behaviors, researchers

implemented a token economy that led to classroom privileges and that indirectly led to

teacher attention. This lack of a direct connection between the determined function and

the intervention implemented calls into question the accuracy of the determined function.

It is impossible to determine if teacher attention was reinforcing the desired behavior or if

it was another element of the token economy. As illustrated through this example, the

body of literature using descriptive assessments gives minimal support to the accuracy or

consistency of the various descriptive assessments.

Studies that employ both descriptive assessments and functional analysis

methodologies also have certain shortcomings. First, researcher influence was

pronounced in these studies that included the time consuming, technically complex

functional analysis process. In many of the studies reviewed, a team of researchers

worked with very few participants (i.e., Doggett, Edwards, Tingstrom & Wilcznski,

2001; Harding et al., 1999; Romaniuk et al., 2002). Another salient difference between









the descriptive methods in the studies that used both descriptive assessments and

functional analysis and the sample that only used descriptive methods is that this sample

was more parsimonious in the use of descriptive assessments. When functional analysis

was also incorporated into the study's protocol, descriptive assessments were only used

to the extent that they could guide the functional or structural analysis procedures. These

studies used the descriptive assessment results to identify times and situations to conduct

analog probes during the times or activities that teachers identified as the most

problematic. For example, in both their studies Broussard and Northup (1995, 1997) only

used direct observations of the participant and a brief interview with the teacher to

determine what would make the analog conditions most representative of the natural

environment. This makes it very difficult to determine what parts of the process are

producing accurate and/ or consistent results. Therefore, the effectiveness and efficiency

of various FBA methodologies are unknown.

There were eight studies that directly evaluated different aspects of the functional

behavior assessment process. This group of studies that attempted to compare the results

of different descriptive assessments with each other and with functional analysis

procedures had several prominent characteristics. Overall, there was agreement between

the different descriptive assessments results. There was also agreement between the

descriptive assessment results and the functional analysis results. However there are three

reasons why these results should be interpreted with caution. First, the descriptive

assessment results informed one another and were not independent from each other as

they were conducted, making direct comparison of separate results impossible. For

example, in Ellingson, Miltenberger, Stricker, Galensky and Garlinghouse (2000),









Murdock, O'Neill and Cunningham (2005) and Newcomer and Lewis (2004), the results

of the indirect assessment directly shaped and guided the direct descriptive observations

thereby compromising their claims of independent consistency across assessment

methods. This makes direct comparison impossible as the results of one assessment

influence the results of another. The second interrelated factor with this assimilation of

assessment methods is that by not being presented in different orders, the order of

administration was not counterbalanced. As a result, the aforementioned three studies and

Yarbrough and Carr (1997) presented their descriptive assessments in the same order,

making it impossible to control for an order effect. In other words, there is a possibility

that a participating teacher's previous responses to an assessment would shape responses

to subsequent assessments.

The third salient factor is the use of the descriptive assessment results to guide the

functional analysis procedures. Within this body of literature, only Cunningham and

O'Neill (2000), and Sasso et al. (1992) conducted functional analysis procedures with

five analog conditions: free play (control), tangible, escape, attention, and alone. Other

studies only tested analog conditions that matched the results of the descriptive

assessments. This difference is subtle yet critical. By only evaluating those hypothetical

functions determined by the descriptive assessment, it is possible to confirm the results as

a function but not the function (because different functions were not compared).

Additionally, two studies, Murdock, et al., (2005) and Ellingson et al. (2000), did not use

any experimental procedures to confirm the functions; this means that their results may

show consistency across descriptive measures, but that does not confirm the accuracy of

the function identified.









In conclusion, all three bodies of FBA literature indicate that there is a need to

evaluate the effectiveness of current assessment procedures. By keeping the different

assessment results separate from each other and from the functional analysis, it was

possible to evaluate each instrument's results independently. This extends current FBA

research that has attempted to evaluate the accuracy of different assessment

methodologies.

Statement of the Problem

No single procedure or combination of procedures for functional behavior

assessment has emerged as the most suitable for use with young children demonstrating

challenging behaviors. Minimal empirical studies have evaluated the effectiveness and

usefulness of various approaches. "Several authors have noted that research directly

comparing the results of different approaches such as interviews and direct observations

is sparse" (Johnston & O'Neill, 2001, p. 206). Failure to establish FBA as a meaningful

process with valid outcomes will adversely impact countless children who demonstrate

challenging behaviors. Absent a highly effective method for determining function and

designing effective interventions, appropriate plans will either go unused or the

interventions that result from the process will be ineffective. Teachers need to be

equipped with strategies that determine the function of a challenging behavior reliably

and accurately for the FBA process to be effective and useful. If the process is ineffective

and does not decrease challenging behaviors, then the FBA process will become another

bureaucratic chore that will not positively impact children (Scott et al., 2004).

Significance of the Study

This fundamental need to meet the demand for effective FBAs that can be

implemented by teachers in a meaningful way leads to an important area of research: to









conduct a series of indirect and direct descriptive assessments and compare the outcomes

of these assessments to each other and to the outcomes of a functional analysis

methodology. This study extends and builds on the current literature that directly

evaluated the results of different descriptive assessments. However, unlike these studies,

the functional analysis in this study was independent of the descriptive assessment

results, thus allowing an unbiased comparison of descriptive and experimental

methodologies. The results of the various direct and indirect descriptive assessments were

recorded independently and separation of the results was maintained until the end of the

study. Additionally, the descriptive assessments' order of administration was

counterbalanced across participants to maintain independence as well. The results of this

research are an empirically validated step towards evaluating descriptive assessments that

can be used both in continuing FBA research and in applied public school and early

childhood settings.

Purpose of the Study

The specific questions this research study addressed were the following:

1. How do the functions of target behaviors identified through various descriptive
functional assessments compare with functions identified through a functional
analysis for young children's challenging behavior?

2. How do the functions of target behaviors identified through various descriptive
functional assessments compare to each other?














CHAPTER 2
LITERATURE REVIEW

The purpose of this chapter is to review the literature on the use of functional

behavior assessment (FBA) that address challenging behaviors in applied settings

demonstrated by young children. This review focuses on the processes that comprise

FBAs, in order to determine the most effective methodologies. First, the significance of

completing FBAs will be traced in the context of an increase in the rate and severity of

challenging behaviors demonstrated by young children. Second, this chapter will define

functional behavior assessment, and outline the current components that comprise FBA.

Third, the theoretical underpinnings will be outlined as they relate to its current

application especially with young children. Fourth, specific aspects of this application

will be examined through a targeted review of the empirical literature. This literature will

be divided into three sections: (1) studies that used only descriptive assessments in the

FBA process (2) studies that used both descriptive assessments and experimental

analyses; and (3) studies that directly examined and compared the results of different

descriptive assessment methodologies. Finally, a general discussion will address the

shortcomings of current empirical literature and outline directions for future research.

Prevalence and Impact of Young Children with Challenging Behaviors

The prevalence of young children exhibiting challenging behaviors is increasing

(Fox, Dunlap, & Powell, 2002), and children are at risk for future development of

behavioral disorders (Webster-Stratton, 1997). The young children under consideration

are defined as children under 8 years of age, in accordance with the age limit specified by









the Division of Early Childhood of the Council for Exceptional Children. Researchers

have documented various prevalence rates across a number of different behavior

disorders. For example, 7% to 25% of preschool-age children currently meet the criteria

for a diagnosis of Oppositional Defiant Disorder (ODD) as defined by the Diagnostic

Statistical Manual IV-TR (DSM IV-TR) (Webster-Stratton, 2000). The DSM defines

Oppositional Defiant Disorder (ODD) as "a recurrent pattern of negativistic, defiant,

disobedient, and hostile behavior toward authority figures that persists for at least 6

months." Additionally, Kaiser, Kai, Hancock, and Foster (2002) tested 332 three-year old

children enrolled in Head Start and found that 39% of boys and 22% of girls scored in the

clinical range for both internalizing and externalizing problem behaviors. As the rate of

challenging behaviors demonstrated by young children continues to increase, the need to

manage and modify behavior in early childhood settings also increases. The inability to

remediate these challenging behaviors will affect young children both now and in the

future.

Behaviors such as tantrums, aggression, and chronic non-compliance exhibited by

young children may have immediate and prolonged effects on academic achievement and

appropriate socialization (Shonkoff & Phillips, 2000). The immediate effects of

challenging behaviors for preschool-aged children are academic under-achievement and

poorly developed social skills (Chandler & Dahlquist, 2002; Webster-Stratton, 1997).

The long-term effects of unaddressed challenging behaviors are an increase in both rate

and severity over time (Reid & Patterson, 1991). Early behavior problems have also been

linked with more severe problems such as substance abuse, unemployment, criminal

behavior, and diagnosis of a psychiatric disorder (Reid and Patterson, 1991).









Chandler and Dahlquist (2002) state that difficult behaviors are often rooted in the

functions that they served at young ages, such as hitting a peer to attain a tangible object

(e.g., a toy), and these behaviors may also increase in frequency, duration, and intensity

as the child gets older. Ascertaining the function of a challenging behavior is the key to

developing an effective intervention to extinguish it (Cooper & Harding, 1993; Iwata et

al., 1982/1994; Gresham, 2003; O'Neill et al., 1997; Sasso et al., 2001) and may prevent

its undesirable immediate and long-term consequences.

Defining Functional Behavior Assessment

The use of functional behavior assessments (FBAs) in school settings represents an

important process in the effort to address increasingly prevalent problems with behavior

in children. Using functional behavior assessment and functionally derived interventions,

as opposed to interventions that are unrelated to the function of the behavior, is supported

by research and considered "best practice" in the field (Arndorfer & Miltenberger, 1993;

Gresham, 2003; Iwata et al., 1982/1994; National Institute of Health, 1989; O'Neill et al.,

1997; Repp, Felce, & Barton, 1988; Sasso et al., 2001). Functional behavior assessment

was designed to identify variables to manipulate through applied interventions to reduce

challenging behaviors that interfere with or prohibit learning, and increase pro-social and

adaptive behaviors that can lead to greater academic and social success (O'Neill et al.,

1997).

In schools, the use of FBA represents a paradigm shift for educators that

traditionally applied a standard system of consequences for all infractions (i.e., time out,

a phone call to parents, a loss of privileges, suspension, expulsion). In early childhood

settings, FBAs represent a valuable tool to help early childhood professionals address the

behaviors that surpass usual challenging behaviors demonstrated by young children (Fox,









Dunlap, & Powell, 2002). With FBA, educators can design interventions that match the

function of a challenging behavior in children. The function of a behavior refers to the

purpose or outcome of the challenging behavior for the child (Scott et al., 2004). In other

words, a function refers to the desired consequence that motivates a challenging behavior,

and function-based interventions refer to environmental manipulations that allow a

student to achieve the desired goal through other means (i.e., appropriate replacement

behaviors). For example, if the function, or purpose, of a student exhibiting a challenging

behavior is to receive teacher attention, then a challenging behavior could be

extinguished by giving the student access to teacher attention by exhibiting appropriate

behaviors and not providing additional teacher attention for the challenging behavior.

Functional behavior assessment, as defined for this literature review, is a process in

which information is gathered regarding a specific behavior or set of behaviors in order to

determine whether that behavior has any predictable relationship with the environment

(O'Neill et al., 1997). The environment is comprised of all factors preceding

(antecedents) and events following (consequences) the behavior. A relationship exists

when these environmental factors predict the occurrence of the behavior and explain why

the behavior is maintained. This assessment helps to understand why behavior occurs, or

how behavior is reinforced and maintained by the specific environmental consequences

(i.e., functions) they elicit. This assessment also identifies the environmental factors that

precede a challenging behavior and make the challenging behavior more or less likely to

occur. The purpose of an FBA is to understand environmental factors that surround the

behavior so that individualized interventions may be designed to replace undesirable









behavior with appropriate alternatives that efficiently match the functions of the

challenging behavior.

Because current empirical studies have applied the term functional behavior

assessment broadly, the definition of FBA for this literature review requires further

specification. As noted, functions typically refer to environmental consequences that

follow a challenging behavior. The experimental evaluation of functions of behavior is

referred to as 'functional analysis' within the literature. However antecedents, (also

referred to as setting events or establishing operations) that precede and increase the

likelihood of challenging behaviors, also play an important role in functional behavior

assessment (Conroy & Stichter, 2003). Within the literature, experimental evaluation of

antecedents is referred to as a 'structural analysis,' but may also simply be included as

another aspect of an FBA. For the purposes of this literature review, studies that evaluate

both the antecedents and consequences of behavior will be included as part of an FBA.

There are essentially three ways to generate functions of a challenging behavior: (1)

indirect descriptive measures that establish hypothetical functions of a challenging

behavior, (2) direct descriptive measures that establish 1yipohicihu al functions and (3)

experimental analysis of an environment that can determine the actual function of a

challenging behavior. However, for the purposes of this literature review direct and

indirect measures will be combined and simply referred to as descriptive assessment

measures in the three sections of literature that are reviewed (Sturmey, 1994). Indirect

measures are those that do not require direct observation of the student and the

challenging behavior. Typically, indirect measures include interviews with parents,

teachers, (and the student if appropriate), questionnaires, rating scales and reviews of









school records. A number of commercially produced comprehensive interviews exist (see

O'Neill et al., 1997 for an example). In addition, commercially produced behavior

checklists, such as the Motivation Assessment Scale (MAS) (Durand & Crimmins, 1992)

are designed for determining the hypothetical function of a behavior. The total number of

indirect assessments of this type is difficult to determine. According to a literature review

conducted by Floyd, Phaneuf and Wilczynski (2005) there are 19 commercially produced

indirect assessments; however, this does not include those indirect assessments that are

created by individual school systems or early childhood centers. Further, as Floyd et al.,

(2005) point out most of these other instruments have minimal research into their efficacy

and validity; only O'Neill's interview and the MAS met their criteria for inclusion in a

literature review of indirect descriptive assessment methods.

In contrast, direct measures require the direct observation of the child and the

challenging behaviors. Direct descriptive assessments typically consist of scatter plots

(Touchette, et al., 1985), and observation worksheets that reflect antecedents, behaviors

and consequences (ABCs) (Bijou, et al., 1968) that are recorded in real time as they are

observed. While some school districts and early childhood centers may adopt

commercially produced FBA measures, they also may have created their own direct and

indirect methods of gathering data to identify these functions. This means that there are a

large number of descriptive methodologies that can be used within the FBA process.

Finally, the third type of procedure is an experimental approach, also referred to as

functional and structural analysis. Functional analysis refers to the use of analogs,

manipulating specific environmental conditions that follow challenging behaviors, to

systematically identify consequences as they relate to the function of a behavior (Sugai,









Horner, & Sprague, 1999). For example, to determine the function of the behavior,

analog situations representing possible consequences are manipulated (i.e., attention,

tangible, escape, and control conditions). Following the manipulation of each of these,

the consequence that produces the highest level of behavior is identified as the function.

For example, to determine the function of a challenging behavior, each analog condition

is conducted, contingently providing the targeted consequence following the occurrence

of the behavior. For example, in the attention condition, the consequence of adult

attention would be provided contingently every time the target behavior is demonstrated

while all other possible antecedents and consequences are controlled to account for any

variance. If the consequence of adult attention results in an increased rate of the behavior,

as compared with other consequences administered in the same controlled environment,

then it can be concluded that the function of this challenging behavior is adult attention.

Structural analysis involves a similar process; however, rather than systematically

manipulating the consequences that follow a behavior, a structural analysis manipulates

the antecedents that precede the behavior to identify variables that set the occasion or

serve an evocative effect.

All three of these assessment procedures have drawbacks. Indirect assessments rely

on the subjective judgments of a third party, and direct observations can be context

specific and inefficient in capturing the occurrence of a challenging behavior and its

antecedents and consequences (Floyd et al., 2005). Direct observations can be inefficient

because the probability of actually observing the challenging behavior may be relatively

low, necessitating long and/or frequent observation periods (Scott et al., 2004). In this

sense, the measures are inefficient because they may demand a considerable amount of









time from the practitioner but do not insure meaningful results. Functional and structural

analysis have been identified as too time consuming, and technically complex for

practitioners to conduct, and may lack external validity due to the high degree of

experimental control (i.e., the highly controlled conditions in which it is conducted)

(Arndorfer & Miltenberger, 1993; Ellingson, et al., 2000).

Regardless of the methodology, identifying and evaluating relationships between

environmental antecedents and consequences and challenging behavior improves the

opportunities to enable all students to meet academic and behavioral goals (Hanley, Iwata

& McCord, 2003). As antecedents and consequences that are related to challenging

behavior are identified, environments can be modified to reduce the occurrence of

challenging behavior. This may be done by reducing the antecedents that increase the

occurrence of the challenging behavior, giving students access to desired reinforcement

through adaptive behaviors and/or limiting access to reinforcement for challenging

behaviors. Creating effective interventions that reduce or eliminate challenging behavior

at a young age can decrease the possibility that they will increase in frequency, duration

and intensity as the child gets older. The theoretical framework for designing

interventions according to function is rooted in the Behaviorist approach to human

psychology. This theoretical framework and an historical perspective supporting FBA are

described in the next section.

Theoretical Framework and Historical Perspective Supporting FBA

The theoretical framework for functional behavior assessment methodology can be

traced back to the work of early behaviorists such as J.B. Watson and, so-called neo-

behaviorists such as B.F. Skinner (1953). In fact, the term function, as it is used in this

context, is often credited to Skinner's book Science and Human Behavior (1953). Skinner









(1953) states, "The external variables of which behavior is a function provide for what

may be called a causal or functional analysis." (p. 35). The subsequent work in the field

of applied behavior analysis, codified by the creation of the Journal ofApplied Behavior

Analysis (JABA), expanded this theoretical framework to work with humans in applied

settings.

Within this framework, researchers address behaviors by examining the

environmental variables that maintain them, and operate on the assumption that

functional relations not only exist, but also can be identified and manipulated. Early

seminal work in JABA included an article by Bijou, Peterson and Ault (1968) that

described a method to integrate descriptive and experimental data within field studies.

Additionally, Sailor, Guess, and Rutherford (1968) conducted early work identifying

"functions" of behavior and reported that for a 9 year old who engaged in frequent

tantrums, the problem behavior "seemed to produce the effect of terminating contact with

other individuals who usually were making some demand of her" (p. 238). While

establishing the function of a behavior can be traced back to the work of early

behaviorists, the emergence of a specific process for determining the behavioral functions

is related to two more recent seminal events; the creation of a functional analysis

methodology and the legal mandate of FBA in the Individuals with Disabilities Education

Act (IDEA, 1997; 2004).

First, Iwata et al. (1982/1994) developed a methodology that used various

environmental conditions, referred to as analog probes, in a clinical setting to identify a

relationship between the environment and the function of self-injurious behavior. These

analog probes included the following conditions: adult attention, escape from a task, an









alone condition and free play (Iwata et al., 1982/1994). The direct, systematic

manipulation of environmental factors in a highly controlled environment in order to

evaluate the function of a challenging behavior, in this case self-injurious behaviors, was

termed 'functional analysis.' This procedure led to highly specific, highly effective

intervention plans that incorporated the function of the challenging behavior and thus,

decreased these behaviors. As the importance of finding the function of a challenging

behavior became apparent, the methodologies broadened beyond experimental

manipulation into more descriptive assessments. Following this groundbreaking research,

a more formalized functional behavior assessment process emerged from the clinic

setting and was implemented in more natural settings, such as the classroom.

The second critical event, the reauthorization of the Individuals with Disabilities

Education Amendments (1997), provided the impetus to extend the methodology by

making the use of functional behavior assessments (FBAs) mandatory for all students

whose behavior warrants consideration for a change in educational placement or

disciplinary measures such as extended suspension or expulsion. Although IDEA (1997)

provided the impetus to expand the methodology of FBA, it did not specify what

procedures should be used by individuals in schools when implementing an FBA

resulting in varying forms of FBAs from state to state, district to district and even school

to school. Despite its continually increasing prevalence in both educational systems and

research settings, the effectiveness of FBA is still hampered by muddled definitions, an

unspecified protocol and components that have not been empirically validated for use

with various populations in numerous settings (Ervin, Ehrhardt, & Poling, 2001;

Gresham, et. al, 2004; Scott et al., 2004; Sugai, et al., 1999).









While IDEA mandates have specific implications for public schools, the legislation

has also continued to stimulate interest in the FBA process as it impacts early childhood

settings such as child care centers (Neilsen & McEvoy, 2004). This prevalence has a clear

relationship to the federal mandates regarding positive behavior support and the need for

evidence-based practices for all ages (see Conroy, Dunlap, Clark, & Alter, 2005 for a

review). As outlined by Amdorfer and Miltenberger (1993), there are specific

implications for FBA in early childhood settings. These implications include examination

of the utility of functional assessment methods in early childhood settings, especially

including the use of various different descriptive assessments (i.e. indirect and direct

assessments) and the viability of functional analysis procedures in early childhood

settings. Thus, there is a need for examination of these specific components, as the role of

FBA in early childhood settings continues to be prevalent in research and practice.

No single procedure or combination of procedures has emerged as the most suitable

for application with young children demonstrating challenging behaviors. In concert with

this shortcoming, minimal empirical studies have evaluated the effectiveness and

usefulness of these various approaches. "Several authors have noted that research directly

comparing the results of different approaches such as interviews and direct observations

is sparse" (Johnston & O'Neill, 2001, p, 206). Failure to establish FBA as a meaningful

process with valid outcomes will adversely impact countless students who demonstrate

challenging behaviors, because a highly effective method for designing effective

interventions will either be unused or the interventions that result from the process will be

ineffective.









The application of various functional behavior assessment procedures in

educational settings has outpaced current research. There is minimal research validating

the effectiveness of different functional behavior assessment processes currently used for

children with emotional and behavioral disorders (Sasso, Conroy, Stichter & Fox, 2001).

"Given the different options available for conducting a functional assessment, the

question regarding which approach practitioners ought to use to assess problem behavior

is paramount" (Yarbrough & Carr, 2000, p, 133). The rising prevalence of young

children's challenging behavior in early childhood settings increases this urgency. Thus,

educators need to be equipped with strategies that determine the function of a challenging

behavior reliably and accurately for the FBA process to effective and useful.

Review of the Current Literature

Current research in the use of FBAs, including functional and structural analyses,

with young children indicates that although there is considerable application of these

methodologies, the use of FBAs with young children with EBD in applied or natural

(e.g., classrooms, home) settings is still in its fledgling state. The first section of this

review examines the use of both direct and indirect descriptive FBA assessment

methodologies in natural settings (i.e., such as childcare centers, public school

classrooms, and homes). The rationale for reviewing this group of empirical studies is

that they can provide insight into the effectiveness of FBAs that are conducted without

any type of systematic environmental manipulation. The second section examines FBAs

that used both types of descriptive assessments and experimental analysis for FBAs.

Examining this literature will provide a clearer picture as to what place, if any, functional

analysis will hold in natural settings and what the ramifications of not using experimental

analysis may be. The third section examines studies that specifically compared the results









of descriptive assessment methodologies with each other and with the results of a

functional analysis. This purpose of this section of the literature review is to examine

what research has been conducted thus far comparing, evaluating and possibly validating

different components of the FBA process.

Method to Select Reviewed Studies

This literature search was completed in three steps. First, the author conducted a

computer search of EbscoHost including Academic Search Premier, PsychlNFO,

Psychology and Behavioral Sciences Collection, and the Professional Development

Collection databases), ERIC, and First Search. The following keyword combinations

were used: functional, assessment, assessment-based, analysis, behavior, experimental,

school, preschool and young children. Second, articles were cross-referenced that were

identified in the computerized search to determine if research articles had been omitted.

Third, the author consulted with researchers who have written extensively in this area to

accumulate any other articles that may have been omitted in the first two steps.

To compare 'descriptive assessments' (direct and indirect) to 'experimental

analysis,' it is necessary to differentiate the two terms. As previously described,

descriptive assessment represents the use of methodologies including: direct observation,

rating scales, interviews, and hypothesis development using antecedent behavior

consequence (ABC) analysis (Bijou et al., 1968). Experimental analysis (i.e., structural

analysis and functional analysis) uses analog probes to purposefully manipulate a

variable within the environment and then develops hypotheses based on these controlled,

manipulated environments. One important distinction between these two terms is that

descriptive assessment methodologies identify only a hypothetical function; whereas,









functional analyses are able to identify a causal relationship (i.e., function) through

experimental analyses (Lerman & Iwata, 1993).

However, certain characteristics of these research articles made separating

descriptive methodologies from functional analysis a complex task in several ways. First,

all descriptive studies included hypothesis testing, the addition or removal of a specific

stimulus in the environment. This type of environmental manipulation is similar to a

traditional FA. However, two features differentiated hypothesis testing and functional

analysis: (1) the type of experimental control and (2) the number of functions tested in

these controlled settings. Control of the environment is the distinguishing characteristic

separating descriptive analysis from functional analysis (Sasso et al., 1992). For

example, if researchers manipulated the environment to keep every other factor

(curriculum, schedule) the same and only systematically changed the analogs, this was

termed a functional analysis. Additionally, if a study tested multiple functions in this

controlled environment then it was termed a functional analysis. For example, if a study

systematically tested tangible items, adult attention and an alone condition, this was

considered a functional analysis. However, if a study only performed a test of one

hypothetical function based on the interviews or direct observations, this was still

considered a validation of the findings from the descriptive assessment (Hanley, et al.,

2003).

Another characteristic that made distinguishing descriptive methodologies from

experimental analysis difficult is the absence of clearly standardized terminology in the

research studies. In certain studies, the actual methodology is obfuscated by an imprecise

use of terms. For example, Kamps et al. (1995) referred to their research as a "functional









analysis." However, because of minimal environmental control, the methodology used in

this study is more accurately judged to be a descriptive analysis, according to Sasso et

al.'s (1992) definition. To counteract the use of imprecise terminology within the

literature, the criterion for categorization was set that if individuals (researchers, teachers,

parents) manipulated the environment to test multiple different functions in any way, in

order to confirm a hypothetical function, then it was deemed an experimental analysis.

There are two final caveats regarding this differentiation. The first issue concerned

studies that implemented interventions-even if interventions were implemented explicitly

to confirm hypothesized functions. Regardless, these studies were still judged to fall

under the 'solely descriptive' section of the literature review. The rationale for this

decision is that while interventions may confirm a functional relationship, they do not

necessarily confirm the functional relationship, unlike a traditional functional analysis.

Other functional relationships may exist that would not be tested simply by implementing

interventions guided by descriptive assessment results. For example, a descriptive

assessment suggests that desire for a tangible object may be the function of the behavior

and an intervention is implemented that provides increased time playing with the toy for

desired behaviors and decreased time playing with the toy for challenging behaviors.

Although this may be an effective way to reduce challenging behaviors, it does not

indicate that this is either the only function or the primary function of the challenging

behavior. The function may actually be the attention the student gets when playing with a

toy, thereby making the real function, attention. As an aside, the efficacy of using

interventions to confirm or disprove functional relationships will be discussed within the

first section of this literature review, studies using solely descriptive assessment studies.









The second caveat was the issue of investigating antecedent-based assessments, at

times referred to as structural analysis, and antecedent-based interventions. The rationale

for including these studies within this review is based on the use of antecedents within

classrooms and other natural settings. It has been noted that antecedent events play an

important role in natural settings because less predictable patterns of behavior occur in

these environments, less control over antecedents occurs, and schedules of reinforcement

may not be consistently applied (Conroy & Stichter, 2003). Furthermore, it is reasonable

to assume that teachers completing the FBA process will not limit the process in the

widespread application of FBAs, solely to consequence-based assessment and

intervention. Thus, studies that addressed antecedents are included in this review.

Additional criteria to determine which research studies would be included were

also determined. First, only studies that occurred in homes, community-based childcare,

and early childhood public school classrooms were used. Because this review is focused

on the use of FBA in real-world settings, its use in highly controlled clinical settings was

not applicable. As a result, a number of articles were excluded due to this potential lack

of external validity. Second, for the first two sections, research studies directed solely

toward populations with low-incidence disabilities, such as autism, or moderate and

severe mental retardation were eliminated, as was research geared toward highly specific

behaviors that are not typical of most young children who demonstrate challenging

behavior, e.g. hair-twirling, eye-poking, breath-holding, were eliminated. Therefore, as

long as at least one participant within the study had been classified as at risk for or having

a high-incidence disability, the study was included within this review. Only the results of

the children with high incidence disabilities were discussed in the first two sections of









this review. Criterion was established to focus on the application of FBA with a broad-

based population that is, young children with or at risk for emotional/behavioral

disorders. However, for the final section of the literature review that consisted of studies

that directly compared and evaluated the results of different descriptive methodologies,

this criterion was expanded due to the dearth in this area of research. For the final section,

results of students who were older or had more severe disabilities were also reported.

Third, in order to focus on FBAs with young children, each study was to include at

least one participant 8 years of age or younger. Fourth, only research published in a peer-

reviewed journal after 1994 was included. Because this methodology has been revised

and refined extensively over the past ten years, articles pre-dating 1994 would not be

applicable to current applications. These criteria were again expanded for the final

section due to the limited number of studies in this area.

Results of the Literature Review

Thirty-two studies were found that met the criteria for inclusion in this review.

Fourteen studies used descriptive methods only to identify the hypothetical function of a

challenging behavior, ten studies used both descriptive methods and functional analysis

to identify the function of a challenging behavior, and eight studies directly compared the

results of descriptive assessments with functional analysis results. In the first two sections

of this review, (1) solely descriptive assessments (indirect and direct), and (2)

experimental analysis and descriptive combined, were evaluated by looking at: (a)

participant and setting characteristics, (b) implementer characteristics, (c) components of

the descriptive approach or analog conditions, (d) antecedents and functions identified,

and (e) interventions implemented. In the final section, each of the eight studies is

presented individually. These studies are discussed in terms of the aforementioned









characteristics, but their findings and implications are also summarized. Patterns of

agreement or disagreement are also presented as well as methodological limitations.

These eight studies are presented in chronological order. The studies in this final section

are presented in greater depth, because their rationales and methodologies provided the

foundation for the current study.

Descriptive Assessment Methodologies

This section of the paper will address the research found on the use of descriptive

assessments in the formulation of FBAs in natural settings. Descriptive assessment

methodologies, both direct and indirect, are procedures that evaluate the function of a

challenging behavior but do not systematically manipulate the environment. Therefore,

there is no experimental validation of the potential antecedents and consequences

identified. In regard to FBAs with young children, descriptive assessments including

interviews, checklists and direct observations (ABC recordings) are far more prevalent

than the more technically complex, time consuming experimental analyses (i.e.,

functional analysis and structural analyses) (Arndorfer & Miltenberger, 1993; Reid &

Nelson, 2002; Scott et al., 2004).

While descriptive assessments are less time consuming and less technically

rigorous, the main drawbacks are that they can only generate hypothetical functions of

behavior. Additionally, descriptive assessments often rely on the judgments of third party

informants. This can be especially problematic as questionnaires and interviews were

subject to the teachers' memories and biases of behavior in different settings (Shores,

Wehby & Jack, 1999). Thus the value of these assessment methodologies hinges on the

informants' accuracy of recall. Despite their current prevalence in educational settings,

there still remains a shortage of literature in regard to studies that used only the









descriptive assessment methodologies. A surprising outcome of this review is that

although IDEA (1997) produced a flurry of position papers (Sasso, et al., 2001), it did not

appear to produce an accompanying flurry of empirical research that met the criteria for

this literature review. In fact, there are six studies on or before 1997 and eight studies

after 1997. A total of fourteen studies met the inclusion criteria (see Table 2-1 for a

summary of information regarding each study). The following sections will highlight the

findings of this group of studies and discuss methodological and practical limitations of

the research.

Participant and Setting Characteristics

In terms of participant characteristics, there were a total of 274 participants in the

solely descriptive studies, however, 210 participants were from a single study (see

Chandler, Dahlquist, Repp, and Feltz, 1999). Because information about ethnicity and

socio-economic level was omitted from most of the studies, particularly those with large

numbers of participants, this level of detail is not discussed in the review. Omitting the

Chandler et al. (1999) study that did not report gender, in the smaller studies that reported

participant gender, only 9 girls (between the ages of 3-6 years old) out of the remaining

64 participants (14%) were evaluated in all of the studies. This is not surprising since

typically boys demonstrate external challenging behaviors at a much higher rate at the

preschool age than girls do (Webster-Stratton, 1997).














Table 2-1. Descriptive Methods for Young Children
Authors/ Year Participant Implementer Descriptive Functions Interventions
of Publication Characteristics Characteristics Methods Identified Implemented
(number, gender, age in
years, label) Setting
characteristics
Dunlap, 3 males, 5-11 yrs**, Researchers DO, HD, I Not stated, Use of choice
dePerczel, EBD and ADHD, self- Lack of in curriculum
Clarke Wilson contained classrooms choice recommended


Wright, White,
& Gomez
(1994)*
Storey, et al.,
(1994)





Clarke, Dunlap,
Foster-Johnson,
Childs, Wilson,
White & Vera
(1995)*

Kamps, Elllis,
Mancina,
Wyble, Greene,
& Harvey
(1995)





Grandy & Peck
(1997)


Umbreit &
Blair (1997)*

Schill,
Kratochowill,
& Elliott
(1998)

Symons,
McDonald &
Wehby (1998)*


1 male, 6 yrs, no label,
reg. ed. Classroom





3 males, 5-11 yrs**,
EBD and ADHD, self-
contained classroom or
separate work room


8 males, 2 females, 3-5
yrs, at risk for EBD,
reg. ed. Classroom or
Head Start classroom






1 male, 6 yrs, no label,
gen. ed. Classroom


1 male, 4 yrs, at risk for
EBD, childcare center

19 students, 12 male, 7
female ages 4-5 yrs, no
label, Head Start
classroom

12 students, all male, 6
yrs, labeled EBD, self-
contained classrooms


Researchers &
Teacher





Researchers &
Teachers


Teachers-
provided with
training
workshops,
Researchers
monitor and
assist


Researchers
and teacher


Researchers &
Teacher

Researchers &
Teachers


Researchers
guided
teachers
through all
FBA
assessments


I, DO


I, RR, DO,
HD


(implied)


Attention, Increased
tangibles specific
reinforcement
using a
reinforcement
system
Lack of Incorporation
interest in of student
assignments interests


DO, HD, Adult Increased
TR attention, supervision,
tangibles reinforcement
for prosocial
behavior,
reduced
attention for
inappropriate
behavior
DO, HD, Teacher Token
SA,TR attention economy (led
to teacher
attention
DO,HD,I, Preference, Pair preferred
SA choice activity with
non-preferred
DO, HD, I, Skill deficit, DRO, Time
CL attention, out, Modeling
tangible


SP, HD, Teacher
SA attention,
Varied with
antecedents


Revised
schedules,
modified
assignments,
restructured
access to
attention












Table 2-1. Continued


Authors/ Year
of Publication




Blair, Umbreit,
Bos (1999)*



Chandler,
Dahlquist,
Repp, & Feltz
(1999)




Kennedy,
Long, Jolivette,
Cox, Tang &
Thompson
(2001)






Roberts,
Marshall,
Nelson &
Albers (2001)


VanDerHeyden
Witt & Gatti
(2001)


Participant
Characteristics
(number, gender, age,
label) Setting
characteristics
4 students, 3 male, 1
female 5 yrs, at risk for
EBD, childcare center


210 students, 3-6 yrs, 3
types of classrooms
special needs (assorted
disabilities), at risk,
early childhood


3 students, 6, 6 and 8
yrs, 2 male, 1 female,
1 identified EBD, 2 'at
risk', self-contained
and gen. ed classrooms






3 students, l-1st grade,
2-4th grade**, all
males, no identified
disability, general ed.
classroom

14 students, ages 2-4
yrs, 2 classrooms,
speech delay, autism,
hypothyroidism


Implementer
Characteristics




Researcher
and teachers



Teachers-
provided with
training
workshops,
Researchers
monitor and
assist
Researchers
and teachers


Researchers
and teachers


Researchers
and teachers


Descriptive Functions Interventions
Methods Identified Implemented


DO, HD, I Attention, Integrate
choice, choice and
preferred preference
task into
curriculum
DO, HD Not stated Varied for
classroom/
child


I, DO, HD Teacher
attention,
Escape,
Topograph-
ically
specific-
functions
varied
across
contexts
I, DO, Escape
Diagnostic motivated
academic
task


I, DO, HD, Teacher
attention
(primary)
escape,
tangible,
peer
attention
(secondary)


Differential
attention,
self-
monitoring,
Eliminate
certain
consequences
(school
suspension)

Used
curriculum
based
assessments
to modify
antecedents
Differential
reinforcement
of alternative
behaviors
(DRA)


Packenham, 2 students, 1 male, 1 Teacher I, HD Teacher Task
Shute & Reid female, 8 & 9 yrs**, no trained and through DO attention, modifications
1( 1 i4) identified disability completed (ABC) escape specific
assessments (only directions,
options) card to access
teacher
attention
Note. DO=direct observation; CL=checklist; I=interview; HD=hypothesis development; RR=record
review; SA=structural analysis; SP=scatterplot; TR=teacher reports; DRO=differential reinforcement of
other behaviors
* indicates a structural or antecedent based assessment used
** indicates only students in age range included









In regard to participants' disabilities, the publication dates of these studies and their

methodologies are suggestive of a type of evolutionary process that occurs with FBA

research, namely the progression of research with more severe, low-incidence disabilities

(autism, profound mental retardation) to participants with less severe, high-incidence

disabilities (EBD, mild mental retardation). Related to participants' disabilities is the

issue of what types of behaviors were being demonstrated and addressed through

functional behavior assessment. Behaviors such as noise making, talking out, property

destruction, excessive off-task behavior and non-compliance represented the majority of

the behaviors that were addressed in this sample of studies. The settings in which these

studies took place are also a highly relevant area of consideration in regard to the

widespread application of the FBA process. The studies in this sample used descriptive

assessments in self-contained special education classrooms (39%), Head Start classrooms

(11%), general education classrooms (39%) or childcare center rooms (11%). Four

studies included different types of classrooms in the same study and Clarke et al., (1995)

also reported using a teacher workroom for part of the assessment process.

The first notable aspect of the literature in regard to practical applications of FBA

in the classrooms is the total number of participants used in this sample of studies.

Omitting the Chandler et al. (1999) study, studies averaged about 6 participants per study,

in terms of all participants (not only the ones that met the inclusion criteria for this

literature review). However, Chandler et al. (1999) identified the importance of training

teachers as the creators and implementers of the FBAs. Through a series of workshops,

teachers were given the requisite knowledge to generate effective FBAs for entire classes.









The emphasis on training for class-wide implementation is indicative of the importance

of implementer characteristics, as discussed in the next section.

The second notable aspect about this sample of studies is in regard to gender. The

minimal representation of female participants is very likely an unavoidable aspect of this

type of research. The omission of girls from all but the largest descriptive methods

studies severely limits the generalizability of the results of the FBA process to this

population. However, this proportion of males to females may be representative of the

population schools and early childhood centers identify as necessitating an FBA.

The third practical consideration that is a strong point in regard to this sample of

FBA studies is the issue of setting. FBAs that are set in highly controlled settings such as

clinics threaten the applicability of this process. This is noted in this sample as all of the

studies were conducted in natural settings. Even the use of separate rooms other than

participants' classrooms threatens the external validity of the FBA process, as it exists

today. The consequence of this artificial setting is that the generalizability of these studies

remains questionable. Because of their countless responsibilities for an entire classroom

of students, it is unlikely that teachers will be able to spend the necessary time required to

conduct one-on-one descriptive assessments of students in separate classrooms.

Furthermore, unless the student's behavior is automatically reinforced, it is likely to be

mediated in some way by social settings. Therefore, conducting assessments away from

the typical classroom or early childhood center milieu may create artificial variables not

typically found in the classroom settings. An additional consideration beyond the FBA

setting is the consideration of who actually conducted and implemented the FBA.









Implementer Characteristics

Because educators will ultimately be responsible for generating effective FBAs, it

is important to note who is generating the FBAs in research studies (Scott et al., 2004). In

all of the studies in this section of the literature review, an effort was made to include

general and special education teachers in the creation and implementation of the FBA

(see Kamps et al., 1995; VanDerHeyden, Witt & Gatti, 2001). However, in 13 out of 14

(93%) of the studies, a number of highly trained researchers and trained graduate and

undergraduate students worked with a small, select group of students (or an individual

student) and the teachers' role appeared to be very limited. It is important to insure that

this process will generalize to the personnel in early childhood centers and schools so that

the process can be implemented after researchers leave. Furthermore, this may be the

opposite of what is currently occurring in a typical classroom, which is a small number of

teachers with minimal training are responsible for creating a large number of FBAs.

While investigating who is completing FBAs is an important consideration, how the

functions of the behavior are determined is the central issue in this literature review.

Components of the Descriptive Approach

In regard to the varying components of the descriptive approach, all fourteen

studies used some combination of descriptive methods to determine a hypothetical

function of a challenging behavior. The descriptive approach may include informant

assessments such as interviews with the teachers, teaching assistants and/or the parents, a

review of permanent records, and behavior checklists. In this sample of articles: 71%

(n=10) used some type of interview, 7% (n=l) used a behavior checklist, 100% (n=14)

used direct observation, 7% (n=l) used a scatter plot and 7% (n=l) explicitly mentioned

performing a records review. All fourteen of the articles also engaged in some type of









hypothesis development in which the researchers and teachers collaborated to generate

hypothetical functions for challenging behavior. It should also be noted that studies that

evaluated antecedents of the challenging behavior used essentially the same methods of

interviewing, behavior checklists, and direct observation for extracting information about

the antecedents that precede challenging behaviors.

While underused in this sample, there are two advantages to the use of checklists.

First, unlike an interview that is open to interpretation by the individual reviewing the

answers, a hypothetical function is explicitly arrived at through the checklist scoring

process. Second, behavior checklists are highly efficient and can be completed quickly

without extra work. However, they do not have the advantage of direct observation of the

child and all information in a behavior checklist is filtered through a third party informant

(Shores, et al., 1999).

Direct observation is another descriptive approach. One of the primary methods of

this descriptive approach is the use of an A-B-C procedure, in which the observer

carefully records the antecedent, behavior, and consequence to establish a pattern of

supporting environmental stimuli (Bijou et al., 1968) that either precede or follow the

challenging behavior. Direct observation and/or generating an ABC (antecedent-

behavior-consequence) hypothesis was used in all fourteen studies. The importance of

direct observation is two-fold. First, it provides basic information that can guide other

descriptive and analog methodologies. Second, it provides the requisite baseline data to

determine if subsequent interventions are effective.

Besides the ABC analysis, another direct observation tool that has been widely

used in research to systematically observe the antecedents and consequences of









challenging behavior is the Functional Assessment Observation Form (FAOF) (O'Neill et

al., 1997). Although more technically complex then the ABC hypothesis sheet, the FAOF

allows for the observation of chains and sequences of multiple behaviors and their

antecedents and consequences. Identifying chains of behaviors can give insight into

whether the challenging behaviors often occur in a hierarchy going from low-intensity to

high-intensity (Chandler & Dahlquist, 2002). For example, a student begins poking his

neighbor to avoid participating in 'Circle Time.' The teacher verbally prompts the student

to stop bothering his neighbor. The student ignores the redirection and begins stomping

the floor with his feet. When the teacher asks the student to get up, the student responds

by screaming at the teacher and flipping over a nearby chair. Without careful

identification of the target behavior, such as 'participating in group activities,' the focus

may be placed solely on the high intensity behavior, when in fact the initial behavior,

avoiding 'Circle Time,' is a more likely target for intervention strategies. In such cases,

interventions can be designed that address the initial challenging behavior early in the

chain before high intensity behavior occurs. In addition, the identification of behavioral

chains provides evidence that is useful in analyzing environmental events that may

precipitate all challenging behaviors. One important methodological consideration noted

in Kennedy et al. (2001) and Grandy and Peck (1997) is that these direct observations

should occur across settings and at different times of the day. The value of this

consideration is that it provides important contextual information regarding the function

of the behavior and reinforces the fact that behavior and its functions are often

contextually specific.









A teacher interview is another prevalent indirect descriptive method. The use of a

semi-structured or unstructured interview with a teacher provided important insight into

the possible functions of a behavior for a number of studies that used only a descriptive

methodology (e.g. Grandy & Peck, 1997; Kamps et al., 1995). The most common

interview used was the Functional Assessment Interview (FAI). The Functional

Assessment Interview (O'Neill et al., 1997) is a multi-step interview form designed to

identify antecedents of the challenging behavior and establish the function of a child's

challenging behavior. This is a comprehensive interview that has some empirical testing

(Kinch, Lewis-Palmer, Hagan-Burke, Sugai, 2001) and is described by Gresham, Watson

and Skinner (2001) as "the most comprehensive, up-to-date treatment of this topic" (p.

162).

Some different methods of assessment such as a careful review of records (Clarke

et al., 1995) and the use of a scatter plot (Symons, McDonald & Wehby, 1998) go

relatively unexplored in regard to research studies. In Packenham, Shute and Reid (2004),

one interesting point in regard to the use of formal descriptive assessments was that the

teacher who participated in the study said "I already do all of this (formal descriptive

assessments), I just never formalized it" (p.23). In light of this consideration, the need for

less formal descriptive assessments and more effective ways to access teacher knowledge

could be an important step in future research. Additionally, continuing to monitor

teachers' perceptions of the FBA process is also an important aspect of current and future

research. In this sample of studies, 6 out of 14 (43%) studies administered some type of

treatment acceptability questionnaire following the study. Finding assessments that









teachers are willing to implement is as important as finding assessments that accurately

identify the functions of challenging behavior.

Functions Identified

Accurately reporting descriptive statistics regarding the determined function for

these studies was not feasible for two reasons. First, in the largest study, Chandler et al.

(1999), the function of the challenging behavior for the 210 participants was not reported.

Second, other studies did not directly report functions of behaviors. For example, some

studies reported possible functions by calculating conditional probabilities between the

target behavior and the consequences (i.e., Kennedy et al., 2001) or other outcomes as

primary and secondary functions (i.e., Van Der Heyden et al., 2001). Therefore, only the

most general observations can be made about patterns in this sample of literature

regarding the three functions of behavior. This is unfortunate as evaluating base rates of

behavioral functions across different populations may be an important area of future

research (Gresham, 2003).

The functions of a target behavior that cause the behavior to increase in frequency

are typically classified into three types. First, external positive reinforcement includes

attention from teacher or peers or tangible rewards (Lalli & Goh, 1993). In this sample,

this was the predominate function of challenging behavior- largely teacher attention,

followed by peer attention and tangible items. Second, external negative reinforcement

includes escape or avoidance of an unpleasant or undesirable stimulus, such as a difficult

academic task. A third type of reinforcement is automatic reinforcement, in which the

reinforcement is hypothesized to be internal and biologic in nature. This type of

reinforcement often explains self-injurious or stereotypic behavior (Mace & Roberts,

1993) and was not a function of behavior for any of the participants in this sample of the









literature. The environmental relationships surrounding a challenging behavior can be

further classified in terms of whether the variable precedes the behavior, which is an

antecedent, or a reinforcer that follows the behavior, which is a consequence or a function

of the behavior. Choice or lack of choice was the predominate antecedent that was

manipulated in the five studies that evaluated antecedents in this sample.

With the exception of the predominance of teacher attention as the function of the

behavior, there was no distinguishable pattern of the fourteen studies investigated in

terms of positive versus negative reinforcement or antecedent versus consequence. The

studies varied in terms of their participants. However, some general practical and

methodological observations can be made. One notable methodological problem in the

research was in regard to the functions determined within this group of studies. These

studies often generated, or at least reported, functions that were very general. For

example, Schill, Kratochwill and Elliott (1998) describe the function of five of their

participants as "attention from adults and/or peers." (p.128). This lack of specificity is

problematic for a number of reasons. First, the lack of specificity makes it difficult to

determine the appropriateness of the intervention. In other words, simply saying that

attention was the function of the challenging behavior leaves the reader to wonder if

access to any type of attention would ameliorate the challenging behavior or does the

student require a specific intervention? To this point, the Schill et al. (1998) study did not

find a statistically significant difference between the effects of function-based

interventions and non-function based interventions. Typically, the greater the amount of

specificity in all areas of applied behavior analysis, the more effectively behavior can be

measured and modified. Determining the function of a behavior should lead to the









identification of an intervention to implement. If a consequence of a challenging behavior

were positive reinforcement through increased adult attention, then an effective

intervention would most likely be decreasing attention for challenging behaviors and

increasing attention for adaptive or pro-social behaviors. Ultimately, the goal of

accurately establishing the function of a behavior is to implement effective interventions.

Interventions Implemented

All of the studies recommended interventions and 93% (n=13) actually

implemented the interventions. Descriptive methods within these studies generated

similar (but participant-specific) interventions and all of the interventions reported at

least initial success at stopping the challenging behaviors to some degree. Examples of

the interventions were increasing teacher supervision (Kamps et al, 1995), increasing

reinforcement for different behaviors (Schill, et al., 1998; VanDerHeyden, et al., 2001),

and embedding preferred activities into existing curriculum (Blair, Umbreit, & Bos,

1999). A more notable aspect of these studies was the follow-up that occurred some time

later. Packenham et al. (2004) reported a continued reduction of the challenging behavior

five weeks later due to suggestion made as a result of the functional behavior assessment.

This finding is surprising, considering Reid (2000) questions whether or not teachers will

persist with interventions that are generated by functional assessment over time.

One criticism of specific articles reviewed was a lack of direct congruence between

the identified function and the intervention implemented. For example, the intervention

of increasing attention to desired behaviors and decreasing attention to challenging

behaviors, (e.g. Kennedy et al. 2001; Storey, et al., 1994), while effective, did little to

highlight the need for an individual FBA unless this intervention directly matches the

antecedents and consequences identified through the FBA process surrounding the









individual target child. A different version of the same criticism was the use of multiple

interventions that may or may not address the same function, but are implemented at the

same time making it impossible to determine what intervention was actually addressing

the identified function.

This lack of distinct function-based interventions is also seen with the insertion of a

mediating intervention between the desired behavior and the reinforcement that addressed

the function of the behavior. For example, Grandy and Peck (1997) determined that the

function of a challenging behavior was teacher attention. However, rather then implement

an intervention that directly linked teacher attention with desired behaviors, the

researchers established a token economy, which targeted obtaining a tangible item, but

indirectly led to increased teacher attention. This lack of direct connection between the

determined function and intervention implemented calls into question the accuracy of the

determined function or the potency of reinforcers that do not necessarily match function

(e.g. tokens). It is impossible to determine if teacher attention was reinforcing the desired

behavior or if it was another element of the token economy.

Finally, the use of interventions to confirm hypothesized functions is not

methodologically problematic. However, what are problematic are the conclusions that

authors of these studies draw from the use of these interventions. For example, by using a

reversal design in the implementation of a function-based intervention, Kamps et al.

(1995) failed to differentiate between a functional relationship that was established and

proven effective and the actual functional relationship that exists for a student's

challenging behavior. In other words, just because an intervention worked does not mean

it addressed the primary function of the challenging behavior, it simply means that this









intervention is powerful enough to be effective even if it does not directly address the

actual function of the challenging behavior. While this is a subtle distinction, it points to

one of the cornerstones of applied behavior analysis which is the need for precision in

language and in all aspects of empirical studies, or as Baer, Wolf and Risley (1968)

described it, the technological dimension of empirical studies.

Summary

Descriptive assessment methods represent the vast majority of the assessments

being used in FBAs. The wide range of descriptive assessments being used singularly or

in varying combinations for the preschoolers in this review also indicate that no one

assessment method or combination of assessment methods has emerged as the

preeminent process for young children with or at risk for EBD. There are two additional

conclusions that can be drawn from a review of these studies.

First and foremost, the issue of methods for accurately determining the function of

challenging behavior remains unresolved. There are two major obstacles to resolving the

question of what will be the most effective protocol of descriptive assessments for FBAs.

First, there is the continued combining of different assessment results in a single study. In

other words, studies continue to use interviews, checklists and direct observations

without separating what procedures are giving consistent information with other

assessments. The second obstacle is the fact that descriptive assessments can only yield

hypothetical functions and there is no way to confirm the accuracy of any method without

providing some type of environmental manipulation. In order to determine what is best in

the widespread application of FBA, it is necessary to determine what components are the

most accurate when compared with these environmental manipulations (i.e., a functional

analysis).









Second after the most accurate ways to determine function in widespread

applications are determined, research should address the issue of function-based

interventions. This must be done after the process to accurately determine function has

been empirically tested, because function-based interventions are unattainable without

accurately determining the function first. This current research sample indicates that there

is a gap between the research that has been done to accurately identify the function of a

challenging behavior and the research that investigates the use of function-based

interventions. Future research needs to concentrate on the implementation of

recommended interventions and the rationales that they are based on. It is imperative that

researchers delineate between individualized function-based interventions and the

recommendation of improved classroom management skills. Researchers and

practitioners must recognize that even if a methodology is highly accurate at determining

the function of a challenging behavior, but does not lead to highly effective interventions,

than the methodology is essentially useless. This leads to the next section of this literature

review that includes empirical studies that combined descriptive assessments with

functional analysis.

Descriptive Assessments and Functional and Structural Analyses within FBAs

The use of experimental analysis has been widely documented over the past twenty

years. While analog assessment is a frequent design in studies in many behavioral

journals, particularly the prestigious Journal ofApplied Behavior Analysis, it has not

crossed the research to practice gap in the widespread application of FBA in public

schools (McKerchar & Thompson, 2004; Sasso et al., 2001). Explanations for this lack of

widespread application vary, but clearly the technical complexities and the additional

time and staffing demands that are required to complete these procedures are difficult to









overcome. Furthermore, aspects of functional analysis may seem counterintuitive to

many practitioners, especially when conducting functional analysis that manipulate the

environment in an effort to elicit, as opposed to prohibit, certain behaviors. Thus, there

continues to be a limited amount of research in the use of functional analysis in

classroom settings (Ervin et al., 2001).

Illustrating this point, the research base did not translate to a large number of

studies within this literature review. Although many studies have used functional or

structural analyses with young children, the two criteria that eliminated the majority of

the studies were the use of these analyses in natural settings and their use with children

with high incidence disabilities. Functional analysis, particularly, continues to be used

largely in clinical settings and with more individuals with severe developmental

disabilities. In three of the studies that are reviewed, researchers included participants

with high-incidence and low incidence disabilities (Ellis & Magee, 1999; Harding et al.,

1999; Romaniuk et al., 2002); only the high-incidence participants are discussed in this

review. Ten studies met the criteria for inclusion in this section of the literature review

(see Table 2-2 for a summary of information regarding each study). The following

sections will outline the findings and discuss methodological and practical limitations of

the research.











Table 2-2. Descriptive Methods and Functional Analysis for Young Children
Authors/ Year Participant Implementer Descriptive Antecedent Interventions
of Publication Characteristics Characteristics Methods and Functions Implemented
(number, gender, Functional Identified
age, label) Setting Analysis
characteristics


Broussard &
Northrup,
(1995)


Umbreit
(1996)


Broussard &
Northrup
(1997)

Ellis &
Magee (1999)


Harding,
Wacker,
Cooper,
Asmus,
Jensen-
Kovalan, &
Grisolano
(1999)*
Blair,
Umbreit &
Eck (2000)

Boyajian,
Dupaul,
Handler,
Eckert, &
McGoey
(2001)
Doggett,
Edwards,
Moore,
Tingstrom &
Wilcski(2001)


3 males, 6-9 yrs, at
risk for ADHD or
EBD, empty
classroom and
primary classroom


1 male, 5 yrs, mild
MR, daycare center


4 males, 6-9 yrs,
ADHD or at risk for
EBD or ADHD,
primary classroom
3 males, 6-10 yrs,
mild autism, PDD,
EBD, ADHD,
separate classroom
and primary
classroom


3 males, 4 yrs.,
ADD, sight and
hearing, at risk for
EBD, primary
classroom




1 male, 4 yrs., at
risk for EBD,
primary classroom

3 males, 4-5 yrs., at
risk for ADHD,
primary classroom




2 males, 6 & 7 yrs.
old, no identified
disability, primary
classroom


Teachers &
researchers


Teachers &
researchers
(only training &
data collection)
Teachers &
Researchers


Researchers
(separate room
FAs) and
teachers (in
class FAs)




Teachers &
researchers







Teachers &
researchers


Researchers and
teachers





Teachers
(teachers
implemented
BFAs) &
researchers


DO, HD,
BFA, 3 analog
settings based
on descriptive
analysis


HD, I, DO,
BFA-4
conditions

DO, HD, and
BFA


DO, HD, FA-4
conditions
(incl. peer
attn.,
competition
for teacher
attn.)

DO, SA, BFA-
4 analog
conditions






DO, I, HD,
FA- 5
conditions

DO, I, BFA- 4
conditions





DO, I, BFAs-2
conditions


Teacher
attention,
peer
attention,
escape
academic
task
Escape
difficult
tasks

Peer
attention


Peer and
adult
attention,
escape
from
demands


Specific v.
general
directions,
social
attention,
reprimands


Contingency
reversal No
extended
intervention
implemented


Modified
curriculum,
Social skills
training
Token
economy,
DRO

Token
economy,
Social skills
training, task
modification


More specific
directions,
increased
attention for
prosocial
behavior


Attain Social Skills
tangibles training, No
other
interventions
Attention, Contingency
Tangible, reversal over
Escape longer period
of time (11-18
days)


Teacher
attention,
peer
attention


No
interventions










Table 2-2. Continued
Authors/ Year Participant Implementer Descriptive Antecedent Interventions
of Publication Characteristics Characteristics Methods and Functions Implemented
(number, gender, Functional Identified
age, label) Setting Analysis
characteristics

Romaniuk, 7 participants, 5-10 Researchers and I, Preliminary Choice, No No
Miltenberger, yrs, identified for therapists FA, BFAs-4 Choice, interventions
Conyers, problem behaviors conditions Attention,
Jenner, also some Escape
Jergens & participants Mosaic
Ringenberger Down syndrome or
(2002) mental retardation,
separate classroom,
Stichter, 1 male, 7 yrs., Researchers, DO, BSA- 3 Low social Integrate
Sasso & EBD, primary spec. Teachers structural demands, antecedents
Jolivette ed. and gen. ed. provide variables high that decrease
(I'"14)* classrooms descriptive structure challenging
information behavior into
school day
Note. DO=direct observation; CL=checklist; I=interview; HD=hypothesis development; FA=functional
analysis; BFA=brief functional analysis; RR=records review; SA=structural analysis; DRO=differential
reinforcement of other behaviors
* indicates a structural analysis

Participant and Setting Characteristics

The total number of participants for all ten of the studies was 28 participants or an

average of 2.8 participants per study with a range of 1-7 participants. Only three of the

participants in all ten of the studies were female 11% (n=3). Furthermore, while all

studies included at least one participant who met the criteria of a high-incidence

disability, three of the studies included participants with low-incidence disabilities. For

example, in Ellis and Magee (1999) 1 participant had Attention Deficit-Hyperactivity

Disorder (ADHD), but the other 2 participants were labeled with mild autism and

pervasive developmental disorder (PDD). In Romaniuk et al. (2002) 3 of the 7

participants were diagnosed with Down syndrome or mental retardation. In Harding, et

al. (1999) 1 participant had sight and hearing disabilities and 1 was identified with

developmental disabilities. This indicates that functional analysis may have a wider range









of applicability with individuals with high-incidence disabilities than the research base

reflects. In regard to practical limitations of this literature, the limited number of

participants within these research studies is the first indication of the impracticality in

making functional analysis an integral part of early childhood settings. The study with the

most participants, Romaniuk et al., (2002) (n=7), is one of the most recent studies in this

literature review. While this may suggest the possibility of widespread use of functional

analysis within the classroom, there is very little other empirical research. Typically,

functional analysis is time intensive and is completed on one individual child at a time.

One possible explanation is that a low participant size is maintained as researchers place

more emphasis on honing the methodology, exercising environmental control, or

initiating novel interventions than addressing the need for widespread application.

Another possibility is that functional analysis is only appropriate for students needing the

most intensive assessments and interventions. It should be noted that studies ten years

ago (e.g., Broussard & Northup, 1995) acknowledged the need for widespread

application to more students in naturalized settings, although they continued to use very

small sample sizes.

The second observation in regard to participants and settings is that the use of

analogs in primary classrooms during regular classroom activities indicates some

modification from the controlled, structured nature of analogs to more natural, less

restrictive settings. One study that clearly modeled this assimilation is the Ellis and

Magee (1999) study that conducted environmental manipulations both in the classroom

and in a separate environment.









Implementer Characteristics

The roles of researchers and teachers took many forms within this sample of

studies. In six (60%) of the studies, the participants' teachers or other school personnel

were trained and conducted the analog conditions of the functional analysis. For three of

the studies (30%), the researcher conducted the analog conditions. For the Ellis and

Magee (1999) study, the researcher conducted the analogs in the separate room and the

teacher conducted them in the classroom.

While researcher influence was also observed as a limitation for 'descriptive

assessment only' studies, it was even more pronounced in functional and structural

analysis trials. Between training teachers and/or conducting analog conditions,

researcher influence was even more evident in these studies than the studies in the

previous section. This is not surprising due to the highly technical demands of these types

of analysis. However, this is important because the greater the influence researchers have

on the design and implementation, the more difficult it will be to generalize the findings

to teachers as implementers and determine external validity. Some studies included in the

third section of this literature review have created protocols that teach teachers how to

implement a functional analysis. Such creative methodologies and their effectiveness

may be an important part of future research (Sasso et al., 1993).

In fact, the three studies that used researchers to implement the analog conditions

introduced considerable bias considering the number of differences that a researcher

conducting the analog conditions may have versus the teacher conducting analog

conditions. For example, they did not evaluate whether adult attention from a novel

person (the researcher) is more or less powerful than adult attention from a teacher. This

ignores all of the idiosyncratic and qualitative variables that may be influencing the









occurrence for the behavior (i.e., tone of voice, appearance, person preference and so

forth) that are introduced with researchers conducting the analogs. Finally, the issue of

external validity is brought into question. Since these studies were not designed to

address long-term widespread application, it is not necessarily a criticism of these

specific studies but rather a shortcoming of the research base as a whole. The studies that

used researchers to conduct the analogs were actually focused on a specific disability,

such as ADHD (Boyajian et al, 2001) or the effects of antecedents as an environmental

manipulation (Stichter, Sasso & Jolivette, 2004) and were not directed toward examining

the implementation of FBA in school settings. Rather these studies were geared to

investigate other aspects of structural and functional analysis and their applicability with

individuals with a specific disability or the overall value of the use of antecedent

manipulations. The next consideration is how the actual analog procedure took place and

what components comprised its protocol.

Components of the Descriptive Approach and Analog Conditions

The preliminary descriptive components were similar to the sample of studies that

only used descriptive methodologies. All of the studies (100%) used direct observations

and some type of hypothesis development. Five of the studies (50%) also used a

structured interview with the teacher and/ or other individuals who have contact with the

participants. The only salient difference between the descriptive methods in this sample

of studies and the sample of studies that only used descriptive methods is that this sample

of studies was more parsimonious in its use of descriptive assessments. In evaluating the

descriptive assessments listed in these 10 studies, it appeared that the descriptive

assessments were only selected and used to the extent that they could guide the functional

or structural analysis procedures. These studies used the descriptive assessment results to









identify times and situations to conduct analog probes during the times or activities that

teachers identified as the most problematic. For example, both Broussard and Northup

(1995, 1997) only used direct observations of the participant and a brief interview with

the teacher to determine what analog conditions to conduct.

In studies using functional analysis, the five 'traditional' analog conditions were as

varied as the participants with whom they were being implemented. The studies often

included variations of the following conditions (1) alone or ignoring, (2) attention, (3)

tangible reward, (4) escape or avoidance of a task and (5) free play. These are the classic

conditions established by Iwata (1982/1994) and other researchers using functional

analysis. Besides using two different settings, classroom and separate room, Ellis and

Magee (1998) described another notable variation of their analog conditions. Two

alternative analog settings were implemented: (1) peer attention with the use of a

confederate peer attending the target behavior and (2) peer competition for teacher

attention. These two contrived analogs represent another step toward integrating

functional analysis into more natural settings and provide a more accurate means of

assessment using functional analysis. The results indicated that peer attention did play a

role in the function of one participant's target behavior.

Romaniuk et al. (2002) provided another addition to the classic analogs by the

testing choice or 'control' as a precursor of challenging behavior; this manipulation is

considered a structural analysis. To test control, this study offered the participant the

option to choose the activity in one condition and did not allow them to choose the

activity in another condition. Not being able to choose the activity was the setting event

or antecedent of the participant's challenging behaviors. The examination of choice or









'control' as a possible pre is similar to the Blair, Umbreit and Bos (1999), a study

reviewed in the 'descriptive methods only' section of this literature review.

The methodology of all of the functional analyses studies differed from those that

were originally conceived of in Iwata et al. (1982/1994); because they took far less time,

used descriptive methods to precede the analog settings, and were subject to more of the

situations that occur in natural settings. Thus, while the roots of functional analysis can

be traced back to Iwata et al. (1982/1994), the process has clearly evolved. This evolution

also includes the use of structural analyses, manipulating antecedents in much the same

way that a functional analysis manipulates consequences. This process was used with

two studies, Harding et al. (1999) and Stichter et al. (2004). In Harding et al. (1999),

variables such as general directions versus specific directions and group activity versus

independent activity were manipulated. For Stichter et al. (2004), variables such as noise,

structure, and levels of socialization were manipulated. Whether investigating

antecedents or consequences, the important issue in these studies is determining the

environmental factors that are related to the challenging behavior so that the functions

may be accurately identified.

Functions Identified

As noted in Broussard and Northup (1995), the most prevalent functions typically

identified, in general, are teacher attention, peer attention, and escape from academic

demands. An overview of this sample of studies confirms this assertion and, similar to the

sample of studies that used descriptive assessments only, it is impossible to descriptively

summarize the most prevalent functions using descriptive statistics. However, it can be

observed that no other pattern of identified functions emerged. In other words, there was

not a clear difference in the number of participants whose function was either positively









reinforced (i.e., attention or tangible) or negative reinforced (i.e., escape). Furthermore,

the explanation for why it is so difficult to make summative statements is a notable

strength of this sample of studies. Namely, the high degree of specificity of function

makes summative statements across participants difficult. A difference between this

group of studies and the 'descriptive methods only' studies is that the functions that

functional analysis determines are more specifically and precisely defined. Proponents of

functional analysis view the lack of specificity with descriptive assessments as

problematic. They assert that as functions are more generally defined, the intervention

becomes more idiosyncratic and less effective (Sasso et al., 1992). Harding et al. (1999)

noted that in their comparison of the two methodologies, the descriptive methodology

produced results that were "difficult to interpret" (p. 329) and "brief experimental

analyses were needed to clarify functional relations" (p.330). They did note that this was

a result of not being able to observe the target behavior during descriptive analysis.

Another direction of the research within this sample of studies is the combining of

structural analysis and functional analysis. As Romaniuk et al. (2002) points out,

combinations of antecedents and established functions have not been extensively

investigated through research methods. In this study, the antecedent of choice or control

over tasks had different effects determined by whether the function of the challenging

behavior was attention or escape. This study indicated that choice could serve as a

powerful antecedent when the function of a challenging behavior was escape. Harding et

al. (1999) used a similar design of measuring both antecedents and consequences. The

next logical step after the accurate identification of the functions of challenging behavior









is the implementation of effective interventions that use the information gathered through

these methods.

Interventions Implemented

Multiple types of interventions were used with these studies. Functional analysis

methodology creates a complex issue when evaluating the interventions used with these

studies. Within the ten articles reviewed, following the functional and structural analysis,

an environmental manipulation was conducted that resulted in reducing the challenging

behavior using a contingency reversal. There were also a variety of other interventions

that occurred following the functional analysis. Five studies (50%) reported reducing the

occurrence of challenging behaviors during the intervention phase, also referred to as the

contingency reversal phase. However, the other 5 studies used other interventions

including differential reinforcement of other behaviors (DRO), and implementing token

economies.

It should be noted that Boyajian et al. (2001) reported a continued reduction of the

challenging behavior six months later due to implementing contingency reversal elements

into the milieu; thus, the contingency reversal became the intervention that was

implemented over time. Contingency reversal involves establishing an environment that

increases the rate of challenging behavior and implementing a phase of the opposite

condition. For example, because a challenging behavior increases during a 'no choice,

high academic demand' condition, the reversal contingency would be 'participant choice

and low academic demand.' While this reversal accomplished the desired result, it may

not be a situation that is appropriate across situations in the typical preschool

environment. The reason this situation may not be appropriate to generalize to typical

preschool environments is that at some point it is necessary to give children tasks that are









not preferred and to ask children to participate in activities that are not preferred. It

should also be noted that even if an antecedent based intervention is implemented that

decreases the probability of the challenging behavior, one will still have to implement an

intervention that directly addresses the occurrence of the challenging behavior at some

point.

The fact that a reversal contingency may not be appropriate for implementation in

an early childhood environment illustrates a practical limitation in studies with this type

of design. As Blair, Umbreit and Eck (2000) indicate, the reversal of the problematic

contingency was not feasible in preschool settings, because it is not always possible to

put children in the situations that will reduce challenging behaviors (i.e., always making

sure there were a positive ratio of available toys to peers). In the natural setting of a

preschool classroom, there will be times when there are not enough toys for everyone, or

when access to a preferred toy may not be possible because another child is playing with

it. Further, insuring that this does not ever happen to the student does not allow the

participant to gain appropriate adaptive or social skills. It may be more appropriate to

increase the amount of social skill instruction that precedes a functional behavior

assessment to alleviate its use and then to have specific interventions in place that address

the challenging behaviors after they occur.

A notable aspect of these studies was the follow-up that occurred some time later.

Harding et al. (1999), Boyajian et al. (2001), and Ellis and Magee (1999) reported

continued decreases in challenging behaviors for the participants as they followed up to

six months later. The durability of these prescribed interventions is encouraging. As FBA

research continues, more and more examples of effective interventions will be devised









and illustrated, allowing practitioners to make direct connections between the identified

function and the intervention that is implemented.

Summary

The use of functional and structural analysis with young children with or at risk for

EBD remains an important area of future research. Clearly, the research is still in the

process of evolving from its setting in clinics with individuals with severe disabilities to

settings with children in classrooms with mild disabilities. However, it is difficult to

determine what role the methodical manipulation of the environment will play in the

widespread application of functional analysis methodology in FBAs. Thus far, this

methodology has not become a part of the protocol in the formulation of FBAs in

schools. The concept of teachers deliberately manipulating the environment and tracking

the increase or decrease of challenging behavior may never be feasible. In this sample of

studies, functional analysis was used as one more method, albeit typically the most

decisive and trustworthy method, to triangulate assessment methods and accurately

determine the function of challenging behavior. There are three salient points that can be

drawn from this sample of the literature.

First, similar to the first set of studies, no study with functional analysis procedures

could be conducted with large numbers of children in the same setting, thus there

continues to be small sample sizes and little replication has occurred. Researcher

involvement and highly specific but varied methodologies make it difficult to foresee

what role functional analysis will play in school and early childhood settings based on

this sample of literature. The issue of data collection is also an important consideration.

In all of these studies, functional analysis was a multi-person job and, while it was

encouraging to see teachers taking an active role in conducting the analogs, it was still









necessary to have trained observers count and record the occurrences of the behavior. It is

impossible to know whether early childhood centers and schools will be able to dedicate

the appropriate amount of personnel for training and data collection that is necessary to

both conduct the analog conditions but also to record the occurrences of behaviors. To

the second point, if the interventions are not feasible or lack social validity, this

dedication of time, personnel, and resources to functional analysis may not even be

desirable.

Second, in each of these studies, functional analysis procedures depended on

descriptive assessments and discerning the results from the different descriptive methods

and the functional analysis were impossible. This leads to the question 'Was the

functional analysis, as part of an FBA protocol, necessary if descriptive assessments

arrived at the same function?' Additionally, a core principle of functional analysis is

environmental control (Sasso et al., 2001) and functional analysis is only as effective as

the amount of control that can be obtained. The difficulty is that classrooms by nature are

free operant environments, especially compared with clinics (Scott et al. 2004) and

establishing necessary environmental control may be difficult for educators with limited

time and resources. For this reason, functional analysis may not be appropriate as one

more component of a FBA protocol. It is important for future research to continue to use

more classroom settings and to increase the amount of teacher influence to evaluate the

role of functional analysis in classroom protocols.

Third, while functional analysis can develop more specific functions that lead to

more effective interventions (Harding et al., 1999), they can also lead to interventions

that are not practical or appropriate for implementation in the natural setting (Blair,









Umbreit & Eck, 2000). In consideration of these points, the ultimate role of functional

analysis may be as an effective method to evaluate the accuracy of less obtrusive

descriptive methodologies, however, it is difficult to see its widespread adoption as a

component of a standard FBA protocol. Unlike the studies reviewed in this section that

combined functional analysis and descriptive methodologies as a protocol to determine

function, the next section will separately evaluate different descriptive methodologies

using functional analysis as the standard to determine their accuracy.

Direct Evaluation of Descriptive Methodologies

A search for articles that directly evaluated different descriptive methodologies

indicates a dearth of research in this area. Only eight studies were identified that

evaluated the reliability of different descriptive methodologies when comparing the

results with each other, such as comparing the results of a direct observation and a

teacher interview. Furthermore, only six out of these eight evaluated the validity of these

descriptive methodologies by comparing them to a functional analysis. Murdock, et al.,

(2005) and Ellingson et al. (2000) only compared different types of descriptive

assessments with each other. This paucity in the research is particularly evident with

children who are diagnosed or at risk for EBD (Gresham, 2003; Gresham et al. 2004;

Fox, Conroy, Heckaman, 1998; Sasso et al., 2001). Due to the limited research in this

area, it was necessary to expand the exclusion criteria beyond young children with EBD

to include all ability and age levels of children. In fact, all but one of the studies included

participants with more severe developmental disabilities and the age range was expanded

to even include young adults, some of whom were evaluated in residential and vocational

settings. Additionally, the criterion regarding the age of the research was also expanded

from ten years to 15 years.









Finally, unlike the previous two bodies of literature, the interventions were not

evaluated within this section of the study, because the focus of this review is to determine

how accurately functions are identified via various descriptive methodologies. Evaluating

function-based interventions will be an important area of future research but that area is

premature until the methods for accurately identifying functions have been evaluated.

Further, the rationale for omitting this aspect of the research study was to provide greater

focus on the methodologies used to evaluate the assessments rather than evaluating the

resulting interventions; therefore, the focus of this section is to examine the portion of the

study that attempted to determine a primary function of the challenging behavior.

Function-based interventions will be an important area of future research; however,

research methods must first empirically determine whether the assessment methods can

accurately arrive at the function of the challenging behavior.

All other criteria that were established for this review of the literature were

maintained. One important criterion that was maintained between this body of research

and the other two areas of research was limiting the specific target behaviors to exclude

research geared toward highly specific behaviors that are not typical of most young

children who demonstrate challenging behavior (e.g. self-injurious behavior, stereotypy,

hair-twirling, eye-poking, breath-holding) were eliminated. Therefore, studies that

compared descriptive assessments with functional analyses for bizarre speech (Mace &

Lalli, 1991) stereotypy (Crawford, Brockel, Schauss, & Miltenberger, 1992; Paclawskyj,

Matson, Rush, Smalls & Vollmer, 2001) or self-injurious behavior (Belfiore, Browder &

Lin, 1993; Lerman & Iwata, 1993; Paclawskyj et al., 2001) for the majority of the

participants were excluded.









Because the different assessment methods were evaluated separately and, in some

cases, compared with the results of a functional analysis, these eight studies are unlike the

other studies that used descriptive assessments and functional analysis in the FBA

process. The studies in the previous two sections simply used various descriptive

assessment methods and assimilated the results to develop a hypothetical function and/or

guide a functional analysis. The studies that evaluated the different descriptive

assessment methods listed the results of each method separately and reported agreement

or disagreement between the methods. Because these studies more closely resemble the

study being proposed, these eight studies will be presented in a study-by-study basis, in

reverse chronological order, presenting the most recent study first (see Table 2-3 for a

summary of information regarding each of these studies).

In Murdock, et al., (2005), eight junior high school students (7 classified as EBD, 1

classified as LD) were evaluated using three descriptive assessment methods. The three

methods were a teacher team interview, a student interview and classroom observations.

No functional analysis procedures were used. First, a team of teachers who worked with

an individual student was assembled and the first author used a questionnaire to guide the

interview process regarding specific challenging behaviors and the environmental factors

that surrounded the behaviors. The responses were condensed into summary statements

that included the challenging behaviors, and both antecedents and hypothetical functions.















Table 2-3. Direct Evaluations of Descriptive Assessments
Studies Participants: Setting Who completed the What descriptive Agreement How was the Agreement
Age, Gender, FBA? assessments were between functional between FA
Disability used? descriptive analysis and descriptive
assessments completed? assessments
Murdock, O'Neill -8 junior high General ed and -Teacher Team Teacher FAI, 9/ 14 (64%) N/A N/A


& Cunningham
(2005)


Newcomer &
Lewis (2'1"4)









Cunningham &
O'Neill (2000)


school age
males
-7 classified
EBD
-1 classified LD


-9 years, Male,
OHI
-11 years, Male,
at-risk
-11 years,
Female,
at risk



-3 years, Male,
Autism
-5 years,
Male, Autism
-5 years Male,
Autism


resource
classrooms


General education
classrooms









Self-contained
classrooms


completed
questionnaire
-Student
questionnaire
-Researchers and
paraprofessional
conducted direct
observations


Researchers
administered FAI
and PBQ to
teachers and
students.
Researchers
completed direct
observations and
FAs

Researchers
administered
descriptive
assessments to
teacher and
classroom assistant.
Teacher completed
the FAO.
Researchers
completed FAs.


Student-FAI and
FAOF


FAI (S-FAI),
PBQ, Scatterplot,
ABC








MAS, FAI, FAO,
Informal
question


agreement on all
three
13/ 14 (93%)
agreement
between teacher
FAI and FAOF


Assessments
were consistent
except for PBQ
results with 2/3
participants






Good consistency
across methods
and participants.
Used rank orders
for hypothetical
functions


Based on
descriptive
assessments,
each participant
placed in one
indicated and
one control
condition


Each participant
placed in 5
analog
conditions in a
separate school
multi-purpose
room


Agreement
between
descriptive
assessments
and functional
analysis for 2/3
participants



Rank order
indicates good
agreement of
primary
function
between
descriptive
assessments
and FA for 2/3
participants















Table 2-3. Continued
Studies Participants: Setting Who completed the What descriptive Agreement How was the Agreement
Age, Gender, FBA? assessments were between functional between FA
Disability used? descriptive analysis and descriptive
assessments completed? assessments
Ellingson, -19 years, Self-contained Researchers Questionnaire, Questionnaires N/A N/A


Miltenberger et al.
(2000)








Yarbrough & Carr
(2000)


Female, Severe
M.R.
-18 years, Male,
Angelman's
syndrome &
profound M.R.
-12 years
Male, Severe
M.R.
-18.33 years
Male, Autism &
Profound M.R.
-11.5 years.
Female, Autism
& Profound
M.R.
-12.4 years,
Female
DeLange
Syndrome &
Moderate M.R.


classrooms


Self-contatined
classrooms,
vocational and
community
placements


administered
questionnaire to
teachers and
teachers completed
questionnaire
independently.
Conducted direct
observations

Researchers
administered
adapted FAI to
teachers and
classroom
assistants and
conducted
functional analysis


Interview, and
ABC


and interviews
consistent with
each other.
Results from
ABC were not
consistent with
questionnaire
results


FAI followed by
a Likert scale
rating the
resulting
hypothesized
functions. FAI
also generated
high and low
likelihood
antecedents


Based on FAI
information
high and low
likelihood
antecedents
manipulated in
natural settings
and one
hypothesized
function is
implemented
across
antecedents


Agreement
between FAI
and FAs
consistent for
high likelihood
but not for low
likelihood
antecedents















Table 2-3. Continued
Studies Participants: Setting Who completed the What descriptive Agreement How was the Agreement
Age, Gender, FBA? assessments were between functional between FA
Disability used? descriptive analysis and descriptive
assessments completed? assessments


Calloway &
Simpson (1998)


Amdorfer,
Miltenberger,
Woster, Rortvedt,
& Gaffeney
(1994)


-3 years,
Male,
Developmental
delays 4 years,
Male, Speech &
Cognitive
delays
-4 years,
Male,
Speech &
Language,
Chronic
misbehavior

-3 years,
Female, Speech
delays
-13 years, Male,
Down syndrome
-5 years,
Male,
Autism & Mild
MR
-2 years,
Male,
FAS
-5 years,
Male, Bipolar
disorder


Self-contained
classrooms


Participants'
homes


Researchers
conducted FAs and
administered
descriptive
assessment to
teacher and
paraeducators


-Researchers
administered MAS
and FAI to parents.
Researchers
completed A-B-C
direct observation
assessments. -
Researchers
directed parents in
FA conditions


Informal ranking
process of 4
possible
functions of
misbehavior,


MAS, FAI, ABC


N/A


-MAS consistent
with FAI for 2/5*
-FAI and A-B-C
consistent 5/5.
MAS and A-B-C
consistent 2/5


Based on
preliminary
assessments
each participant
placed in 1-3
analog
conditions.
Analysis
occurred in
small work
room or isolated
classroom


Each participant
was placed in
two analog
conditions either
escape or
attention**


Minimal
agreement
between
informal
assessment and
FA. Only one
paraeducator
for one
participant
agreed with the
FA results for
primary
functions

Consistent
agreement
between ABC,
FAI and FA.
Inconsistent
agreement for
MAS and FA.














Table 2-3. Continued
Studies Participants: Setting Who completed the What descriptive Agreement How was the Agreement
Age, Gender, FBA? assessments were between functional between FA
Disability used? descriptive analysis and descriptive
assessments completed? assessments
Sasso et al. (1992) -7.5 years, Self-contained Teachers ABC direct N/A Researcher Consistent
Female, Autism classrooms completed all observation conducted agreement
-13 years, aspects of FBA formal FA. Each between ABC
Male, Autism teacher results and
conducted 5 FAs
analog
conditions in
classroom
*For the other 3 participants, the second function on the MAS was the same as the primary function for the FAI.
**One participant was placed in an interrupted play versus non-interrupted play condition.









The team scored the accuracy of the statements on a four-point Likert scale and

then ranked the summary statements from least to most problematic for the student in the

classroom. This final step of scoring summary statements for accuracy was developed in

another study included in this review, Yarbrough and Carr (2000). Second, the first

author conducted the same interview process with each individual student. Third, the first

author and a paraprofessional conducted direct classroom observations using the

Functional Assessment Observation Form (FAOF). Using the results from the teacher

team interview as a guide, students were observed for 20-30 minutes, 1-3 times per week

for three weeks.This resulted in a total range of 5-15 hours of direct observation. Inter-

observer agreement was conducted for 19% of the sessions and averaged 80% agreement.

Participating teachers also completed a social validity scale to assess the three-step

process. There were four possible hypothetical functions (attention, escape, tangible and

self-stimulation) that the three assessment methods could identify. However, tangible

and self-stimulation were almost never identified in teacher and student summary

statements. Results of the study indicated agreement of a hypothesized function for all

three methods 9/14 (64%) of the time when either escape or attention was identified.

When there was not agreement across all three methods, the teacher team interview and

the direct observations still agreed four out of the other five times. Escape from a task or

activity was the most prevalent function identified.

There are four conclusions that can be drawn from this study. First, these

investigators achieved a moderate level of agreement across descriptive assessments.

Second, although there was agreement across the results of different assessment methods

(including a measure of confidence from participating teachers and students), the validity









of the results remains unconfirmed because no experimental procedure occurred. Thus,

the results are correlational in nature. Without an experimentally rigorous functional

analysis, the consistency across instruments does not necessarily indicate validity of the

function. In other words, all of the descriptive assessment may agree with each other that

the function is attention, but this may not actually be the function. This is especially true

if the same individual is completing all of the descriptive assessments. Third, the nature

of the participants limited the possible hypothetical functions that could be determined. It

is unlikely that in this setting junior high school aged students with disabilities such as

LD and EBD would engage in self-stimulation or have acquiring tangibles as the function

of their behavior. In other words, this sample of students was unlikely to engage in

stereotypy and also unlikely to argue over toys (although other tangibles could cause

challenging behavior). By limiting possible functions to attention and escape, this

increased the possibility for agreement across measures. Fourth, the findings of the

descriptive assessments were not kept separate in this study, the direct observations were

informed by the teacher team interviews "the behaviors listed on the observation form

came directly from those listed on the teacher interview form." (p.10). This makes the

level of agreement between the teacher interviews and the direct observations less

compelling.

In Newcomer and Lewis (2004), three elementary-aged students (one classified as

Other Health Impaired, two not classified with any specific disability) were first

evaluated using five descriptive assessment methods followed by a brief functional

analysis was conducted. The five assessment methods were a teacher interview, a student

interview, a teacher checklist, a scatter plot, and classroom observations. The functional









analysis procedures presented a limited number of conditions (attention v. no attention,

preferred v. non-preferred peer groups, easy task v. hard task). These conditions were

guided by the descriptive assessment results and while the article refers to all three

experimental procedures as functional analyses, the manipulations of peer group and task

difficulty is more aptly described as a structural analysis, because it is a manipulation of

antecedents rather than consequences.

The five descriptive assessments were presented in the same order for all

participants. First, the teachers were interviewed using an adapted version of the

Functional Assessment Interview (O'Neill et al., 1997). After a hypothesis regarding

function was formulated using the interview results, teachers were asked to complete the

checklist, the Problem Behavior Questionnaire (PBQ), and a scatter plot for a minimum

of five days. It should be noted that unlike the other assessment methods used in this

study, the scatter plot does not lead to a hypothetical function, it is a direct observational

tool that identifies patterns between occurrences of the behavior and times of day and/ or

activities (Touchette, et al., 1985). Targeted students were then interviewed using the

student version of the FAI. Finally, the student was observed using the Antecedent

Behavior Consequence (A-B-C) sheet. The observation periods were determined by

results from the other descriptive assessments. The amount of time and number of

observations for this descriptive assessment were not reported. These descriptive

assessment results were assimilated to identify a primary hypothetical function that was

tested through experimental analysis.

The experimental analysis was constructed using a single-case, alternating

treatment design. The analyses occurred in natural classroom settings during both









academic and low structure tasks and targeted individually identified challenging

behaviors. Each child was exposed to 10-minute analog probes featuring the

hypothesized function condition, according to the descriptive assessments and a control

or contra-indicated condition. There was a range of 13-22 sessions over 2- 7 days. The

results of the experimental analysis showed clear differentiation of rates of challenging

behavior across conditions for 2 of the 3 participants, supporting the summative results of

the descriptive assessments for these 2 participants. There was only slight differentiation

between the conditions for the third participant offering limited support to the result of

the descriptive assessment.

The results of the descriptive assessments were also compared with each other.

Newcomer and Lewis state that all of the descriptive assessments were consistent with

two exceptions. For one student, the teacher interview and the checklist indicated

attention was the primary function, but all of the other assessments including the

experimental analysis indicated that escape was the primary function. Similarly, for

another participant, the checklist identified attention as the primary function when all

other assessments supported escape as the primary function. It is notable that neither of

these 2 participants was the same individual that had limited support from the functional

analysis. Therefore, all 3 participants had some type of assessment result that was

inconsistent with other assessment results.

There are three conclusions that can be drawn regarding this research study. First,

although there is some evidence to support the reliability and validity of descriptive

assessments, in that they had fairly consistent results that were supported by an

experimental procedure, the function of behavior in this case does appear contextually









relative and somewhat fluid. Second, the use of descriptive assessments as a 'package'

with a specific order and one assessment informing the next one does not allow for

scrutiny of the individual assessments and does make the results vulnerable to an order

effect. Third, by limiting the functional analysis conditions to those suggested by the

descriptive assessments, no other functions were experimentally tested. This means that if

additional function of the behavior exists, but was not indicated by the majority of the

descriptive assessments, then it was not evaluated experimentally.

In Cunningham and O'Neill (2000), three preschool aged children with autism

were evaluated using four descriptive assessments and an experimental functional

analysis. The four descriptive assessments were: an informal question to the teacher and

classroom assistant (e.g. "Why do you think Dan bites?"), a teacher interview (the

Functional Assessment Interview), a behavior checklist (the Motivation Assessment

Scale), and descriptive observations (the Functional Assessment Observation Form). The

functional analysis was then conducted replicating the procedure outlined in Iwata,

Dorsey, Slifer, Bauman and Richman (1994/1982) presenting 5 analog conditions:

demand (escape), attention, tangible, alone and free play.

The procedure for this study consisted of conducting the indirect descriptive

assessments (the informal question, the interview, and the behavior checklist) with both

the teacher and the classroom assistant after class hours. These were conducted separately

for each teacher/classroom assistant dyad to determine if the two adults present in the

classroom had similar perceptions of the challenging behavior and its antecedents and

consequences. The MAS scores were then combined to give an average score for the

primary function. The primary classroom teacher then completed the direct observations









for two hours a day for 4-5 days, resulting in a total of 10-12 hours of direct observation

for each participating student. Inter-observer agreement was conducted for 20% of these

sessions and ranged from 80%-100% agreement. The design for the experimental

functional analysis was a single subject, multi-element design with 15-minute sessions

being conducted in separate rooms that were made available to the researchers, e.g. a

multi-purpose room adjacent to the gymnasium.

The results for the various different descriptive assessments were presented as a

rank ordering of hypothetical functions of the target behavior. For the descriptive

assessments, there was perfect agreement for the informal question and the interview for

all dyads of teachers and assistants. The other assessments also had fairly good agreement

in which all the assessments varied only somewhat, for example, the MAS listed escape

as the primary function and tangible as secondary and the FAI listed tangible as primary

and escape as secondary. For the functional analysis, 2 of the 3 participants had little

differentiation between tangible and escape. However, this mixed result actually confirms

the descriptive results that had similar mixed primary functions across the different

assessments.

There are three conclusions regarding this research study. First, soliciting

information from both the teacher and the classroom assistant determined that the two

adults present in the room had similar perceptions regarding behavioral functions (i.e.,

inter-rater reliability). Second, the use of ranking the functions for each student and each

assessment allowed for students whose challenging behavior may have dual functions but

could have complex results. Subsequently complex interventions that may be difficult to

implement in classrooms are the end product. Third, one aspect of the functional analysis









that may make its results less compelling was the low rate of challenging behavior

demonstrated across all analog conditions. But because the greatest rate for the students,

means of 10%, 10% and 9%, was consistent with the descriptive assessment results, the

functional analysis appears accurate.

In Yarbrough and Carr (2000) 3 participants between the ages of 11 and 18 years

old were evaluated. Two participants were classified autistic and the third participant was

diagnosed with DeLange syndrome. There was only one descriptive assessment used. An

adapted version of a Functional Assessment Interview was used with two professionals

who had had up to two weeks of contact with the participating student. This descriptive

assessment was followed by a brief functional analysis. The brief functional analysis

manipulated both antecedents and consequences that had been identified by the

descriptive assessment results.

The FAI was adapted in three ways. First, after information had been gathered

using the standard FAI questions, summary statements were generated for each

participant that outlined setting events, antecedents and consequences. Each professional

who was interviewed then verified each statement for accuracy and rated each summary

statement on a 7-point Likert scale for likelihood of the situation occurring. Next, the

second professional would then evaluate and agree or disagree with these hypothesized

summary statements. Lastly, these same steps would be taken with the roles of the two

professionals reversed. If statements were generated that the second professional could

not confirm witnessing, then a third professional, such as another classroom assistant,

would be asked to confirm the statement and its likelihood rating. Following this









descriptive assessment process, a brief functional analysis was conducted using the

results.

In the functional analysis, a reversal design was used in which the different

hypothesized situations were compared with control conditions that regulated the

presence of the hypothesized contributory variables. The 3 participants were placed in

seven situations while one or two different hypothetical functions were tested in each

condition. For example, the first participant was placed in a situation in which he had to

clean a table. The hypothetical function was that his challenging behavior of self-injury

was maintained by escaping difficult tasks. A control condition consisted of assigning the

participant to a task he found easier, object identification. The frequency of the

challenging behavior in these two conditions was then compared. Each participant was

placed in two or three high and low-likelihood situations with one or two hypothetical

functions of the challenging behavior being tested for each participant.

When the information gathered through the descriptive assessments was tested

using an experimental functional analysis, a clear pattern emerged. In those situations that

had a high-likelihood of eliciting problem behaviors, the informants accurately identified

the surrounding environmental variables. For the situations that had a low likelihood of

eliciting challenging behaviors, the functional analysis did not validate the informants'

hypotheses of the antecedents and the functions of the challenging behaviors. The authors

offer four interrelated factors that may explain this pattern: (1) infrequent opportunities to

observe the low-likelihood situations; (2) stimulus complexity; (3) perceived strength of

reinforcement; (4) the role of setting events. These factors should be interpreted with

caution, as they are not experimentally validated. In terms of evaluating the efficacy and









efficiency of the descriptive assessment method (i.e., the FAI adaptation), it can be

concluded that this method is effective for identifying the function for certain frequent

behavior problems, but limited in detecting more subtle variables in low-likelihood

situations. Unfortunately, the authors did not conduct a treatment acceptability measure,

so it cannot be determined if this benefit is worth the time and energy investment for the

participating professionals.

In Ellingson, et al., (2000), 3 participants, aged 19, 18 and 12, were evaluated using

two different descriptive assessments. No experimental functional analysis procedures

were used. The participants were all classified as either severely or profoundly mentally

retarded and one participant was also diagnosed with Angelman's syndrome. All of the

assessments occurred in special education classrooms in public schools. A second phase

of this study investigated function-based interventions, but this phase of the research is

outside the scope of this literature review.

The two descriptive assessments were each used in two different ways. The first

descriptive assessment was an interview that was presented as both an interview with the

first author asking follow-up questions and as a questionnaire that the teacher completed

independently. The second descriptive assessment was a direct assessment that used the

Antecedent-Behavior-Consequence (ABC) worksheet. These worksheets were

simultaneously completed by both teachers and experts trained in Applied Behavior

Analysis. The assessments were always presented in the same order: (1) the teacher

completed the questionnaire independently, (2) the teacher was interviewed using the

same questions as the questionnaire by the first author except follow-up questions were

also asked, (3) using the information provided by the interviews and questionnaires an









ABC checklist was formulated and direct observations with the ABC instrument were

conducted at the same time by both the classroom teacher and a trained expert. Two other

trained behavior analysts conducted reliability checks on the hypothetical function

determined for the interview and the questionnaire.

The results of the descriptive assessments were determined by the authors'

evaluations of the interview/questionnaire and the ABC checklist results. All of the

hypothetical functions were determined to be either attention or escape. For one

participant, all of the different methods of assessment, including the second rater who

supplied reliability data, agreed that the function of the challenging behavior was

attention. For the second participant there was agreement that the function of the behavior

was attention for all of the assessments except for the second rater of the questionnaire.

For the third participant there was agreement from the indirect assessments; the interview

and questionnaire agreed that the function of the problem behavior was both attention and

escape. All of the direct observation results determined that the function of the problem

behavior was attention.

There are three conclusions regarding this research. First, the agreement between

the questionnaire, the interview and the direct observations are encouraging but in

consideration for the amount of influence the questionnaire and interview had over

structuring the direct observation, the agreement is less compelling. The construction of

an ABC checklist using indirect assessments requires further testing, especially against a

less structured ABC observation sheet and functional analysis. Second, the agreement

between the questionnaire and the interview is a step toward making the process more

efficient, (i.e., it only requires one person rather than two people); however, this result









may be idiosyncratic to the skills of the interviewer. Some interviewers may elicit

valuable information through follow-up questions that a questionnaire completed

independently could never draw out. Third, it is unclear whether the possible functions

were in some way limited to the choices of attention or escape, or if this was simply a

characteristic of these three specific participants.

In Calloway and Simpson (1998), three preschool aged children were evaluated

using two descriptive assessments, an experimental analysis and an informal question

posed to teachers and para-educators. The result of the descriptive assessment and the

experimental analysis were then compared with the responses to the informal question.

The 3 participants were male, ages 4, 4 and 3 and three were in an inclusive community

early childhood special education classroom. Two participants were in the same

classroom and one was in a different classroom; all assessments took place in the

classroom setting. Two children had speech and language and cognitive delays and one

participant was diagnosed with developmental delays and suspected autism. All three had

been referred to this study for repeated demonstration of various problem behaviors.

The procedure for this study occurred in three phases. The first phase consisted of

descriptive classroom observations and completion of the MAS. The researchers

interpreted this information and generated a hypothetical function. Manipulating through

analog conditions tested this hypothesized function. A reversal design was used with a

contra-indicated activity to compare behavioral frequencies. These reversals took place

every two or five minutes for either one 10 or 20 minute session. The second phase

involved implementing an intervention directly related to this function. This was

implemented not only to reduce the problem behaviors for the individual child, but also









to provide further validation of the function of the behavior as determined by the

descriptive assessment and functional analysis.

In the third phase of the study, three classroom professionals, either the teacher

and/or the classroom para-educators, were asked to identify a hypothetical function of the

problem behavior. They had no knowledge of the results of the formal analyses. The

classroom professionals were asked to identify the behavioral function using a ranking

system of 1-4 for these choices: (a) avoidance, (b) attention, (c) self-expression, and (d)

power/control.

According to the formal analyses, one student's function was confirmed as

attention, the second student's function was escape and the third student's results

indicated that the problem behavior was maintained by multiple functions: attention,

power/control and escape. The results of the ranking system indicated that the classroom

professional agreed with each other six out of nine opportunities (66%) in identifying the

primary function. However when comparing the formal and informal analysis results, the

classroom professionals only agreed with the formal analysis one out of nine times

(11%).

There are three salient aspects about this research study. First, the use of 'self-

expression' and 'power/control' as possible functions is divergent from most current

literature regarding functions of problem behavior. Self-expression was the function

chosen five out of nine times by the teachers and paraprofessionals, and it was not

confirmed by any of the formal analyses. Further, the functions of self-expression and

power/control are not even options on the MAS; therefore, a direct comparison between

the interview and the MAS for some participants was not possible. Second, agreement









between the analog conditions and the descriptive assessments is difficult to determine

since the analogs directly applied the results of the descriptive assessments, making it

impossible to separately evaluate them. Third, in regard to the results of the informal

analysis, this point to a larger consideration, namely that consistency does not equal

accuracy. The professionals were consistent with each other 66% of the time, but they

were only accurate in the sense that they agreed with the functional analysis 11% of the

time.

In Arndorfer, Miltenberger, Woster, Rorvedt and Gaffeney (1994), five children

ages 2 to 13 years old were evaluated using three descriptive assessments and a brief

experimental analysis. Unlike the other studies included in this section of the literature

review, all of the assessments occurred in the participants' homes. Parents served as the

source of information for the indirect assessments and also conducted direct assessments.

The 5 participants had been diagnosed with different disabilities including: (1) Down

syndrome, (2) fetal alcohol syndrome, (3) autism and mild mental retardation, (4)

developmental speech and cognitive delays, and (5) bipolar disorder. All of the

participants were referred to this study due to repeated demonstrations of problem

behaviors.

There were two phases of this research study. The first phase consisted of the

parents completing the Motivation Assessment Scale (MAS) two times with one to two

week delays between administrations. Repeated administrations were done to evaluate

reliability over time. Before the MAS was scored, the first author conducted the FAI with

the parents and the first and second authors independently reviewed the responses and

generated hypothetical functions. The MAS was then scored to also determine a









hypothetical function. Then, a third researcher and the parents of the participant, who

were unaware of the indirect assessment results, conducted three to five 30-60 minute

direct observations using the FAOF. The parents received a brief training session on

conducting a direct observation using the FAOF. A scatter-plot assessment determined

specific times target behaviors were likely to occur and this guided the times the direct

observations were conducted. Inter-observer agreement was conducted for 33% of all

direct descriptive assessments. Using the results of direct observations, hypothetical

functions of the problem behavior were formulated.

The second phase consisted of experimentally manipulating antecedents and

consequences that had been identified through the descriptive assessments. A reversal

design consisting of high-likelihood antecedents and consequences and low-likelihood

antecedents and consequences was conducted. For example, a participant whose

descriptive assessments indicated that attention was the primary function of her behavior

had a session where her mother looked at, talked to, or touched her at least twice every

minute and all problem behaviors were ignored. In the next condition, the mother looked

at, talked to, or touched the participant's siblings at least twice every minute but gave

contingent attention based on the participant exhibiting problem behaviors. These

conditions were presented in 15-30 minute sessions on the same day for three of the

children and over two to three days for the other two children. Only one session of each

phase was conducted for each child in an effort to make the experimental analysis as

brief, and socially valid, as possible.

The result of the descriptive agreement indicated agreement 16 out of 20

opportunities (four descriptive assessment results for 5 participants). For three out of the









four disagreements, the MAS was the instrument with divergent results. Additionally,

agreement between the two administrations of the MAS indicated agreement for

individual questions ranging from 56% to 75% (M=64%). The experimental analysis

supported the results of the interview and the direct observations for all five of the

participants (100%). There was differentiation in the frequency of the problem behaviors

throughout the reversal of the conditions.

There are three conclusions regarding this research study. First, this study bolsters

the use of the FAI and the direct observations to generate accurate hypothetical functions

of behavior, but draws the validity of the results of the MAS into question. Second, while

the experimental analysis does offer support for the descriptive assessments, because it

only tested one function, it is not possible to know if other functions of the behavior exist

without testing multiple functions. Third, by only conducting sessions of each condition

on one day for 3 of the 5 participants, there are contextual factors, and setting events that

may not be accounted for and may impact the function of the problem behavior. It seems

possible that this study may have sacrificed a degree of accuracy in an effort to increase

efficiency and social validity of the experimental process.

In Sasso et al. (1992), 2 participants, ages 7 and 13 years old, were evaluated using

a conventional functional analysis, a direct observation descriptive assessment and a

teacher conducted functional analysis. Both participants were classified as autistic and the

descriptive assessment and teacher conducted functional analysis occurred in self-

contained special education classrooms. The conventional functional analysis occurred

outside the classroom setting. The direct observation assessment was the ABC worksheet.

However, it was completed during specific activities that are similar to the conditions of









the functional analysis conditions (i.e., presentation of a difficult task). The functional

analyses consisted of a multi-element design based on Iwata et al. (1982/1994) featuring

four conditions: alone, escape, attention, and free play and also including a tangible

condition.

There were three phases of this study. First, researchers conducted a conventional

functional analysis. Second, the first teacher was trained to complete an ABC

observation. Then, the teacher conducted four 15-minute observations of the student. The

environment was varied for the observations and included solitary play, a high demand,

low teacher attention condition and a low demand and high teacher attention condition.

After the direct descriptive observations were completed, a brief functional analysis

within the classroom was conducted. The classroom teacher modified the environment to

test the five analog conditions. The same procedure occurred for the second participant-

the only difference was that the first teacher, rather then the researcher, taught the second

teacher to conduct the descriptive assessment and the classroom based functional

analysis. The results of this study indicated that in all three methods of assessment,

escape and tangible were the functions of the problem behavior for one participant and

escape was the function of the problem behavior for the other participant. Thus, there was

100% agreement for both participants.

There are three conclusions that can be drawn from this research study. First, this

research article demonstrates flexibility in the functional analysis process, in that the

same procedures can be applied across different settings and produce similar results.

Second, this study suggests that functional analysis procedures may have broader

applicability for classroom teachers. This runs counter to the concept that functional









analysis procedures are too complex to have broad-based appeal. Third, this idea that

teachers can effectively train teachers suggests of the possibility of a much broader

training model.

There are two conclusions that can be drawn from a summative evaluation of these

eight studies that directly compared descriptive assessments with each other and/or some

type of functional analysis. First, no descriptive assessment method has emerged as the

preeminent method to accurately determine a hypothetical function. It appears that for

indirect assessments, teacher interviews, both formal and informal, play an important role

in identifying potential functions in seven out of eight of these studies. Their overall

accuracy is further suggestion of the conclusion drawn earlier that research should

continue to focus on the knowledge teachers already have about their students' behaviors

and effectively and efficiently hone methodologies to access this knowledge. For direct

descriptive assessments, ABC assessment sheets completed by both teachers and

researchers do reveal accurate antecedents and functions. However, this conclusion is

tainted by the second conclusion about the overall state of this research. Namely, that all

of the methods, descriptive and experimental, were not conducted independently of each

other with a high degree of control.

This second conclusion, that the assimilation of results for all of the methodologies

muddled the final results, is pervasive throughout all eight of the studies. This is true

both for the descriptive assessments and for the link between the descriptive assessments

and the functional analyses. The end result is that no conclusions can be drawn about

what single descriptive methodology is the most accurate. This occurred for a variety of

reasons. First, some studies used a ranking system between raters that allowed for









multiple functions to be enumerated within the results. The second and most influential

reason is that the descriptive assessments were used to guide the functional analysis. By

not maintaining separation of the descriptive assessments from the functional analysis,

the functional analysis can only verify or refute the hypothetical function that the

descriptive assessments reveal. Conducted independently, functional analysis can

establish the function of the challenging behavior (Iwata et al., 1982/1994). In future

research, these results could then be taken back and compared with the independent

descriptive assessments to determine which ones accurately arrived at the function.

In summary, the overall lack of comparative studies that evaluated descriptive

assessments suggests that descriptive methods have not been validated by research. If

descriptive assessments, such as rating scales, interviews, and hypothesis development

using antecedent, behavior, consequence (ABC) analysis (Bijou, et al., 1968) indicate an

inability to accurately determine the function of a young child's disruptive behavior, then

their use is ill-advised. However, if descriptive methods adequately arrive at the same

function as experimental analysis, then the use of these simpler, less time-consuming

methods has been supported. Regardless of the outcome of the project, it is incumbent

upon researchers to prepare school and early childhood center personnel with the best,

experimentally validated procedures.

Comprehensive Summary of Research Literature

"Whether FBA can be effectively and efficiently accomplished outside a research

setting remains to be seen"(Reid & Nelson, 2002, p.23). As Reid and Nelson suggest, the

use of FBA as a durable, highly effective strategy in widespread applications in applied

settings remains to be seen. A review of 32 studies that used FBA, predominately with

young children with high incidence disabilities in natural settings, indicates three









conclusions that can be drawn from the current state of the research. First, in regard to the

participants and settings, the use of functional behavior assessment is still clearly

evolving in terms of its use with large numbers of young students with mild disabilities in

classroom settings. Second, interventions generated by FBA, while effective in the

studies in this sample, remain idiosyncratic and, at times, the relationship between the

derived function and the function-based intervention could be described as tenuous at

best. The third and the most important factor is that in terms of the FBA process the

individual components that would make up a universal FBA protocol have not been

established as valid and reliable for the widespread application of functional behavior

assessment.

The first conclusion in regard to the research in this literature review concerns the

participants and settings of FBA. Two considerations regarding the practical limitations

of all of these studies are the types of disabilities that are being addressed and their use of

large numbers of students. A third ancillary consideration is the type of behaviors that are

being addressed within the FBA literature. What may be attributed to IDEA (1997) was

this use of FBAs for children with disabilities other than developmental disabilities. In

1999, Nelson, Roberts, Mathur and Rutherford estimated that 80%-90% of functional

assessment research had been conducted with children with developmental disabilities.

The trend, however, to apply it to students with other disabilities or no disability at all is

proliferating according to this sample of studies. This is encouraging and indicates that

FBA is a highly flexible process with multiple levels of applications.

Combined with the problem of determining the ability to generate large numbers of

FBAs, the problem of researcher influence is an important consideration. The question of









'how well do these studies indicate what will occur in generalized application after the

study?' remains. This threat to external validity points to an important area of future

research. There is a need for further research in this area that uses longitudinal

methodology providing initial teacher training on the use of descriptive assessments, a

series of temporally determined probes to evaluate implementation fidelity, and a final

measure of effectiveness. Studies using a design such as this will allow researchers to

minimize the impact that they are having on teacher implementation of the FBA process.

This is a problematic shortcoming of the research that has not been answered to date.

A third point regarding this first conclusion is in regard to what types of behaviors

research has addressed with FBA methodology. Studies clearly addressed a

predominance of externalizing behaviors. While this is not surprising, as teachers are

more likely to identify externalizing behaviors as problematic, it does point to another

shortcoming in literature thus far. This is the question 'can FBA procedures effectively

address problematic internalizing behaviors?' Internalizing behaviors, social withdrawal,

isolation, and flat affect can be as problematic for students as externalizing behaviors

(Chandler & Dahlquist, 2002). However, current research has not addressed whether

these behaviors can be effectively addressed through functional behavior assessment.

The second conclusion is in regard to function-based interventions. While some

interventions led to highly specific interventions directly related to the function of the

challenging behavior, other interventions were only indirectly related to the function or

were not related at all. Researchers and teachers implemented a number of effective

interventions, manipulating both antecedents and consequences to reduce challenging

behaviors; however, one practical and methodological limitation is evident from this









sample of literature. In regard to methodology, even studies that were directly linked to

the intervention received far less emphasis than the process that arrived at the function of

the behavior. Empirical research has not produced a 'menu' of validated interventions

that directly match functions. However, the overall lack of emphasis on interventions is

surprising as the ultimate goal of FBAs is the formulation of effective intervention and

essentially FBA methodology will be judged on its effectiveness. In regard to practical

limitations, a limitation that this sample of studies did not address is the question of

teachers being able to sustain prescribed, function-based interventions for a number of

students over time. This will be an important area of future research as widespread

application of function-based interventions continues to broaden into different types of

classrooms and different populations of students.

The third and most important conclusion regarding this literature is the lack of

validity and reliability studies that directly evaluate the components that make up the

FBA process. The reason this conclusion is the most important is that before widespread

application for many children in applied settings with a variety of behaviors can occur,

the process must be honed so that it is not only reliable and valid, but also accessible to

minimally trained school personnel and highly efficient. Similarly, although effective

function-based interventions are paramount to the process, it will be impossible to derive

those without reliable and valid components. Unfortunately, even in the third section of

this literature review, a section that was comprised of studies that directly evaluated

descriptive assessments, useful evaluation of the components did not occur. There was

still a lack of separation of the findings between the individual descriptive assessments

and the descriptive assessments and the functional analysis that was often used to validate









the hypothetical functions. As a result in these research studies, it was very difficult to

determine what components were most effective.

This leads to an important area of research namely, conducting a series of indirect

and direct descriptive assessments and comparing the outcomes of these assessments to

the outcomes of a functional analysis methodology. However, to rectify limitations in the

existing research, independence of the descriptive assessment results and the functional

analysis results is important to determine if the descriptive instruments are able to obtain

valid functions independent of the functional analyses. Additionally, the results of the

various direct and indirect descriptive assessments would need to be kept separate and

counterbalanced across participants to maintain control for carry over and order effects.

These assessments should occur in natural settings with consideration of total time of

procedures and necessary levels of training with the goal of maximizing efficiency

without sacrificing validity. While this literature review has indicated a number of

problems with the widespread application of FBA in natural settings, it is vital that the

methodologies be evaluated before issues of implementation, teacher training, and

effective interventions are addressed. The results of this research are an empirically

validated step towards establishing a universal protocol that can be applied consistently

both in continuing FBA research and in applied settings.

Research Questions

The following questions were addressed in this study:

1. How do the functions of target behaviors identified through various descriptive
functional assessments compare to each other?

2. How do the functions of target behaviors identified through various descriptive
functional assessments compare with functions identified through a functional
analysis for young children's challenging behavior?














CHAPTER 3
METHODS

This chapter outlines the methodology used to conduct this study. The goals of this

study were to compare the functions determined by three descriptive assessments with

each other and with the functions determined by functional analyses (FA) for young

children who demonstrate challenging behaviors. Teachers' perceptions of the process

and its results were also evaluated. This chapter presents the following information: (a)

participants, (b) settings, (c) materials, (d) measurement procedures including dependent

measures, the independent variables, data collection and type of data recording, and

interobserver agreement. This chapter then outlines the experimental procedures

including recruitment and training, pre-experimental assessments, descriptive

assessments, and functional analysis. Finally, the methods of measuring social validity,

procedural integrity, and the design and data analysis are presented.

Participants

Four teacher-child dyads participated in this study. In accordance with the policies

set forth by the University of Florida Institutional Review Board, the principal

investigator obtained informed consent from the participating teachers and caregivers of

participating children (see Appendix A for approved IRB forms and participant informed

consent forms). All participating teachers and the child participants' parents received a

$25 dollar gift card. Child participants were four children who met the following criteria.

* Children were between the ages of 3 and 6 years old.









* They were currently enrolled in a Day Care Center, Pre-kindergarten, Kindergarten,
or First Grade class (general education or special education) in North Central
Florida.

* They had been identified through teacher nominations, in that teachers have
reported that these children are having challenging behaviors that are interfering
with their learning and this behavior warrants an FBA.

* They had a Total Problem Score of no less then 60 on the Child Behavior Checklist
(CBCL-TRF 1.5-5 or 6-18 years old) (Achenbach, 1992).

* They had consistent attendance at school, defined as 80% attendance or 4 out of 5
days a week, documented through school records.

* Teacher reports indicated average cognitive functioning.

Teacher participants were four individuals who met the following criteria:

* They taught children ages 3-6 years old.
* They had at least one child in the classroom who met the criteria above.
* They held at least an Associate of Arts degree,
* They served as the child participant's primary teacher for at least one month.

The flow chart of the teacher and child participant recruiting process is outlined in

Appendix B. The following is a brief description of all four participant dyads.

Ramon. Ramon is a 4-year old Hispanic male. His primary teacher, Rose, is a

Hispanic female. She has thirty years of experience in early childhood care settings; she

has had 7 months of contact with the participating child, and holds a Bachelor of Arts

degree. His teacher reported average cognitive functioning. The teacher report form of

the Child Behavior Checklist (CBCL-TRF 1.5-5) (Achenbach, 1992) indicated that

Ramon has a total score on the Total Problem scale of 110 (>97 percentile) with sub-

scores in the clinically significant range for emotionally reactive behavior,

anxious/depressed, withdrawn, attention problems and aggressive behavior. Ramon's

challenging behaviors were disruptive vocalizations defined as raising his voice above









the classroom volume level, asking more than three questions in a 10-second interval or

using profanity or "bathroom talk."

Anthony. Anthony is a 5-year old African-American male. His teacher reported

that Anthony has average cognitive functioning. His primary teacher, Terry, is a white

female. She has 4 years of experience in early childhood care settings, has known the

participating child for 6 months and has completed her Associate of Arts degree. The

teacher report form of the Child Behavior Checklist (CBCL-TRF 1.5-5) (Achenbach,

1992) indicated that Anthony has a total score on the Total Problem Scale of 88 (>97

percentile) with sub-scores in the clinically significant range for emotionally reactive

behavior, sleep problems, and aggressive behavior. Anthony's challenging behaviors

were crying, yelling "Mom" or "Mama", rolling around on the ground, raising his voice

above the classroom volume level and leaving his assigned area (defined as being more

than 6 feet away from his assigned area).

Jimmy. Jimmy is a 5-year old African-American male. His teacher reported

average cognitive functioning. His primary teacher, Amy, is a white female. She has 25

years of experience in an early childhood care setting, has known the participant for 7

months and holds a Bachelor of Arts degree. The teacher report form of the Child

Behavior Checklist (CBCL-TRF 1.5-5 years old) (Achenbach, 1992) indicated that

Jimmy has a total score on the Total Problem Scale of 61 (87 percentile) with sub-scores

in the borderline range for aggressive behavior. Jimmy's challenging behaviors were

disruption defined as verbal defiance screaming, teasing, saying 'No', crying or whining.

Greg. Greg is a 6-year old African-American male. Teacher reports indicated

average cognitive functioning. His primary teacher, Alison, is a Hispanic female. She has









1 year of experience in the classroom; she has had 7 months of contact with the

participating child and has a Bachelor of Arts degree. The teacher report form of the

Child Behavior Checklist (CBCL-TRF 6-18) (Achenbach, 1992) indicated that Greg has

a total score on the Total Problem Scale of 95 (>98 percentile) with sub-scores in the

borderline range for social problems and rule-breaking behavior and in the clinical range

for thought problems and aggressive behavior. Greg's challenging behaviors were

disruptive noises defined as tapping his pencil, whistling and raising his voice above the

classroom volume level. Child and teacher information are summarized in Tables 3-1 and

3-2. All CBCL sub-scores are reported in Appendix C.

Table 3-1. Child Participant Information
Child Participant Age Race Gender CBCL-TRF Total Score
Ramon 4.9 Hispanic Male 110 (97th percentile)

Anthony 5.0 African-American Male 88 (>97th percentile)

Jimmy 5.5 African-American Male 61 (87th percentile)

Greg 6.4 African-American Male 95 (>98th percentile)


Table 3-2. Teacher Participant Information
Teacher Child Years of
Participant participant Experience

Rose Ramon 30 years

Terry Anthony 4 years

Amy Jimmy 25 years

Alison Greg 1 year


Race


Hispanic

White

White

Hispanic


Months of
Contact with
Child
7 months

6 months

7 months

7 months


Current Level
of Education

Bachelor of
Arts
Associate of
Arts
Bachelor of
Arts
Bachelor of
Arts









Settings

The settings for all of the assessment methods were the primary classrooms; a

primary classroom setting was defined as the room in which the child participant spends

the majority of his school day. For the first 3 participants listed (Ramon, Jimmy and

Anthony), the setting was the 4 and 5-year-old classroom of a university-sponsored early

childhood care center. The fourth participant's (Greg) primary setting was at a public

elementary school multi-grade, self-contained, special education classroom. The pre-

experimental assessments and teacher training occurred in the primary classroom with the

teacher during non-student contact times, (i.e., planning periods, nap times or after

dismissal). The indirect descriptive methods (interview and checklists were also

conducted in the primary classroom of the individual child during non-student contact

times, (i.e., planning periods, or naps).

The direct descriptive observations occurred during different classroom activities

with the majority of the observations in the child participant's primary classroom setting.

For the three participants in the early childhood center, the activities were primarily

Centers Time and Free Play. For the student in Kindergarten participant (Greg), the

activity was mainly independent work time. However, observations also occurred on the

playground for all 4 participants and in the cafeteria for Greg. The functional analyses

also occurred in the primary classroom setting. For the functional analysis, an area within

the primary classroom was designated where other children were not located, e.g. an area

in the corner of the room. Table 3-3 lists summaries of the phases, the assessments and

the settings in which they took place.




Full Text

PAGE 1

COMPARATIVE ANALYSIS OF FUNCTIONAL ASSESSMENT METHODOLOGIES AND FUNCTIONAL ANALYSIS FOR YOUNG CHILDREN WITH OR AT RISK FOR EMOTI ONAL OR BEHAVIORAL DISORDERS By PETER JAY ALTER JR. A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Peter Jay Alter Jr.

PAGE 3

With love to Erin and Betsy.

PAGE 4

iv ACKNOWLEDGMENTS I am extremely grateful to all for the he lp and support that I received throughout this program. It has been fulfilling, challenging and enriching. Most of all, I need to thank my wife, Erin. I can truly say that without he r I would not have gotte n to experience this, and I am eternally grateful.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT....................................................................................................................... ..x CHAPTER 1 INTRODUCTION........................................................................................................1 Overview....................................................................................................................... 1 Statement of the Problem..............................................................................................8 Significance of the Study..............................................................................................8 Purpose of the Study.....................................................................................................9 2 LITERATURE REVIEW...........................................................................................10 Prevalence and Impact of Young Child ren with Challenging Behaviors...................10 Defining Functional Behavior Assessment................................................................12 Theoretical Framework and Histor ical Perspective Supporting FBA........................17 Review of the Current Literature................................................................................21 Method to Select Reviewed Studies...........................................................................22 Results of the L iterature Review................................................................................26 Descriptive Assessment Methodologies.....................................................................27 Participant and Setting Characteristics................................................................28 Implementer Characteristics................................................................................33 Components of the Descriptive Approach..........................................................33 Functions Identified.............................................................................................37 Interventions Implemented..................................................................................39 Summary..............................................................................................................41 Descriptive Assessments and Functional a nd Structural Analyses within FBAs.......42 Participant and Setting Characteristics................................................................45 Implementer Characteristics................................................................................47 Components of the Descriptive Approach and Analog Conditions....................48 Functions Identified.............................................................................................50 Interventions Implemented..................................................................................52

PAGE 6

vi Summary..............................................................................................................54 Direct Evaluation of De scriptive Methodologies.......................................................56 Comprehensive Summary of Research Literature......................................................80 Research Questions.....................................................................................................84 3 METHODS.................................................................................................................85 Participants.................................................................................................................85 Settings....................................................................................................................... 89 Materials.....................................................................................................................9 0 Measurement Procedures............................................................................................92 Dependent Measures...........................................................................................92 Independent Variables.........................................................................................92 Data Collection Procedures.................................................................................95 Interobserver Agreement.....................................................................................97 Experimental Procedures............................................................................................99 Recruitment and Training....................................................................................99 Phase I--Pre-experimental Assessments............................................................100 Phase II--Descriptive Assessments....................................................................101 Phase III--Functional Analysis..........................................................................101 Social Validity..........................................................................................................105 Procedural Integrity..................................................................................................105 Design and Data Analysis.........................................................................................106 Design................................................................................................................106 Data Analysis.....................................................................................................106 4 RESULTS.................................................................................................................110 Individual Child Participant Results.........................................................................111 Ramon................................................................................................................112 Anthony.............................................................................................................113 Jimmy................................................................................................................114 Greg...................................................................................................................116 Comparison of Outcomes of Descriptive Assessments............................................118 Social Validity Questionna ire and Interview Results...............................................121 Social Validity Questionnaire............................................................................121 Social Validity Interview...................................................................................123 Summary...................................................................................................................124 5 DISCUSSION...........................................................................................................126 Limitations................................................................................................................128 Interpretations of Results..........................................................................................135 Comparison of Descriptive Assessments..........................................................135 Comparison of Descriptive Assessm ents with Functional Analyses................138 Social Validity...................................................................................................142 Future Research........................................................................................................143

PAGE 7

vii Implications for Practice...........................................................................................147 Summary...................................................................................................................150 APPENDIX A APPROVED IRB FORMS AND PARTICIPANT INFORMED CONSENT FORMS.....................................................................................................................152 B FLOW CHART OF TEACHER AND CHILD PARTICIPANT RECRUITING PROCESS.................................................................................................................162 C STUDENT PARTICIPANT CHILD BEHAVIOR CHECKLIST (CBCL) SCORES...................................................................................................................163 D INSTRUMENTS......................................................................................................164 E SOCIAL VALIDITY MEASURES.........................................................................165 F TEACHER TRAINING OUTLINE.........................................................................169 G FUNCTIONAL ANALYSIS PROCEDURES.........................................................173 H FINAL SCORING SUMMARY CHART................................................................178 LIST OF REFERENCES.................................................................................................179 BIOGRAPHICAL SKETCH...........................................................................................189

PAGE 8

viii LIST OF TABLES Table page 2-1 Descriptive Methods for Young Children................................................................29 2-2 Descriptive Methods and Functi onal Analysis for Young Children........................44 2-3 Direct Evaluations of Descriptive Assessments.......................................................59 3-1 Child Participant Information...................................................................................88 3-2 Teacher Participant Information...............................................................................88 3-3 Summary of Assessments and Their Settings For Each Phase................................90 3-4 Summary of Child Particip antsÂ’ Preferred Items, Neutral Activities and Difficult Tasks.........................................................................................................................9 1 4-1 Summary of the Functions of Ch allenging Behaviors Across Assessment Methods..................................................................................................................112 4-2 Summary of Agreements Between Descriptive Assessments Compared With Each Other and with a Functional Analysis...........................................................119 4-3 Results of the Social Validity Questionnaire.........................................................121

PAGE 9

ix LIST OF FIGURES Figure page 4-1 Challenging Behaviors per Minute During Functional Analysis for Ramon.........113 4-2 Challenging Behaviors per Minute During Functional Analysis for Anthony......114 4-3 Challenging Behaviors per Minute During Functional Analysis for Jimmy.........116 4-4 Challenging Behaviors per Minute During Functional Analysis for Greg............117

PAGE 10

x Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARATIVE ANALYSIS OF FUNCTIONAL ASSESSMENT METHODOLOGIES AND FUNCTIONAL ANALYSIS FOR YOUNG CHILDREN WITH OR AT RISK FOR EMOTI ONAL OR BEHAVIORAL DISORDERS By Peter Jay Alter Jr. August 2006 Chair: Vivian I. Correa Cochair: Maureen A. Conroy Major Department: Special Education The prevalence of functional behavior asse ssments (FBAs) in schools continues to increase as a result of federal mandates a nd widespread school policy reform. As the use of FBAs in schools increases, it is necessary to improve their current condition by evaluating the concurrent and construct valid ity of the process components. Failure to establish FBA as a process with valid outcomes will adversely impact students who demonstrate challenging behaviors b ecause a method for designing effective interventions will either be unus ed or the interventions that result from the process will be ineffective. There are a large number of de scriptive methodologies that can be used within the FBA process. However, the out comes of many of these methodologies have not been compared with each other or vali dated through functional analysis. Thus, there is a continuing need to empirically co mpare and validate the outcomes of these descriptive assessment methods The two purposes of this investigation were: (1) to

PAGE 11

xi compare the results of three commercially produced descriptive assessments with each other to determine concurrent validity, a nd (2) to compare the descriptive assessment results with the results of a functional analysis to determine construct validity. Descriptive assessments including one checkli st, an interview, and a direct observation were completed for four young children (ages 46) with or at risk for emotional and behavioral disorders. Teachers and researcher s then completed functional analyses within the classroom setting of the same participan ts. Results of the assessments were kept independent from each other until completion. Teachers also completed a social validity questionnaire and interview. This study indi cated that descriptive assessment results had low consistency with each other, 8/24 (33%) agreements between assessments, and that the outcomes of two of the three descrip tive assessments (i.e., the interview and checklist) also had low agreement with the functional analyses with 3/8 (38%) agreements. However, the direct observati ons agreed with the functional analyses conclusions for all 4/4 (100%) of the part icipants. These results do not support the concurrent validity of the three descriptive assessments or the cons truct validity of the checklist or the interview, but do support the construct valid ity of direct observations.

PAGE 12

1 CHAPTER 1 INTRODUCTION This chapter provides a brief overview of the literature and presents the need to examine the validity of various assessments in the functional behavior assessment (FBA) process. This brief overview outlines the us e of FBA with preschool-aged children who demonstrate challenging behavior and also highlights articles that have directly evaluated the results of these various assessments. The ar ticles that have dire ctly evaluated various assessments compare the reported results of th e descriptive assessments to each other and to results of functional analyses. The literat ure provides a synthesis of the evolution of the FBA process, the need for distinct comparisons of the various assessments and a rationale for the current study. In addition, th e contributions this study makes to improve the effectiveness of the FBA methodology, in both research and pract ice, are identified. Finally, this chapter presents the specific res earch questions that we re addressed in this study. Overview As the utilization of functional behavior assessments (FBAs) increases, it is incumbent on researchers to improve FB AsÂ’ current practices by evaluating the effectiveness and the efficiency of this proc ess. The increasing prevalence of FBAs is a direct result of two interrelat ed factors. The first reason for more widespread application of FBAs is the increased number of childre n who are engaging in challenging behaviors (Chandler & Dahlquist, 2002). The number of young child ren exhibiting challenging behaviors is rising. Many of these childr en may ultimately meet the criteria for

PAGE 13

2 emotional/behavioral disorder s (Fox, Dunlap, & Powell, 2002). According to WebsterStratton (2000), 7% to 25% of preschool-aged children currently meet criteria that place them at risk for future development of oppos itional defiant disorder or a conduct disorder. For example, Kaiser, Kai, Hancock and Fost er (2002) tested 332 th ree-year old children enrolled in Head Start and found that 39% of boys and 22% of girls demonstrated behavioral deficits that placed them in th e clinical range for both internalizing and externalizing problem behaviors on the Ch ild Behavior Checklist (Achenbach, 1992). FBAs are one strategy that can be used to address childrenÂ’ s needs; however, many of the FBA procedures used are problematic (Johnston & OÂ’Neill, 2001). Increased effectiveness of FBAs are needed to appr opriately meet the increasing needs of young children with challenging behaviors. Second, the reauthorization of the Indi viduals with Disabi lities Education Amendments (IDEA, 2004) has mandated the us e of functional behavior assessments for all students who have been suspended for ove r ten days, who demonstrate behaviors that warrant consideration of a change in educat ional placement, or disciplinary measures such as placement in a more restrictive se tting (i.e., an interim a lternative educational setting) or expulsion. As a result, school dist ricts now view the comp letion of an FBA as the first step in any signifi cant intervention. This IDEA mandate has also impacted early childhood personnel working with young children ages 3 to 6 served under Section 619 of Part B of the act. The prevalence of FBAs in early childhood settings has been stimulated both by IDEA legislation (Nie lsen & McEvoy, 2004) and recognition that the process of FBA represents an importa nt methodology in addressing challenging behaviors demonstrated by young child ren (Fox, Dunlap & Powell, 2002).

PAGE 14

3 Functional behavioral assessment is a pro cess in which information is gathered regarding a specific behavior or set of behaviors in orde r to determine whether that behavior has any predictable relationship with the environm ent (OÂ’Neill et al., 1997). Relationships exist when these environmen tal factors both predict the behavior and explain why it is maintained. Identification of these relationships increases understanding of why a particular behavior o ccurs or how behavior is rein forced and maintained by the specific environmental consequences (i.e., f unctions) they elicit. The FBA must identify the environmental factors surrounding challengi ng behaviors. Once environmental factors are identified and understood thr ough an effective FBA, indi vidualized interventions may be designed to replace undesirable behavior with appropriate alternatives that efficiently obtain the same functions of th e challenging behavior. For example, if a child engages in challenging behaviors to receive teacher attention, an intervention is designed that gives the child increased access to teacher atten tion by demonstrating desired behaviors and decreased access to teacher at tention for challenging behaviors. In most educational settings, information for FBAs is gathered us ing a combination of bot h indirect and direct descriptive assessment measures (Johnston & OÂ’Neill, 2001). Indirect measures are those that do not require direct obser vation of the student. Indirect measures include interviews with parents, teachers, and children via the us e of questionnaires, rating scales and record reviews. In contrast, direct measures typi cally consist of sca tter plots (Touchette, MacDonald, & Langer, 1985) and observation worksheets that reflec t a running narrative of antecedents, behaviors and consequences (ABCs) (Bijou, Peterson, & Ault, 1968) that are recorded as they are obs erved in natural settings.

PAGE 15

4 An alternative method for gathering info rmation to determine the function of a challenging behavior is a f unctional analysis (Sasso, Conroy, Stichter, & Fox, 2000). Functional analysis refers to the use of analog conditions, manipulating specific environmental conditions that follow challengi ng behaviors, to systematically identify consequences as they relate to the functi on of a behavior (Sugai, Horner, & Sprague, 1999). Following the manipulation of each of these conditions, the consequence that produces the highest level of behavior is identified as the function. The difference between the former two descri ptive methods (direct and indi rect) and the third method of functional analysis is that the descriptive methods generate correlational results while a functional analysis identifies causal relations hips between behavior and the maintaining consequences (Lerman & Iwata, 1993). No combination of these three methods has been accepted as the standard protocol either by re search literature or federal mandates (Scott et al., 2004). A literature review of the research of the use of FBAs with young children demonstrating challenging behaviors revealed three distinct groupings : (a) studies that only employed indirect and di rect descriptive methodologies to determine hypothetical functions; (b) studies that combined descriptive methodol ogies with functional analysis methodologies to determine functions; and (c) st udies that directly compared the results of different assessments with each other and/or a functional analysis. Although some FBA methodologies have been available fo r use by teachers for over 20 years (Ervin, Ehrhardt & Poling, 2001), FBA me thods are still in a fledglin g state of development. Studies that only used desc riptive assessment methods in the FBA process remain the most prevalent for young children demons trating challenging behavior. By using only

PAGE 16

5 descriptive assessments, and not using functional analysis, th e function identified in this body of research remains hypothetical. Interv entions based on the hypothetical functions that are not experimentally ma nipulated (i.e., a reversal desi gn) only provide correlational support (Kazdin, 1982). Additionally, in studies that only used descriptive assessments, many of the interventions implemented within th e studies were either not directly related to the hypothetical function or included additional interventions that muddled whether or not the hypothetical function id entification was accurate. Fo r example, Grandy and Peck (1997) hypothesized that the function of the challenging behavior was teacher attention. However, rather than creating an interv ention that limited teacher attention for challenging behavior and increased teacher at tention for desired behaviors, researchers implemented a token economy that led to cla ssroom privileges and that indirectly led to teacher attention. This lack of a direct c onnection between the determined function and the intervention implemented calls into ques tion the accuracy of the determined function. It is impossible to determine if teacher attent ion was reinforcing the desired behavior or if it was another element of the token economy. As illustrated through this example, the body of literature using descriptive assessments gives minimal support to the accuracy or consistency of the various descriptive assessments. Studies that employ both descriptive assessments and functional analysis methodologies also have certain shortcom ings. First, researcher influence was pronounced in these studies th at included the time cons uming, technically complex functional analysis process. In many of th e studies reviewed, a team of researchers worked with very few participants (i.e., Doggett, Edwards, Tingstrom & Wilcznski, 2001; Harding et al., 1999; Romaniuk et al., 2002). Another salient difference between

PAGE 17

6 the descriptive methods in the studies that used both descriptive assessments and functional analysis and the sample that only us ed descriptive methods is that this sample was more parsimonious in the use of descri ptive assessments. When functional analysis was also incorporated into the studyÂ’s protoc ol, descriptive assessments were only used to the extent that they could guide the functional or structur al analysis procedures. These studies used the descriptive assessment results to identify times and situations to conduct analog probes during the times or activitie s that teachers identified as the most problematic. For example, in both their studies Broussard an d Northup (1995, 1997) only used direct observations of the participan t and a brief interview with the teacher to determine what would make the analog condi tions most representa tive of the natural environment. This makes it very difficult to determine what parts of the process are producing accurate and/ or cons istent results. Therefore, th e effectiveness and efficiency of various FBA method ologies are unknown. There were eight studies that directly evaluated different aspects of the functional behavior assessment process. This group of st udies that attempted to compare the results of different descriptive assessments with each other and with functional analysis procedures had several prominent characterist ics. Overall, there was agreement between the different descriptive assessments result s. There was also agreement between the descriptive assessment results and the functiona l analysis results. However there are three reasons why these results should be interp reted with caution. Firs t, the descriptive assessment results informed one another and were not independent from each other as they were conducted, making direct comparis on of separate results impossible. For example, in Ellingson, Miltenberger, St ricker, Galensky and Garlinghouse (2000),

PAGE 18

7 Murdock, OÂ’Neill and Cunningham (2005) and Ne wcomer and Lewis (2004), the results of the indirect assessment dir ectly shaped and guided the direct descriptive observations thereby compromising their claims of i ndependent consistency across assessment methods. This makes direct comparison impo ssible as the results of one assessment influence the results of another. The second in terrelated factor with this assimilation of assessment methods is that by not being pres ented in different orders, the order of administration was not counterbalanced. As a result, the aforementioned three studies and Yarbrough and Carr (1997) presented their desc riptive assessments in the same order, making it impossible to control for an order eff ect. In other words, there is a possibility that a participating teacherÂ’s previous respons es to an assessment would shape responses to subsequent assessments. The third salient factor is the use of the descriptive assessment results to guide the functional analysis procedures. Within this body of literature, only Cunningham and OÂ’Neill (2000), and Sasso et al. (1992) conduc ted functional analysis procedures with five analog conditions: free play (control), tangible, escape, attention, and alone. Other studies only tested analog conditions that matched the results of the descriptive assessments. This difference is subtle yet critical. By only eval uating those hypothetical functions determined by the descriptive assessmen t, it is possible to confirm the results as a function but not the function (because different f unctions were not compared). Additionally, two studies, Murdock, et al., (2005) and Ellingson et al. (2000), did not use any experimental procedures to confirm the functions; this means that their results may show consistency across descriptive measures but that does not confirm the accuracy of the function identified.

PAGE 19

8 In conclusion, all three bodies of FBA lite rature indicate that there is a need to evaluate the effectiveness of current assessment procedures By keeping the different assessment results separate from each othe r and from the functional analysis, it was possible to evaluate each instrument’s resu lts independently. This extends current FBA research that has attempted to evaluate the accuracy of different assessment methodologies. Statement of the Problem No single procedure or combination of procedures for functional behavior assessment has emerged as the most suitabl e for use with young ch ildren demonstrating challenging behaviors. Minimal empirical studies have eval uated the effectiveness and usefulness of various approaches. “Several au thors have noted that research directly comparing the results of different approaches such as interviews a nd direct observations is sparse” (Johnston & O’Neill, 2001, p. 206). Failure to establish FBA as a meaningful process with valid outcomes will adversel y impact countless children who demonstrate challenging behaviors. Absent a highly effective method fo r determining function and designing effective interventions, appropria te plans will either go unused or the interventions that result from the process wi ll be ineffective. Teachers need to be equipped with strategies that determine the function of a challengi ng behavior reliably and accurately for the FBA process to be effect ive and useful. If the process is ineffective and does not decrease challeng ing behaviors, then the FBA process will become another bureaucratic chore that will not positivel y impact children (Scott et al., 2004). Significance of the Study This fundamental need to meet the de mand for effective FBAs that can be implemented by teachers in a meaningful way l eads to an important area of research: to

PAGE 20

9 conduct a series of indirect and direct desc riptive assessments a nd compare the outcomes of these assessments to each other and to the outcomes of a functional analysis methodology. This study extends and builds on the current literatu re that directly evaluated the results of different descriptiv e assessments. However, unlike these studies, the functional analysis in this study was independent of the descriptive assessment results, thus allowing an unbiased comp arison of descriptive and experimental methodologies. The results of the various direct and indirect descrip tive assessments were recorded independently and sepa ration of the results was maintained until the end of the study. Additionally, the desc riptive assessmentsÂ’ orde r of administration was counterbalanced across participan ts to maintain independence as well. The results of this research are an empirically validated step to wards evaluating descriptive assessments that can be used both in continuing FBA research and in applied public school and early childhood settings. Purpose of the Study The specific questions this research study addressed were the following: 1. How do the functions of target behavior s identified through various descriptive functional assessments compare with f unctions identified through a functional analysis for young childrenÂ’s challenging behavior? 2. How do the functions of target behavior s identified through various descriptive functional assessments compare to each other?

PAGE 21

10 CHAPTER 2 LITERATURE REVIEW The purpose of this chapter is to review the literature on the use of functional behavior assessment (FBA) that address ch allenging behaviors in applied settings demonstrated by young children. This review focuses on the processes that comprise FBAs, in order to determine the most effec tive methodologies. First, the significance of completing FBAs will be traced in the context of an increase in the rate and severity of challenging behaviors demons trated by young children. Second, this chapter will define functional behavior assessment, and outline the current components that comprise FBA. Third, the theoretical underpinnings will be outlined as they relate to its current application especially with young children. Fourth, specific as pects of this application will be examined through a targeted review of the empirical literature. This literature will be divided into three sections: (1) studies th at used only descriptive assessments in the FBA process (2) studies that used both de scriptive assessments and experimental analyses; and (3) studies that directly exam ined and compared the results of different descriptive assessment methodol ogies. Finally, a general di scussion will address the shortcomings of current empirical literature and outline directions for future research. Prevalence and Impact of Young Children with Challenging Behaviors The prevalence of young children exhibiting challenging behavi ors is increasing (Fox, Dunlap, & Powell, 2002), and children ar e at risk for future development of behavioral disorders (Webster-Stratt on, 1997). The young children under consideration are defined as children under 8 years of age, in accordance with the age limit specified by

PAGE 22

11 the Division of Early Childhood of the Counc il for Exceptional Children. Researchers have documented various prev alence rates across a numbe r of different behavior disorders. For example, 7% to 25% of pres chool-age children curren tly meet the criteria for a diagnosis of Oppositional Defiant Diso rder (ODD) as defined by the Diagnostic Statistical Manual IV-TR (DSM IV-TR) (W ebster-Stratton, 2000). The DSM defines Oppositional Defiant Disorder (ODD) as “a re current pattern of negativistic, defiant, disobedient, and hostile behavior toward auth ority figures that persists for at least 6 months.” Additionally, Kaiser, Kai, Hancock, and Foster (2002) tested 332 three-year old children enrolled in Head Start and found that 39% of boys and 22% of girls scored in the clinical range for both internalizing and extern alizing problem behavior s. As the rate of challenging behaviors demonstr ated by young children continues to increase, the need to manage and modify behavior in early childhood settings also increases. The inability to remediate these challenging behaviors w ill affect young children both now and in the future. Behaviors such as tantrums, aggressi on, and chronic non-compliance exhibited by young children may have immediate and prolonged effects on academic achievement and appropriate socialization (Shonkoff & Phil lips, 2000). The immediate effects of challenging behaviors for preschool-aged ch ildren are academic under-achievement and poorly developed social skills (Chandler & Dahlquist, 2002; Webs ter-Stratton, 1997). The long-term effects of unaddressed challenging behaviors are an in crease in both rate and severity over time (Reid & Patterson, 1991). Early behavior problems have also been linked with more severe problems such as substance abuse, unemployment, criminal behavior, and diagnosis of a psychiatri c disorder (Reid and Patterson, 1991).

PAGE 23

12 Chandler and Dahlquist (2002) state that difficult behavior s are often rooted in the functions that they served at young ages, such as hitting a peer to a ttain a tangible object (e.g., a toy), and these behaviors may also incr ease in frequency, duration, and intensity as the child gets older. Ascertaining the func tion of a challenging behavior is the key to developing an effective intervention to ex tinguish it (Cooper & Harding, 1993; Iwata et al., 1982/1994; Gresham, 2003; O’Neill et al., 1997; Sasso et al., 2001) and may prevent its undesirable immediate and long-term consequences. Defining Functional Behavior Assessment The use of functional behavior assessments (FBAs) in school settings represents an important process in the effort to address increasingly prevalent pr oblems with behavior in children. Using functional behavior assessment and functionally derived interventions, as opposed to interventions that are unrelated to the function of the behavior, is supported by research and considered “best practice” in the field (Arndorfer & Miltenberger, 1993; Gresham, 2003; Iwata et al., 1982/1994; National Institute of Health, 1989; O’Neill et al., 1997; Repp, Felce, & Barton, 1988; Sasso et al ., 2001). Functional behavior assessment was designed to identify variab les to manipulate through appl ied interventions to reduce challenging behaviors that interf ere with or prohibit learning, and increase pro-social and adaptive behaviors that can lead to greater academic and social success (O’Neill et al., 1997). In schools, the use of FBA represents a paradigm shift for educators that traditionally applied a standard system of c onsequences for all infractions (i.e., time out, a phone call to parents, a loss of privilege s, suspension, expulsion). In early childhood settings, FBAs represent a valuable tool to help early childhood prof essionals address the behaviors that surpass usua l challenging behaviors dem onstrated by young children (Fox,

PAGE 24

13 Dunlap, & Powell, 2002). With FBA, educator s can design interventions that match the function of a challenging behavi or in children. The function of a behavior refers to the purpose or outcome of the challenging behavior for the child (Scott et al., 2004). In other words, a function refers to the desired cons equence that motivates a challenging behavior, and function-based interventions refer to environmental manipulations that allow a student to achieve the desi red goal through other means (i .e., appropriate replacement behaviors). For example, if the function, or purpose, of a student exhibiting a challenging behavior is to receive teacher attentio n, then a challenging behavior could be extinguished by giving the student access to teacher attention by exhibiting appropriate behaviors and not providing addi tional teacher attention for the challenging behavior. Functional behavior assessment, as defined fo r this literature review, is a process in which information is gathered regarding a specif ic behavior or set of behaviors in order to determine whether that behavior has any pr edictable relationship with the environment (OÂ’Neill et al., 1997). The environment is comprised of all factors preceding (antecedents) and events fo llowing (consequences) the beha vior. A relationship exists when these environmental fact ors predict the occurrence of the behavior and explain why the behavior is maintained. This assessment helps to understand why behavior occurs, or how behavior is reinforced and maintained by the specific environmental consequences (i.e., functions) they elicit. This assessment al so identifies the environmental factors that precede a challenging behavior and make the ch allenging behavior more or less likely to occur. The purpose of an FBA is to unders tand environmental fact ors that surround the behavior so that individua lized interventions may be de signed to replace undesirable

PAGE 25

14 behavior with appropriate alternatives th at efficiently match the functions of the challenging behavior. Because current empirical studies have applied the term functional behavior assessment broadly, the definition of FBA for this literature review requires further specification. As noted, functions typically refer to environmental consequences that follow a challenging behavior. The experimental evaluation of functi ons of behavior is referred to as ‘functional analysis’ within the literature. However antecedents, (also referred to as setting events or establis hing operations) that precede and increase the likelihood of challenging behaviors, also play an important role in functional behavior assessment (Conroy & Stichter, 2003). Within th e literature, experime ntal evaluation of antecedents is referred to as a ‘structural an alysis,’ but may also simply be included as another aspect of an FBA. For the purposes of this literature review, studies that evaluate both the antecedents and consequences of behavior will be included as part of an FBA. There are essentially three ways to generate functions of a cha llenging behavior: (1) indirect descriptive measures that establish hypothetical functions of a challenging behavior, (2) direct descriptive measures that establish hypothetical functions and (3) experimental analysis of an e nvironment that can determine the actual function of a challenging behavior. However, for the purposes of this literature review direct and indirect measures will be combined and si mply referred to as descriptive assessment measures in the three sections of literature that are reviewed (S turmey, 1994). Indirect measures are those that do not require di rect observation of the student and the challenging behavior. Typically, indirect measures include interviews with parents, teachers, (and the student if appropriate), questionnaires, rating scales and reviews of

PAGE 26

15 school records. A number of commercially produced comprehensive interviews exist (see OÂ’Neill et al., 1997 for an example). In addition, commercially produced behavior checklists, such as the Motivation Assessment Scale (MAS) (Durand & Crimmins, 1992) are designed for determining the hypothetical function of a behavior. The total number of indirect assessments of this t ype is difficult to determine. A ccording to a literature review conducted by Floyd, Phaneuf and Wilczynski (2005) there are 19 commercially produced indirect assessments; however, this does not include those in direct assessments that are created by individual school systems or early childhood centers. Furthe r, as Floyd et al., (2005) point out most of these other instrument s have minimal research into their efficacy and validity; only OÂ’NeillÂ’s interview and the MAS met their criteria for inclusion in a literature review of indirect descriptive assessment methods. In contrast, direct measures require th e direct observation of the child and the challenging behaviors. Direct descriptive asse ssments typically consist of scatter plots (Touchette, et al., 1985), and observation work sheets that reflect an tecedents, behaviors and consequences (ABCs) (Bijou, et al., 1968) that ar e recorded in real time as they are observed. While some school districts and early childhood centers may adopt commercially produced FBA measures, they also may have created their own direct and indirect methods of gathering data to identify these functions. This means that there are a large number of descriptive methodologies that can be used within the FBA process. Finally, the third type of procedure is an experimental approach, also referred to as functional and structural analysis. Functiona l analysis refers to the use of analogs, manipulating specific environmental conditions that follow challenging behaviors, to systematically identify consequences as they relate to the function of a behavior (Sugai,

PAGE 27

16 Horner, & Sprague, 1999). For example, to determine the function of the behavior, analog situations representing possible conseq uences are manipulated (i.e., attention, tangible, escape, and control conditions). Fo llowing the manipulation of each of these, the consequence that produces the highest level of behavior is identified as the function. For example, to determine the function of a challenging behavior, each analog condition is conducted, contingently providing the ta rgeted consequence following the occurrence of the behavior. For example, in the a ttention condition, the c onsequence of adult attention would be provided c ontingently every time the target behavior is demonstrated while all other possible antecedents and conse quences are controlled to account for any variance. If the consequence of adult attention results in an increased rate of the behavior, as compared with other consequences admini stered in the same controlled environment, then it can be concluded that the function of this challenging behavior is adult attention. Structural analysis involves a similar process; however, rather than systematically manipulating the consequences that follow a be havior, a structural analysis manipulates the antecedents that precede the behavior to identify variables that set the occasion or serve an evocative effect. All three of these assessment procedures ha ve drawbacks. Indirect assessments rely on the subjective judgments of a third part y, and direct observati ons can be context specific and inefficient in capturing the o ccurrence of a challenging behavior and its antecedents and consequences (Floyd et al., 200 5). Direct observations can be inefficient because the probability of actually observing th e challenging behavior may be relatively low, necessitating long and/or frequent obser vation periods (Scott et al., 2004). In this sense, the measures are inefficient because they may demand a considerable amount of

PAGE 28

17 time from the practitioner but do not insure m eaningful results. Func tional and structural analysis have been identified as too time consuming, and technically complex for practitioners to conduct, and may lack ex ternal validity due to the high degree of experimental control (i.e., the highly cont rolled conditions in wh ich it is conducted) (Arndorfer & Miltenberger, 1993; Ellingson, et al., 2000). Regardless of the methodology, identifying and evaluating relationships between environmental antecedents and consequences and challenging behavior improves the opportunities to enable all stud ents to meet academic and behavioral goals (Hanley, Iwata & McCord, 2003). As antecedents and conseque nces that are related to challenging behavior are identified, envi ronments can be modified to reduce the occurrence of challenging behavior. This may be done by re ducing the antecedents that increase the occurrence of the challenging behavior, givi ng students access to de sired reinforcement through adaptive behaviors and/or limiting access to reinforcement for challenging behaviors. Creating effective in terventions that re duce or eliminate challenging behavior at a young age can decrease the possibility that they will increase in frequency, duration and intensity as the child gets older. The theoretical framework for designing interventions according to function is rooted in the Behaviorist approach to human psychology. This theoretical framework and an historical perspectiv e supporting FBA are described in the next section. Theoretical Framework and Histo rical Perspective Supporting FBA The theoretical framework for functional behavior assessment methodology can be traced back to the work of early behavior ists such as J.B. Watson and, so-called neobehaviorists such as B.F. Ski nner (1953). In fact, the term function as it is used in this context, is often cred ited to SkinnerÂ’s book Science and Human Behavior (1953). Skinner

PAGE 29

18 (1953) states, “The external va riables of which behavior is a function provide for what may be called a causal or functional analysis.” (p. 35). The subsequent work in the field of applied behavior analysis, codified by the creation of the Journal of Applied Behavior Analysis (JABA), expanded this theoretical fram ework to work with humans in applied settings. Within this framework, researchers address behaviors by examining the environmental variables that maintain them, and operate on the assumption that functional relations not only exist, but also can be iden tified and manipulated. Early seminal work in JABA included an articl e by Bijou, Peterson and Ault (1968) that described a method to integrate descriptive a nd experimental data within field studies. Additionally, Sailor, Guess, and Rutherfo rd (1968) conducted early work identifying “functions” of behavior and re ported that for a 9 year ol d who engaged in frequent tantrums, the problem behavior “seemed to pr oduce the effect of terminating contact with other individuals who usually were making some demand of her” (p. 238). While establishing the function of a behavior can be traced back to the work of early behaviorists, the emergence of a specific proc ess for determining the behavioral functions is related to two more recent seminal events; the creation of a functional analysis methodology and the legal mandate of FBA in th e Individuals with Disabilities Education Act (IDEA, 1997; 2004). First, Iwata et al. (1982/1994) devel oped a methodology that used various environmental conditions, referred to as analog probes, in a clinical setting to identify a relationship between the environment and the function of self-injurious behavior. These analog probes included the following conditi ons: adult attention, escape from a task, an

PAGE 30

19 alone condition and free play (Iwata et al., 1982/1994). The direct, systematic manipulation of environmental factors in a highly controlled environment in order to evaluate the function of a challe nging behavior, in this case self-injurious behaviors, was termed ‘functional analysis.’ This procedur e led to highly specific, highly effective intervention plans that incorporated the f unction of the challengi ng behavior and thus, decreased these behaviors. As the importanc e of finding the functi on of a challenging behavior became apparent, the methodol ogies broadened beyond experimental manipulation into more desc riptive assessments. Following this groundbreaking research, a more formalized functional behavior asse ssment process emerged from the clinic setting and was implemented in more natu ral settings, such as the classroom. The second critical event, th e reauthorization of the Indi viduals with Disabilities Education Amendments (1997), provided the impetus to extend the methodology by making the use of functional behavior asse ssments (FBAs) mandatory for all students whose behavior warrants consideration fo r a change in educational placement or disciplinary measures such as extended suspension or expulsion. Although IDEA (1997) provided the impetus to expand the met hodology of FBA, it did not specify what procedures should be used by individual s in schools when implementing an FBA resulting in varying forms of FBAs from state to state, district to district and even school to school. Despite its continua lly increasing prevalence in both educational systems and research settings, the effectiveness of FBA is still hampered by muddled definitions, an unspecified protocol and com ponents that have not been empirically validated for use with various populations in numerous sett ings (Ervin, Ehrhardt, & Poling, 2001; Gresham, et. al, 2004; Scott et al., 2004; Sugai, et al., 1999).

PAGE 31

20 While IDEA mandates have specific impli cations for public schools, the legislation has also continued to stimulate interest in the FBA process as it impacts early childhood settings such as child care centers (Neils en & McEvoy, 2004). This prevalence has a clear relationship to the federal mandates regardi ng positive behavior support and the need for evidence-based practices for all ages (s ee Conroy, Dunlap, Clark, & Alter, 2005 for a review). As outlined by Arndorfer and Milt enberger (1993), there are specific implications for FBA in early childhood settings. These implications include examination of the utility of functional assessment methods in early childhood settings, especially including the use of various different descri ptive assessments (i.e. indirect and direct assessments) and the viability of functiona l analysis procedures in early childhood settings. Thus, there is a need for examination of these specific components, as the role of FBA in early childhood settings continues to be prevalent in research and practice. No single procedure or combin ation of procedures has emerged as the most suitable for application with young children demonstrati ng challenging behaviors. In concert with this shortcoming, minimal empirical studie s have evaluated the effectiveness and usefulness of these various approaches. “Several authors have noted that research directly comparing the results of different approaches such as interviews a nd direct observations is sparse” (Johnston & O’Neill, 2001, p, 206). Failure to establish FBA as a meaningful process with valid outcomes will adversel y impact countless students who demonstrate challenging behaviors, because a highly effective method for designing effective interventions will either be unus ed or the interventions that result from the process will be ineffective.

PAGE 32

21 The application of various functional behavior assessment procedures in educational settings has outpaced current res earch. There is minimal research validating the effectiveness of different functional beha vior assessment processes currently used for children with emotional and behavioral di sorders (Sasso, Conroy, Stichter & Fox, 2001). “Given the different options available for conducting a functional assessment, the question regarding which approach practitione rs ought to use to assess problem behavior is paramount” (Yarbrough & Carr, 2000, p, 133). The rising prevalence of young children’s challenging behavior in early childhood settings increases this urgency. Thus, educators need to be equipped with strategies that determine the function of a challenging behavior reliably and accurately for the FBA process to effective and useful. Review of the Current Literature Current research in the use of FBAs, incl uding functional and structural analyses, with young children indicates th at although there is consider able application of these methodologies, the use of FBAs with young children with EBD in applied or natural (e.g., classrooms, home) settings is still in its fledgling state. The first section of this review examines the use of both direct and indirect descriptive FBA assessment methodologies in natural settings (i.e., such as childcare centers, public school classrooms, and homes). The rationale for revi ewing this group of empirical studies is that they can provide insight into the effectiveness of FBAs that are conducted without any type of systematic environmental mani pulation. The second section examines FBAs that used both types of desc riptive assessments and experi mental analysis for FBAs. Examining this literature will pr ovide a clearer picture as to what place, if any, functional analysis will hold in natural se ttings and what the ramificatio ns of not using experimental analysis may be. The third section examines st udies that specifically compared the results

PAGE 33

22 of descriptive assessment methodologies with each other and with the results of a functional analysis. This purpose of this sec tion of the literature review is to examine what research has been conducted thus far comparing, evaluating a nd possibly validating different components of the FBA process. Method to Select Reviewed Studies This literature search was completed in th ree steps. First, th e author conducted a computer search of EbscoHost including Academic Search Premier, PsychINFO, Psychology and Behavioral Sciences Coll ection, and the Professional Development Collection databases), ERIC, and First Sear ch. The following keyword combinations were used: functional, assessment, assessment-based, analysis, behavior, experimental, school, preschool and young children Second, articles were cro ss-referenced that were identified in the computerized search to de termine if research articles had been omitted. Third, the author consulted with researchers who have written extensively in this area to accumulate any other articles that may ha ve been omitted in th e first two steps. To compare ‘descriptive assessments’ (d irect and indirect) to ‘experimental analysis,’ it is necessary to differentiate the two terms. As previously described, descriptive assessment represents the use of methodologi es including: direct observation, rating scales, interviews, and hypothesis development using antecedent behavior consequence (ABC) analysis (Bijou et al., 1968). Experimental analys is (i.e., structural analysis and functional analysis) uses an alog probes to purposefully manipulate a variable within the environment and then develops hypotheses based on these controlled, manipulated environments. One important dis tinction between these two terms is that descriptive assessment met hodologies identify only a hypothetical function; whereas,

PAGE 34

23 functional analyses are able to identify a causal relationship (i .e., function) through experimental analyses (Lerman & Iwata, 1993). However, certain character istics of these research articles made separating descriptive methodologies from f unctional analysis a complex task in several ways. First, all descriptive studies include d hypothesis testing, the addition or removal of a specific stimulus in the environment. This type of environmental manipulation is similar to a traditional FA. However, two features differentiated hypothesis testing and functional analysis: (1) the type of experimental contro l and (2) the number of functions tested in these controlled settings. Control of the envi ronment is the distingui shing characteristic separating descriptive analysis from functi onal analysis (Sasso et al., 1992). For example, if researchers manipulated the environment to keep every other factor (curriculum, schedule) the same and only sy stematically changed the analogs, this was termed a functional analysis. Additionally, if a study tested multiple functions in this controlled environment then it was termed a functional analysis. For example, if a study systematically tested tangible items, adult attention and an alone condition, this was considered a functional analysis. However, if a study only performed a test of one hypothetical function based on th e interviews or direct obs ervations, this was still considered a validation of the findings from the descriptive assessment (Hanley, et al., 2003). Another characteristic that made disti nguishing descriptive methodologies from experimental analysis difficult is the absen ce of clearly standardi zed terminology in the research studies. In certai n studies, the actual methodology is obfuscated by an imprecise use of terms. For example, Kamps et al. (1995) referred to their research as a “functional

PAGE 35

24 analysis.” However, because of minimal e nvironmental control, the methodology used in this study is more accurately judged to be a descriptive analysis, according to Sasso et al.’s (1992) definition. To counteract the use of imprecise terminology within the literature, the criterion for categorization was se t that if individuals (researchers, teachers, parents) manipulated the environment to test multiple different functions in any way, in order to confirm a hypothetical function, then it was deemed an experi mental analysis. There are two final caveats regarding this differentiation. The fi rst issue concerned studies that implemented interventions-even if interventions were implemented explicitly to confirm hypothesized functions. Regardless, these studies were still judged to fall under the ‘solely descriptive’ section of the liter ature review. The rationale for this decision is that while in terventions may confirm a functional relationship, they do not necessarily confirm the functional relationship, unlike a tr aditional functional analysis. Other functional relationships may exist that would not be tested simply by implementing interventions guided by descri ptive assessment results. For example, a descriptive assessment suggests that desire for a tangible object may be the function of the behavior and an intervention is implemented that provi des increased time playing with the toy for desired behaviors and decreased time playing with the toy for challenging behaviors. Although this may be an effective way to reduce challenging behaviors, it does not indicate that this is either the only func tion or the primary function of the challenging behavior. The function may actually be the atte ntion the student gets when playing with a toy, thereby making the real function, atten tion. As an aside, the efficacy of using interventions to confirm or disprove functional relationships will be discussed within the first section of this literature review, studies using solely descriptive assessment studies.

PAGE 36

25 The second caveat was the issue of investig ating antecedent-based assessments, at times referred to as structural analysis, and antecedent-base d interventions. The rationale for including these studies within this review is based on the use of antecedents within classrooms and other natural settings. It has been noted that antecedent events play an important role in natural setti ngs because less predictable pa tterns of behavior occur in these environments, less control over antecedents occurs, and schedules of reinforcement may not be consistently applied (Conroy & St ichter, 2003). Furthermore, it is reasonable to assume that teachers completing the FBA process will not limit the process in the widespread application of FBAs, solely to consequence-based assessment and intervention. Thus, studies that addressed antecedents are included in this review. Additional criteria to determine which re search studies would be included were also determined. First, only studies that occurred in homes, community-based childcare, and early childhood public school classrooms were used. Because this review is focused on the use of FBA in real-world settings, its use in highly controlled clinical settings was not applicable. As a result, a number of articles were excluded due to this potential lack of external validity. Second, for the first two sections, research studies directed solely toward populations with low-incidence disabi lities, such as autism, or moderate and severe mental retardation were eliminated, as was research geared toward highly specific behaviors that are not typi cal of most young children who demonstrate challenging behavior, e.g. hair-twirling, ey e-poking, breath-holding, were e liminated. Therefore, as long as at least one participant within the study had been classified as at risk for or having a high-incidence disability, the study was include d within this review. Only the results of the children with high incidence disabilities we re discussed in the fi rst two sections of

PAGE 37

26 this review. Criterion was established to fo cus on the application of FBA with a broadbased population – that is, young children with or at risk for emotional/behavioral disorders. However, for the final section of th e literature review that consisted of studies that directly compared and evaluated the re sults of different desc riptive methodologies, this criterion was expanded due to the dearth in this area of research. For the final section, results of students who were older or had mo re severe disabilities were also reported. Third, in order to focus on FBAs with young children, each study was to include at least one participant 8 years of age or younger. Fourth, only re search published in a peerreviewed journal after 1994 was included. Because this methodology has been revised and refined extensively over the past ten years, articles pre-dating 1994 would not be applicable to current applications. These criteria were again e xpanded for the final section due to the limited numb er of studies in this area. Results of the Literature Review Thirty-two studies were found that met the criteria for inclusion in this review. Fourteen studies used descriptive methods onl y to identify the hypothetical function of a challenging behavior, ten studies used both descriptive methods and functional analysis to identify the function of a ch allenging behavior, and eight st udies directly compared the results of descriptive assessments with functional analysis resu lts. In the first two sections of this review, (1) solely descriptive a ssessments (indirect and direct), and (2) experimental analysis and descriptive co mbined, were evaluated by looking at: (a) participant and setting characteristics, (b) im plementer characteristics, (c) components of the descriptive approach or analog conditions (d) antecedents and functions identified, and (e) interventions implemented. In the fi nal section, each of the eight studies is presented individually. These studies are discussed in te rms of the aforementioned

PAGE 38

27 characteristics, but their findings and imp lications are also summarized. Patterns of agreement or disagreement are also pres ented as well as methodological limitations. These eight studies are presente d in chronological order. The st udies in this final section are presented in greater depth, because th eir rationales and methodologies provided the foundation for the current study. Descriptive Assessment Methodologies This section of the paper will address the research f ound on the use of descriptive assessments in the formulation of FBAs in natural settings. Descriptive assessment methodologies, both direct and indirect, are pr ocedures that evaluate the function of a challenging behavior but do not systematically manipulate the environment. Therefore, there is no experimental validation of th e potential antecedents and consequences identified. In regard to FBAs with young children, descriptive assessments including interviews, checklists and direct observations (ABC recordings) are far more prevalent than the more technically complex, time consuming experimental analyses (i.e., functional analysis and struct ural analyses) (Arndorfer & Miltenberger, 1993; Reid & Nelson, 2002; Scott et al., 2004). While descriptive assessments are le ss time consuming and less technically rigorous, the main drawbacks are that they can only generate hypothetical functions of behavior. Additionally, descriptive assessments often rely on the judgments of third party informants. This can be especially problema tic as questionnaires and interviews were subject to the teachersÂ’ memories and biases of behavior in diffe rent settings (Shores, Wehby & Jack, 1999). Thus the value of th ese assessment methodologies hinges on the informantsÂ’ accuracy of recall. Despite their current prevalence in educational settings, there still remains a shortage of literature in regard to studies that used only the

PAGE 39

28 descriptive assessment methodolog ies. A surprising outcome of this review is that although IDEA (1997) produced a flurry of posit ion papers (Sasso, et al., 2001), it did not appear to produce an accompanying flurry of empirical research that met the criteria for this literature review. In f act, there are six studies on or before 1997 and eight studies after 1997. A total of fourteen studies met th e inclusion criteria (see Table 2-1 for a summary of information regarding each study). The following sections will highlight the findings of this group of studi es and discuss methodological and practical limitations of the research. Participant and Setting Characteristics In terms of participant charac teristics, there were a tota l of 274 participants in the solely descriptive studies, however, 210 pa rticipants were from a single study (see Chandler, Dahlquist, Repp, and Feltz, 1999). Because information about ethnicity and socio-economic level was omitted from most of the studies, particularly those with large numbers of participants, this level of detail is not discussed in the review. Omitting the Chandler et al. (1999) study that did not report gender, in the smaller studies that reported participant gender, only 9 girls (between the ages of 3-6 years old) out of the remaining 64 participants (14%) were evaluated in all of the studies. This is not surprising since typically boys demonstrate external challengi ng behaviors at a much higher rate at the preschool age than girls do (Webster-Stratton, 1997).

PAGE 40

29 Table 2-1. Descriptive Methods for Young Children Authors/ Year of Publication Participant Characteristics (number, gender, age in years, label) Setting characteristics Implementer Characteristics Descriptive Methods Functions Identified Interventions Implemented Dunlap, dePerczel, Clarke, Wilson, Wright, White, & Gomez (1994)* 3 males, 5-11 yrs**, EBD and ADHD, selfcontained classrooms Researchers DO, HD, I Not stated, Lack of choice (implied) Use of choice in curriculum recommended Storey, et al., (1994) 1 male, 6 yrs, no label, reg. ed. Classroom Researchers & Teacher I, DO Attention, tangibles Increased specific reinforcement using a reinforcement system Clarke, Dunlap, Foster-Johnson, Childs, Wilson, White & Vera (1995)* 3 males, 5-11 yrs**, EBD and ADHD, selfcontained classroom or separate work room Researchers & Teachers I, RR, DO, HD Lack of interest in assignments Incorporation of student interests Kamps, Elllis, Mancina, Wyble, Greene, & Harvey (1995) 8 males, 2 females, 3-5 yrs, at risk for EBD, reg. ed. Classroom or Head Start classroom Teachersprovided with training workshops, Researchers monitor and assist DO, HD, TR Adult attention, tangibles Increased supervision, reinforcement for prosocial behavior, reduced attention for inappropriate behavior Grandy & Peck (1997) 1 male, 6 yrs, no label, gen. ed. Classroom Researchers and teacher DO, HD, SA,TR Teacher attention Token economy (led to teacher attention Umbreit & Blair (1997)* 1 male, 4 yrs, at risk for EBD, childcare center Researchers & Teacher DO,HD,I, SA Preference, choice Pair preferred activity with non-preferred Schill, Kratochowill, & Elliott (1998) 19 students, 12 male, 7 female ages 4-5 yrs, no label, Head Start classroom Researchers & Teachers DO, HD, I, CL Skill deficit, attention tangible DRO, Time out, Modeling Symons, McDonald & Wehby (1998)* 12 students, all male, 6 yrs, labeled EBD, selfcontained classrooms Researchers guided teachers through all FBA assessments SP, HD, SA Teacher attention, Varied with antecedents Revised schedules, modified assignments, restructured access to attention

PAGE 41

30 Table 2-1. Continued Authors/ Year of Publication Participant Characteristics (number, gender, age, label) Setting characteristics Implementer Characteristics Descriptive Methods Functions Identified Interventions Implemented Blair, Umbreit, Bos (1999)* 4 students, 3 male, 1 female 5 yrs, at risk for EBD, childcare center Researcher and teachers DO, HD, I Attention, choice, preferred task Integrate choice and preference into curriculum Chandler, Dahlquist, Repp, & Feltz (1999) 210 students, 3-6 yrs, 3 types of classrooms special needs (assorted disabilities), at risk, early childhood Teachersprovided with training workshops, Researchers monitor and assist DO, HD Not stated Varied for classroom/ child Kennedy, Long, Jolivette, Cox, Tang & Thompson (2001) 3 students, 6, 6 and 8 yrs, 2 male, 1 female, 1 identified EBD, 2 ‘at risk’, self-contained and gen. ed classrooms Researchers and teachers I, DO, HD Teacher attention, Escape, Topographically specificfunctions varied across contexts Differential attention, selfmonitoring, Eliminate certain consequences (school suspension) Roberts, Marshall, Nelson & Albers (2001) 3 students, 1-1st grade, 2-4th grade**, all males, no identified disability, general ed. classroom Researchers and teachers I, DO, Diagnostic academic task Escape motivated Used curriculum based assessments to modify antecedents VanDerHeyden Witt & Gatti (2001) 14 students, ages 2-4 yrs, 2 classrooms, speech delay, autism, hypothyroidism Researchers and teachers I, DO, HD, Teacher attention (primary) escape, tangible, peer attention (secondary) Differential reinforcement of alternative behaviors (DRA) Packenham, Shute & Reid (2004) 2 students, 1 male, 1 female, 8 & 9 yrs**, no identified disability Teacher trained and completed assessments I, HD through DO (ABC) Teacher attention, escape (only options) Task modifications specific directions, card to access teacher atttention Note DO=direct observation; CL=checklist; I=interview; HD=hypothesis development; RR=record review; SA=structural analysis; SP= scatterplot; TR=teacher reports; DR O=differential reinforcement of other behaviors indicates a structural or an tecedent based assessment used ** indicates only students in age range included

PAGE 42

31 In regard to participantsÂ’ di sabilities, the publication date s of these studies and their methodologies are suggestive of a type of e volutionary process that occurs with FBA research, namely the progression of research w ith more severe, low-incidence disabilities (autism, profound mental retardation) to part icipants with less severe, high-incidence disabilities (EBD, mild mental retardation). Related to participantsÂ’ disabilities is the issue of what types of behaviors were being demonstrated and addressed through functional behavior assessment. Behaviors such as noise making, talking out, property destruction, excessive off-task behavior a nd non-compliance represen ted the majority of the behaviors that were addressed in this sa mple of studies. The settings in which these studies took place are also a hi ghly relevant area of consid eration in regard to the widespread application of the FBA process. Th e studies in this sample used descriptive assessments in self-contained special educat ion classrooms (39%), Head Start classrooms (11%), general education cl assrooms (39%) or childcare center rooms (11%). Four studies included different type s of classrooms in the same study and Clarke et al., (1995) also reported using a teacher workroom for part of the assessment process. The first notable aspect of the literature in regard to practical applications of FBA in the classrooms is the total number of par ticipants used in this sample of studies. Omitting the Chandler et al. (1999) study, studies averaged about 6 participants per study, in terms of all participants (not only the ones that met th e inclusion criteria for this literature review). However, Chandler et al (1999) identified the importance of training teachers as the creators and implementers of the FBAs. Through a series of workshops, teachers were given the requisite knowledge to generate effective FBAs for entire classes.

PAGE 43

32 The emphasis on training for class-wide impl ementation is indicative of the importance of implementer characteristics, as discussed in the next section. The second notable aspect about this sample of studies is in regard to gender. The minimal representation of female participants is very likely an unavoidable aspect of this type of research. The omission of girls fr om all but the larges t descriptive methods studies severely limits the generalizability of the results of the FBA process to this population. However, this proportio n of males to females may be representative of the population schools and early childhood centers identify as necessitating an FBA. The third practical consideration that is a strong point in regard to this sample of FBA studies is the issue of setting. FBAs that are set in highly controlled settings such as clinics threaten the applicability of this process. This is noted in this sample as all of the studies were conducted in natu ral settings. Even the use of separate rooms other than participantsÂ’ classrooms threatens the external validity of the FBA pr ocess, as it exists today. The consequence of this ar tificial setting is that the ge neralizability of these studies remains questionable. Because of their countle ss responsibilities fo r an entire classroom of students, it is unlikely that teachers will be able to spend the necessary time required to conduct one-on-one descriptive assessments of students in separate classrooms. Furthermore, unless the studentÂ’s behavior is automatically reinforced, it is likely to be mediated in some way by social settings. Therefore, conducting assessments away from the typical classroom or early childhood cente r milieu may create ar tificial variables not typically found in the classroom settings. An additional consideration beyond the FBA setting is the consideration of who actu ally conducted and implemented the FBA.

PAGE 44

33 Implementer Characteristics Because educators will ultimately be res ponsible for generating effective FBAs, it is important to note who is gene rating the FBAs in research studies (Scott et al., 2004). In all of the studies in this section of the liter ature review, an effort was made to include general and special education teachers in the creation and implementation of the FBA (see Kamps et al., 1995; VanDerHeyden, Witt & Gatti, 2001). However, in 13 out of 14 (93%) of the studies, a number of highly tr ained researchers and trained graduate and undergraduate students worked with a small, select group of students (or an individual student) and the teachersÂ’ role appeared to be very limited. It is important to insure that this process will generalize to the personnel in early childhood centers and schools so that the process can be implemented after research ers leave. Furthermore, this may be the opposite of what is currently occurring in a ty pical classroom, which is a small number of teachers with minimal training are responsib le for creating a large number of FBAs. While investigating who is completing FBAs is an important consideration, how the functions of the behavior are determined is the central issue in this literature review. Components of the Descriptive Approach In regard to the varying components of the descriptiv e approach, all fourteen studies used some combination of descri ptive methods to determine a hypothetical function of a challenging behavior. The de scriptive approach may include informant assessments such as interviews with the teacher s, teaching assistants and/or the parents, a review of permanent records, and behavior ch ecklists. In this sample of articles: 71% (n=10) used some type of interview, 7% (n=1) used a behavior checklist, 100% (n=14) used direct observation, 7% (n=1) used a sc atter plot and 7% (n=1 ) explicitly mentioned performing a records revi ew. All fourteen of the articles also engaged in some type of

PAGE 45

34 hypothesis development in which the research ers and teachers collaborated to generate hypothetical functions for challenging behavior. It should also be noted that studies that evaluated antecedents of the challenging beha vior used essentially the same methods of interviewing, behavior checkli sts, and direct observation fo r extracting information about the antecedents that precede challenging behaviors. While underused in this sample, there are tw o advantages to the use of checklists. First, unlike an interview that is open to interpretation by the i ndividual reviewing the answers, a hypothetical function is explicitly arrived at through the checklist scoring process. Second, behavior check lists are highly efficient a nd can be completed quickly without extra work. However, they do not have the advantage of direct observation of the child and all information in a behavior checklis t is filtered through a third party informant (Shores, et al., 1999). Direct observation is another descriptive approach. One of the primary methods of this descriptive approach is the use of an A-B-C procedure, in which the observer carefully records the antecedent, behavior, and consequence to es tablish a pattern of supporting environmental stimuli (Bijou et al., 1968) that either precede or follow the challenging behavior. Direct observation and/or generating an ABC (antecedentbehavior-consequence) hypothesi s was used in all fourteen studies. The importance of direct observation is two-fold. First, it prov ides basic information that can guide other descriptive and analog methodol ogies. Second, it provides the requisite baseline data to determine if subsequent inte rventions are effective. Besides the ABC analysis, a nother direct observation tool that has been widely used in research to systematically observe the antecedents and consequences of

PAGE 46

35 challenging behavior is the Functional Assessment Observation Form (FAOF) (O’Neill et al., 1997). Although more technically complex then the ABC hypothesis sheet, the FAOF allows for the observation of chains and sequences of multiple behaviors and their antecedents and consequences. Identifying ch ains of behaviors can give insight into whether the challenging behaviors often occur in a hierarchy going fr om low-intensity to high-intensity (Chandler & Da hlquist, 2002). For example, a student begins poking his neighbor to avoid participating in ‘Circle Time.’ The teacher verbally prompts the student to stop bothering his neighbor. The student ignores the redi rection and begins stomping the floor with his feet. When the teacher asks the student to get up, the student responds by screaming at the teacher and flippi ng over a nearby chair. Without careful identification of the target be havior, such as ‘participating in group activities,’ the focus may be placed solely on the high intensity beha vior, when in fact the initial behavior, avoiding ‘Circle Time,’ is a more likely target for intervention strate gies. In such cases, interventions can be designed that address the initial chal lenging behavior early in the chain before high intensity beha vior occurs. In addition, the identification of behavioral chains provides evidence that is useful in analyzing environmental events that may precipitate all challe nging behaviors. One important methodological consideration noted in Kennedy et al. (2001) and Gr andy and Peck (1997) is that these direct observations should occur across settings and at differe nt times of the day. The value of this consideration is that it provides important contextual information regarding the function of the behavior and reinforces the fact that behavior and it s functions are often contextually specific.

PAGE 47

36 A teacher interview is another prevalent indirect descriptive method. The use of a semi-structured or unstructured interview with a teacher provided important insight into the possible functions of a behavior for a numbe r of studies that used only a descriptive methodology (e.g. Grandy & Peck, 1997; Kamp s et al., 1995). The most common interview used was the Functional Assessment Interview (FAI). The Functional Assessment Interview (O’Neill et al., 1997) is a multi-step interview form designed to identify antecedents of the challenging beha vior and establish the function of a child’s challenging behavior. This is a comprehensive interview that has some empirical testing (Kinch, Lewis-Palmer, Hagan-Burke, Sugai, 2001) and is described by Gresham, Watson and Skinner (2001) as “the most comprehensiv e, up-to-date treatment of this topic” (p. 162). Some different methods of assessment such as a careful review of records (Clarke et al., 1995) and the use of a scatter plot (Symons, McDonald & Wehby, 1998) go relatively unexplored in regard to research studies. In Pack enham, Shute and Reid (2004), one interesting point in regard to the use of formal descriptive assessments was that the teacher who participated in the study said “I already do all of this (formal descriptive assessments), I just never formalized it” (p.23). In light of this consideration, the need for less formal descriptive assessments and more effective ways to access teacher knowledge could be an important step in future re search. Additionally, continuing to monitor teachers’ perceptions of the FBA process is also an important aspect of current and future research. In this sample of studies, 6 out of 14 (43%) studie s administered some type of treatment acceptability questionnaire following the study. Finding assessments that

PAGE 48

37 teachers are willing to implement is as importa nt as finding assessments that accurately identify the functions of challenging behavior. Functions Identified Accurately reporting descriptive statistic s regarding the determined function for these studies was not feasible for two reasons. First, in the largest study, Chandler et al. (1999), the function of the challenging behavior for the 210 participants was not reported. Second, other studies did not directly report fu nctions of behaviors. For example, some studies reported possible func tions by calculating conditiona l probabilities between the target behavior and the consequences (i.e ., Kennedy et al., 2001) or other outcomes as primary and secondary functions (i.e., Van Der Heyden et al., 2001). Therefore, only the most general observations can be made about patterns in this sa mple of literature regarding the three functions of behavior. This is unfortunate as evaluating base rates of behavioral functions across di fferent populations may be an important area of future research (Gresham, 2003). The functions of a target behavior that cau se the behavior to increase in frequency are typically classified into three types. Fi rst, external positive reinforcement includes attention from teacher or peers or tangible rewards (Lalli & Goh, 1993). In this sample, this was the predominate function of chal lenging behaviorlargel y teacher attention, followed by peer attention and tangible items. Second, external negative reinforcement includes escape or avoidance of an unpleasant or undesirable stimulus, such as a difficult academic task. A third type of reinforcement is automatic reinforcement, in which the reinforcement is hypothesized to be internal and biologic in nature. This type of reinforcement often explains self-injurious or stereotypic behavi or (Mace & Roberts, 1993) and was not a function of behavior for any of the participants in this sample of the

PAGE 49

38 literature. The environmental relationships surrounding a challenging behavior can be further classified in terms of whether the variable precedes the behavior, which is an antecedent, or a reinforcer that follows the behavior, which is a consequence or a function of the behavior. Choice or lack of choice was the predominate antecedent that was manipulated in the five studies that evaluated antecedents in this sample. With the exception of the predominance of teacher attention as the function of the behavior, there was no distingu ishable pattern of the fourt een studies investigated in terms of positive versus negative reinforcement or antecedent versus consequence. The studies varied in terms of their particip ants. However, some general practical and methodological observations can be made. On e notable methodological problem in the research was in regard to the functions de termined within this group of studies. These studies often generated, or at least reported, functions that were very general. For example, Schill, Kratochwill and Elliott (1998 ) describe the function of five of their participants as “attention from adults and/or peers.” (p.128). This lack of specificity is problematic for a number of r easons. First, the lack of specificity makes it difficult to determine the appropriateness of the interven tion. In other words, simply saying that attention was the function of the challenging behavior leaves the reader to wonder if access to any type of attention would amelio rate the challenging behavior or does the student require a specific interv ention? To this point, the Sc hill et al. (1998) study did not find a statistically significant differen ce between the effects of function-based interventions and non-function based interventi ons. Typically, the greater the amount of specificity in all areas of applied behavior an alysis, the more effec tively behavior can be measured and modified. Determining the f unction of a behavior should lead to the

PAGE 50

39 identification of an intervention to implement If a consequence of a challenging behavior were positive reinforcement through increased adult atten tion, then an effective intervention would most likely be decreasi ng attention for challenging behaviors and increasing attention for adaptive or pro-so cial behaviors. Ultimately, the goal of accurately establishing the functi on of a behavior is to implem ent effective interventions. Interventions Implemented All of the studies recommended inte rventions and 93% (n=13) actually implemented the interventions. Descriptive methods within these studies generated similar (but participant-specific) interventi ons and all of the inte rventions reported at least initial success at stopping the challengi ng behaviors to some degree. Examples of the interventions were increas ing teacher supervision (Kamps et al, 1995), increasing reinforcement for different behaviors (Schil l, et al., 1998; VanDerHeyden, et al., 2001), and embedding preferred activ ities into existing curriculu m (Blair, Umbreit, & Bos, 1999). A more notable aspect of these studies was the follow-up that occurred some time later. Packenham et al. (2004) reported a c ontinued reduction of the challenging behavior five weeks later due to suggestion made as a result of the functional behavior assessment. This finding is surprising, considering Reid (2000) questions whether or not teachers will persist with interventions that are gene rated by functional assessment over time. One criticism of specific articles reviewed was a lack of direct congruence between the identified function and the intervention implemented. For example, the intervention of increasing attenti on to desired behaviors and decr easing attention to challenging behaviors, (e.g. Kennedy et al. 2001; Storey, et al., 1994), while effective, did little to highlight the need for an individual FBA unl ess this intervention directly matches the antecedents and consequences identified through the FBA process surrounding the

PAGE 51

40 individual target child. A different version of the same criticism was the use of multiple interventions that may or may not address th e same function, but are implemented at the same time making it impossible to determin e what intervention was actually addressing the identified function. This lack of distinct functi on-based interventions is also seen with the insertion of a mediating intervention between the desired beha vior and the reinforcement that addressed the function of the behavior. For example, Grandy and Peck (1997) determined that the function of a challenging behavior was teache r attention. However, rather then implement an intervention that directly linked teacher attention with desired behaviors, the researchers established a token economy, whic h targeted obtaining a tangible item, but indirectly led to increased teacher attention. This lack of direct connection between the determined function and intervention implemente d calls into question the accuracy of the determined function or the potency of reinfo rcers that do not nece ssarily match function (e.g. tokens). It is impossible to determine if teacher attention was reinforcing the desired behavior or if it was another element of the token economy. Finally, the use of interventions to confirm hypothesized functions is not methodologically problematic. However, what are problematic are the conclusions that authors of these studies draw from the use of these interventions. For example, by using a reversal design in the impl ementation of a function-based intervention, Kamps et al. (1995) failed to differentiate between a func tional relationship that was established and proven effective and the actua l functional relationship that exists for a studentÂ’s challenging behavior. In other words, just because an inte rvention worked does not mean it addressed the primary function of the chal lenging behavior, it simply means that this

PAGE 52

41 intervention is powerful enough to be effectiv e even if it does not directly address the actual function of the challengi ng behavior. While this is a su btle distinction, it points to one of the cornerstones of app lied behavior analysis which is the need for precision in language and in all aspects of empirical st udies, or as Baer, Wolf and Risley (1968) described it, the technological dimension of empirical studies. Summary Descriptive assessment methods represent the vast majority of the assessments being used in FBAs. The wide range of desc riptive assessments being used singularly or in varying combinations for the preschoolers in this review also indicate that no one assessment method or combination of assessment methods has emerged as the preeminent process for young children with or at risk for EBD. There are two additional conclusions that can be drawn fr om a review of these studies. First and foremost, the issue of methods for accurately determining the function of challenging behavior remains unr esolved. There are two major obstacles to resolving the question of what will be the most effective protocol of descriptive assessments for FBAs. First, there is the continued combining of diffe rent assessment result s in a single study. In other words, studies continue to use inte rviews, checklists and direct observations without separating what pro cedures are giving consistent information with other assessments. The second obstacle is the fact that descriptive asse ssments can only yield hypothetical functions and there is no way to confirm the accuracy of any method without providing some type of environmental manipulati on. In order to determine what is best in the widespread application of FBA, it is ne cessary to determine what components are the most accurate when compared with these e nvironmental manipulations (i.e., a functional analysis).

PAGE 53

42 Second after the most accurate ways to determine function in widespread applications are determined, research s hould address the issue of function-based interventions. This must be done after the process to accurately determine function has been empirically tested, because function-ba sed interventions are unattainable without accurately determining the function first. This current research sample indicates that there is a gap between the research that has been done to accurately identify the function of a challenging behavior and the research that investigates the use of function-based interventions. Future research needs to concentrate on the implementation of recommended interventions and the rationales th at they are based on. It is imperative that researchers delineate between individualiz ed function-based interventions and the recommendation of improved classroom management skills. Researchers and practitioners must recognize th at even if a methodology is highly accurate at determining the function of a challenging behavior, but doe s not lead to highly effective interventions, than the methodology is essentially useless. This leads to the next secti on of this literature review that includes empirical studies that combined descriptive assessments with functional analysis. Descriptive Assessments and Functional and Structural Analyses within FBAs The use of experimental analysis has been widely documented over the past twenty years. While analog assessment is a freque nt design in studies in many behavioral journals, particularly the prestigious Journal of Applied Behavior Analysis it has not crossed the research to practice gap in th e widespread application of FBA in public schools (McKerchar & Thompson, 2004; Sasso et al ., 2001). Explanations for this lack of widespread application vary, but clearly the technical comp lexities and the additional time and staffing demands that are required to complete these procedures are difficult to

PAGE 54

43 overcome. Furthermore, aspects of functiona l analysis may seem counterintuitive to many practitioners, especially when conducti ng functional analysis that manipulate the environment in an effort to elicit, as opposed to prohibit, certain behaviors. Thus, there continues to be a limited amount of research in the use of functional analysis in classroom settings (Ervin et al., 2001). Illustrating this point, the research base did not tran slate to a large number of studies within this literature review. A lthough many studies have used functional or structural analyses with young children, the two criteria that eliminated the majority of the studies were the use of these analyses in natural settings and th eir use with children with high incidence disabilities. Functional analysis, particularly, continues to be used largely in clinical settings and with mo re individuals with severe developmental disabilities. In three of the studies that are reviewed, res earchers included participants with high-incidence and low in cidence disabilities (Ellis & Magee, 1999; Harding et al., 1999; Romaniuk et al., 2002); only the high-incide nce participants are discussed in this review. Ten studies met the criteria for inclus ion in this section of the literature review (see Table 2-2 for a summary of informa tion regarding each study). The following sections will outline the findings and discu ss methodological and practical limitations of the research.

PAGE 55

44 Table 2-2. Descriptive Methods and Functional Analysis for Young Children Authors/ Year of Publication Participant Characteristics (number, gender, age, label) Setting characteristics Implementer Characteristics Descriptive Methods and Functional Analysis Antecedent Functions Identified Interventions Implemented Broussard & Northrup, (1995) 3 males, 6-9 yrs, at risk for ADHD or EBD, empty classroom and primary classroom Teachers & researchers DO, HD, BFA, 3 analog settings based on descriptive analysis Teacher attention, peer attention, escape academic task Contingency reversal – No extended intervention implemented Umbreit (1996) 1 male, 5 yrs, mild MR, daycare center Teachers & researchers (only training & data collection) HD, I, DO, BFA-4 conditions Escape difficult tasks Modified curriculum, Social skills training Broussard & Northrup (1997) 4 males, 6-9 yrs, ADHD or at risk for EBD or ADHD, primary classroom Teachers & Researchers DO, HD, and BFA Peer attention Token economy, DRO Ellis & Magee (1999) 3 males, 6-10 yrs, mild autism, PDD, EBD, ADHD, separate classroom and primary classroom Researchers (separate room FAs) and teachers (in class FAs) DO, HD, FA-4 conditions (incl. peer attn., competition for teacher attn.) Peer and adult attention, escape from demands Token economy, Social skills training, task modification Harding, Wacker, Cooper, Asmus, JensenKovalan, & Grisolano (1999)* 3 males, 4 yrs., ADD, sight and hearing, at risk for EBD, primary classroom Teachers & researchers DO, SA, BFA4 analog conditions Specific v. general directions, social attention, reprimands More specific directions, increased attention for prosocial behavior Blair, Umbreit & Eck (2000) 1 male, 4 yrs., at risk for EBD, primary classroom Teachers & researchers DO, I, HD, FA5 conditions Attain tangibles Social Skills training, No other interventions Boyajian, Dupaul, Handler, Eckert, & McGoey (2001) 3 males, 4-5 yrs., at risk for ADHD, primary classroom Researchers and teachers DO, I, BFA4 conditions Attention, Tangible, Escape Contingency reversal over longer period of time (11-18 days) Doggett, Edwards, Moore, Tingstrom & Wilcski(2001) 2 males, 6 & 7 yrs. old, no identified disability, primary classroom Teachers (teachers implemented BFAs) & researchers DO, I, BFAs-2 conditions Teacher attention, peer attention No interventions

PAGE 56

45 Table 2-2. Continued Authors/ Year of Publication Participant Characteristics (number, gender, age, label) Setting characteristics Implementer Characteristics Descriptive Methods and Functional Analysis Antecedent Functions Identified Interventions Implemented Romaniuk, Miltenberger, Conyers, Jenner, Jergens & Ringenberger (2002) 7 participants, 5-10 yrs, identified for problem behaviors also some participants Mosaic Down syndrome or mental retardation, separate classroom, Researchers and therapists I, Preliminary FA, BFAs-4 conditons Choice, No Choice, Attention, Escape No interventions Stichter, Sasso & Jolivette (2004)* 1 male, 7 yrs., EBD, primary spec. ed. and gen. ed. classrooms Researchers, Teachers provide descriptive information DO, BSA3 structural variables Low social demands, high structure Integrate antecedents that decrease challenging behavior into school day Note. DO=direct observation; CL=checklist; I=interview; HD=hypothesis development; FA=functional analysis; BFA=brief functional analysis; RR=records review; SA=structural an alysis; DRO=differential reinforcement of other behaviors indicates a structural analysis Participant and Setting Characteristics The total number of participants for all te n of the studies was 28 participants or an average of 2.8 participants pe r study with a range of 1-7 part icipants. Only three of the participants in all ten of the studies were female 11% (n=3). Furthermore, while all studies included at least one participan t who met the criteria of a high-incidence disability, three of the studies included participants with low-incidence disabilities. For example, in Ellis and Magee (1999) 1 partic ipant had Attention Deficit-Hyperactivity Disorder (ADHD), but the other 2 participan ts were labeled with mild autism and pervasive developmental disorder (PDD). In Romaniuk et al. (2002) 3 of the 7 participants were diagnosed with Down syndr ome or mental retardation. In Harding, et al. (1999) 1 participant had sight and hearing disabilitie s and 1 was identified with developmental disabilities. This indicates that functional analysis may have a wider range

PAGE 57

46 of applicability with individua ls with high-incidence disabili ties than the research base reflects. In regard to practical limitations of this literature, the limited number of participants within these research studies is the first indication of the impracticality in making functional analysis an integral part of early childhood settings. The study with the most participants, Romaniuk et al., (2002) (n=7), is one of the most recent studies in this literature review. While this may suggest the possibility of widespread use of functional analysis within the classroom, there is very little other empirical research. Typically, functional analysis is time intensive and is completed on one individual child at a time. One possible explanation is that a low particip ant size is maintained as researchers place more emphasis on honing the methodology, ex ercising environmental control, or initiating novel interventions than addressing the need for widespread application. Another possibility is that f unctional analysis is only approp riate for students needing the most intensive assessments and interventions. It should be noted th at studies ten years ago (e.g., Broussard & Northup, 1995) acknow ledged the need for widespread application to more students in naturalized settings, although they continued to use very small sample sizes. The second observation in regard to particip ants and settings is that the use of analogs in primary classrooms during regul ar classroom activities indicates some modification from the controlled, structured nature of analogs to more natural, less restrictive settings. One study that clearly m odeled this assimilation is the Ellis and Magee (1999) study that conducted environmen tal manipulations both in the classroom and in a separate environment.

PAGE 58

47 Implementer Characteristics The roles of researchers and teachers t ook many forms within this sample of studies. In six (60%) of the studies, the part icipants’ teachers or other school personnel were trained and conducted the analog conditions of the functional an alysis. For three of the studies (30%), the researcher conducted the analog conditions. For the Ellis and Magee (1999) study, the researcher conducted th e analogs in the separate room and the teacher conducted them in the classroom. While researcher influence was also obs erved as a limitation for ‘descriptive assessment only’ studies, it was even more pronounced in functional and structural analysis trials. Between training teac hers and/or conducting analog conditions, researcher influence was even more evident in these studies than the studies in the previous section. This is not surprising due to the highly t echnical demands of these types of analysis. However, this is important becau se the greater the influence researchers have on the design and implementation, the more di fficult it will be to generalize the findings to teachers as implementers and determine ex ternal validity. Some studies included in the third section of this literature review have created protocols that teach teachers how to implement a functional analysis. Such cr eative methodologies and their effectiveness may be an important part of futu re research (Sasso et al., 1993). In fact, the three studies that used res earchers to implement the analog conditions introduced considerable bias considering the number of diffe rences that a researcher conducting the analog conditions may have versus the teacher conducting analog conditions. For example, they did not eval uate whether adult attention from a novel person (the researcher) is more or less power ful than adult attention from a teacher. This ignores all of the idiosyncratic and qualitat ive variables that ma y be influencing the

PAGE 59

48 occurrence for the behavior (i.e., tone of voice, appearance, person preference and so forth) that are introduced with researchers conducting the analogs. Fi nally, the issue of external validity is brought into question. Since these st udies were not designed to address long-term widespread application, it is not necessarily a criticism of these specific studies but rather a shor tcoming of the research base as a whole. The studies that used researchers to conduct the analogs were actually focused on a specific disability, such as ADHD (Boyajian et al, 2001) or the e ffects of antecedents as an environmental manipulation (Stichter, Sasso & Jolivette, 2004) and were not directed toward examining the implementation of FBA in school settings Rather these studies were geared to investigate other aspects of st ructural and functional analysis and their applicability with individuals with a specific disability or the overall va lue of the use of antecedent manipulations. The next consideration is how the actual analog pro cedure took place and what components comprised its protocol. Components of the Descriptive Approach and Analog Conditions The preliminary descriptive components were similar to the sample of studies that only used descriptive methodologies. All of th e studies (100%) used direct observations and some type of hypothesis development. Five of the studies (50%) also used a structured interview with the teacher and/ or other individuals who have contact with the participants. The only salient difference between the descriptive methods in this sample of studies and the sample of studies that only used descriptive methods is that this sample of studies was more parsimonious in its use of descriptive assessmen ts. In evaluating the descriptive assessments listed in these 10 studies, it appeared th at the descriptive assessments were only selected and used to th e extent that they c ould guide the functional or structural analysis procedures. These studies used the descriptive assessment results to

PAGE 60

49 identify times and situations to conduct an alog probes during the times or activities that teachers identified as the most problematic. For example, both Broussard and Northup (1995, 1997) only used direct observations of th e participant and a brief interview with the teacher to determine what analog conditions to conduct. In studies using functional analysis, the fi ve ‘traditional’ analog conditions were as varied as the participants with whom they were being implemented. The studies often included variations of the fo llowing conditions (1) alone or ignoring, (2) a ttention, (3) tangible reward, (4) escape or avoidance of a task and (5) free play. These are the classic conditions established by Iwata (1982/1994) and other researchers using functional analysis. Besides using two different setti ngs, classroom and separate room, Ellis and Magee (1998) described another notable variation of their analog conditions. Two alternative analog settings were implemented: (1) peer attention with the use of a confederate peer attending the target beha vior and (2) peer competition for teacher attention. These two contrive d analogs represent another step toward integrating functional analysis into more natural sett ings and provide a more accurate means of assessment using functional analysis. The results indicated that peer attention did play a role in the function of one pa rticipant’s target behavior. Romaniuk et al. (2002) provi ded another addition to the classic analogs by the testing choice or ‘control’ as a precursor of challenging behavior; this manipulation is considered a structural analys is. To test control, this st udy offered the participant the option to choose the activity in one condition and did not allow them to choose the activity in another cond ition. Not being able to choose th e activity was the setting event or antecedent of the participant’s challengi ng behaviors. The examination of choice or

PAGE 61

50 ‘control’ as a possible pre is similar to the Blair, Umbreit and Bos (1999), a study reviewed in the ‘descriptive methods only’ section of this literature review. The methodology of all of the functional an alyses studies differed from those that were originally conceived of in Iwata et al. (1982/1994); becau se they took far less time, used descriptive methods to precede the analog settings, and were subject to more of the situations that occur in natural settings. Thus while the roots of functional analysis can be traced back to Iwata et al. (1982/1994), the process has cl early evolved. This evolution also includes the use of structural analyses manipulating antecedents in much the same way that a functional analysis manipulates c onsequences. This process was used with two studies, Harding et al. ( 1999) and Stichter et al. (2004) In Harding et al. (1999), variables such as general dir ections versus specific direc tions and group activity versus independent activity were manipulated. For Sticht er et al. (2004), variables such as noise, structure, and levels of socialization were manipulated. Whether investigating antecedents or consequences, the important issue in these studies is determining the environmental factors that are related to the challenging beha vior so that the functions may be accurately identified. Functions Identified As noted in Broussard and Northup (1995), the most prevalent functions typically identified, in general, are teacher atten tion, peer attention, and escape from academic demands. An overview of this sample of studies confirms this assertion and, similar to the sample of studies that used descriptive asse ssments only, it is impossible to descriptively summarize the most prevalent functions using de scriptive statistics. However, it can be observed that no other pattern of identified functions emerge d. In other words, there was not a clear difference in the number of partic ipants whose function was either positively

PAGE 62

51 reinforced (i.e., attention or tangible) or nega tive reinforced (i.e., escape). Furthermore, the explanation for why it is so difficult to make summative statements is a notable strength of this sample of studies. Namel y, the high degree of specificity of function makes summative statements across participants difficult. A difference between this group of studies and the ‘descr iptive methods only’ studies is that the functions that functional analysis determines are more spec ifically and precisely defined. Proponents of functional analysis view the lack of speci ficity with descriptive assessments as problematic. They assert that as functions are more gene rally defined, the intervention becomes more idiosyncratic and less effective (Sasso et al., 1992). Ha rding et al. (1999) noted that in their comparison of the two methodologies, the descriptive methodology produced results that were “difficult to in terpret” (p. 329) and “brief experimental analyses were needed to clar ify functional relations” (p.330). They did note that this was a result of not being able to observe the target behavior during descriptive analysis. Another direction of the research within th is sample of studies is the combining of structural analysis and f unctional analysis. As Romaniuk et al. (2002) points out, combinations of antecedents and establis hed functions have not been extensively investigated through research methods. In this study, the antecedent of choice or control over tasks had different effects determined by whether the function of the challenging behavior was attention or escape. This st udy indicated that choice could serve as a powerful antecedent when the function of a ch allenging behavior was escape. Harding et al. (1999) used a similar design of measuri ng both antecedents and consequences. The next logical step after the accu rate identification of the functions of challenging behavior

PAGE 63

52 is the implementation of effective interventi ons that use the information gathered through these methods. Interventions Implemented Multiple types of interventions were used with these studies. Functional analysis methodology creates a complex issue when eval uating the interventions used with these studies. Within the ten articles reviewed, fo llowing the functional and structural analysis, an environmental manipulation was conducted that resulted in re ducing the challenging behavior using a contingency reversal. There were also a variety of other interventions that occurred following the functional analysis Five studies (50%) reported reducing the occurrence of challenging behavi ors during the intervention phase also referred to as the contingency reversal phase. However, the other 5 studies used other interventions including differential reinforcement of othe r behaviors (DRO), and implementing token economies. It should be noted that Boya jian et al. (2001) reported a continued reduction of the challenging behavior six months later due to implementing c ontingency reversal elements into the milieu; thus, the contingency reversal became the intervention that was implemented over time. Contingency reversal involves establishing an environment that increases the rate of challenging behavi or and implementing a phase of the opposite condition. For example, because a challengi ng behavior increases during a ‘no choice, high academic demand’ condition, the reversal contingency would be ‘participant choice and low academic demand.’ While this reversal accomplished the desired result, it may not be a situation that is appropriate across situations in the typical preschool environment. The reason this situation may not be appropriate to generalize to typical preschool environments is that at some point it is necessary to give children tasks that are

PAGE 64

53 not preferred and to as k children to participate in activ ities that are not preferred. It should also be noted that even if an ant ecedent based intervention is implemented that decreases the probability of the challenging behavior, one will still have to implement an intervention that directly addresses the occu rrence of the challenging behavior at some point. The fact that a reversal c ontingency may not be appropr iate for implementation in an early childhood environment illustrates a practi cal limitation in studies with this type of design. As Blair, Umbreit and Eck (2000) indicate, the reversal of the problematic contingency was not feasible in preschool se ttings, because it is not always possible to put children in the situations that will reduce challenging behavior s (i.e., always making sure there were a positive ratio of available toys to peers). In the natural setting of a preschool classroom, there will be times when there are not enough toys for everyone, or when access to a preferred toy may not be possi ble because another child is playing with it. Further, insuring that this does not ev er happen to the student does not allow the participant to gain appropriate adaptive or so cial skills. It may be more appropriate to increase the amount of social skill instruction that prec edes a functional behavior assessment to alleviate its use and then to ha ve specific interventions in place that address the challenging behavior s after they occur. A notable aspect of these studies was the fo llow-up that occurred some time later. Harding et al. (1999), Boya jian et al. (2001), and Ellis and Magee (1999) reported continued decreases in challeng ing behaviors for the participants as they followed up to six months later. The durability of these pres cribed interventions is encouraging. As FBA research continues, more and more examples of effective interventions will be devised

PAGE 65

54 and illustrated, allowing practitioners to ma ke direct connections between the identified function and the intervention that is implemented. Summary The use of functional and stru ctural analysis with young chil dren with or at risk for EBD remains an important area of future rese arch. Clearly, the resear ch is still in the process of evolving from its setting in clinics with individuals with severe disabilities to settings with children in classrooms with m ild disabilities. However, it is difficult to determine what role the methodical manipula tion of the environment will play in the widespread application of functional anal ysis methodology in FBAs. Thus far, this methodology has not become a part of the pr otocol in the formulation of FBAs in schools. The concept of teachers deliberatel y manipulating the environment and tracking the increase or decrease of cha llenging behavior may never be feasible. In this sample of studies, functional analysis was used as one more method, albeit typically the most decisive and trustworthy method, to triangu late assessment methods and accurately determine the function of challenging behavior. There are three salient points that can be drawn from this sample of the literature. First, similar to the first set of studies, no study with functional analysis procedures could be conducted with large numbers of children in the same setting, thus there continues to be small sample sizes and lit tle replication has occurred. Researcher involvement and highly specifi c but varied methodologies make it difficult to foresee what role functional analysis will play in school and early chil dhood settings based on this sample of literature. The issue of data collection is also an important consideration. In all of these studies, functional analys is was a multi-person job and, while it was encouraging to see teachers taking an active role in conducting the analogs, it was still

PAGE 66

55 necessary to have trained observers count and re cord the occurrences of the behavior. It is impossible to know whether early childhood centers and schools will be able to dedicate the appropriate amount of personnel for training and data collection th at is necessary to both conduct the analog conditions but also to record the occurrences of behaviors. To the second point, if the interventions are not feasible or lack social validity, this dedication of time, personnel, and resources to functional analysis may not even be desirable. Second, in each of these studies, func tional analysis procedures depended on descriptive assessments and discerning the re sults from the different descriptive methods and the functional analysis were impossibl e. This leads to the question ‘Was the functional analysis, as part of an FBA prot ocol, necessary if descriptive assessments arrived at the same function?’ Additionally, a core principl e of functional analysis is environmental control (Sasso et al., 2001) and functional analysis is only as effective as the amount of control that can be obtained. Th e difficulty is that classrooms by nature are free operant environments, especially comp ared with clinics (Scott et al. 2004) and establishing necessary environm ental control may be difficult for educators with limited time and resources. For this reason, functional analysis may not be appropriate as one more component of a FBA protocol. It is importa nt for future research to continue to use more classroom settings and to increase the am ount of teacher influence to evaluate the role of functional analysis in classroom protocols. Third, while functional analysis can devel op more specific functions that lead to more effective interventions (Harding et al., 1999), they can also l ead to interventions that are not practical or a ppropriate for implementation in the natural setting (Blair,

PAGE 67

56 Umbreit & Eck, 2000). In consideration of thes e points, the ultimate role of functional analysis may be as an effective method to evaluate the accuracy of less obtrusive descriptive methodologies, however it is difficult to see its widespread adoption as a component of a standard FBA protocol. Unlike the studies reviewed in this section that combined functional analysis and descriptiv e methodologies as a pr otocol to determine function, the next section will separately ev aluate different descriptive methodologies using functional analysis as the stan dard to determine their accuracy. Direct Evaluation of Descriptive Methodologies A search for articles that directly ev aluated different desc riptive methodologies indicates a dearth of research in this area. Only eight studies were identified that evaluated the reliability of different de scriptive methodologies when comparing the results with each other, such as compari ng the results of a di rect observation and a teacher interview. Furthermore, only six out of these eight evaluated the validity of these descriptive methodologies by co mparing them to a functiona l analysis. Murdock, et al., (2005) and Ellingson et al. (2000) only co mpared different types of descriptive assessments with each other. This paucity in the research is particularly evident with children who are diagnosed or at risk for EBD (Gresham, 2003; Gresham et al. 2004; Fox, Conroy, Heckaman, 1998; Sasso et al., 2001). Due to the limited research in this area, it was necessary to e xpand the exclusion criteria be yond young children with EBD to include all ability and age levels of childre n. In fact, all but one of the studies included participants with more severe developmenta l disabilities and the age range was expanded to even include young adults, some of whom we re evaluated in residential and vocational settings. Additionally, the criterion regarding the age of the research was also expanded from ten years to 15 years.

PAGE 68

57 Finally, unlike the previous two bodies of literature, the interventions were not evaluated within this section of the study, because the focus of this review is to determine how accurately functions are identified via various descriptive methodologies. Evaluating function-based interventions will be an important area of future research but that area is premature until the methods for accurately iden tifying functions have been evaluated. Further, the rationale for omitting this aspect of the research study was to provide greater focus on the methodologies used to evaluate the assessments rather than evaluating the resulting interventions; therefore, the focus of this section is to examine the portion of the study that attempted to determine a primar y function of the ch allenging behavior. Function-based interventions will be an im portant area of future research; however, research methods must first empirically determine whether the assessment methods can accurately arrive at the functi on of the challenging behavior. All other criteria that were established for this review of the literature were maintained. One important criterion that was maintained between this body of research and the other two areas of research was limitin g the specific target behaviors to exclude research geared toward highly specific be haviors that are not typical of most young children who demonstrate challe nging behavior (e.g. self-inj urious behavior, stereotypy, hair-twirling, eye-poking, br eath-holding) were eliminate d. Therefore, studies that compared descriptive assessments with func tional analyses for bi zarre speech (Mace & Lalli, 1991) stereotypy (Crawford, Brockel, Sc hauss, & Miltenberger, 1992; Paclawskyj, Matson, Rush, Smalls & Vollmer, 2001) or self-i njurious behavior (Belfiore, Browder & Lin, 1993; Lerman & Iwata, 1993; Paclawskyj et al., 2001) for the majority of the participants were excluded.

PAGE 69

58 Because the different assessment methods we re evaluated separately and, in some cases, compared with the results of a functiona l analysis, these eight studies are unlike the other studies that used desc riptive assessments and func tional analysis in the FBA process. The studies in the previous two sections simply used various descriptive assessment methods and assimilated the result s to develop a hypothetical function and/or guide a functional analysis. The studies that evaluated the different descriptive assessment methods listed the results of each method separately and reported agreement or disagreement between the methods. Because these studies more closely resemble the study being proposed, these eight studies will be presented in a st udy-by-study basis, in reverse chronological order, presenting the most recent st udy first (see Table 2-3 for a summary of information regard ing each of these studies). In Murdock, et al., (2005), ei ght junior high school student s (7 classified as EBD, 1 classified as LD) were evaluated using thr ee descriptive assessment methods. The three methods were a teacher team interview, a student interview and classroom observations. No functional analysis procedures were used. First, a team of teachers who worked with an individual student was assembled and the fi rst author used a quest ionnaire to guide the interview process regarding specific challe nging behaviors and the environmental factors that surrounded the behaviors. The responses were condensed into summary statements that included the challenging behaviors, a nd both antecedents and hypothetical functions.

PAGE 70

59Table 2-3. Direct Evaluations of Descriptive Assessments Studies Participants: Age, Gender, Disability Setting Who completed the FBA? What descriptive assessments were used? Agreement between descriptive assessments How was the functional analysis completed? Agreement between FA and descriptive assessments Murdock, OÂ’Neill & Cunningham (2005) -8 junior high school age males -7 classified EBD -1 classified LD General ed and resource classrooms -Teacher Team completed questionnaire -Student questionnaire -Researchers and paraprofessional conducted direct observations Teacher FAI, Student-FAI and FAOF 9/ 14 (64%) agreement on all three 13/ 14 (93%) agreement between teacher FAI and FAOF N/A N/A Newcomer & Lewis (2004) -9 years, Male, OHI -11 years, Male, at-risk -11 years, Female, at risk General education classrooms Researchers administered FAI and PBQ to teachers and students. Researchers completed direct observations and FAs FAI (S-FAI), PBQ, Scatterplot, ABC Assessments were consistent except for PBQ results with 2/3 participants Based on descriptive assessments, each participant placed in one indicated and one control condition Agreement between descriptive assessments and functional analysis for 2/3 participants Cunningham & OÂ’Neill (2000) -3 years, Male, Autism -5 years, Male, Autism -5 years Male, Autism Self-contained classrooms Researchers administered descriptive assessments to teacher and classroom assistant. Teacher completed the FAO. Researchers completed FAs. MAS, FAI, FAO, Informal question Good consistency across methods and participants. Used rank orders for hypothetical functions Each participant placed in 5 analog conditions in a separate school multi-purpose room Rank order indicates good agreement of primary function between descriptive assessments and FA for 2/3 participants

PAGE 71

60Table 2-3. Continued Studies Participants: Age, Gender, Disability Setting Who completed the FBA? What descriptive assessments were used? Agreement between descriptive assessments How was the functional analysis completed? Agreement between FA and descriptive assessments Ellingson, Miltenberger et al. (2000) -19 years, Female, Severe M.R. -18 years, Male, AngelmanÂ’s syndrome & profound M.R. -12 years Male, Severe M.R. Self-contained classrooms Researchers administered questionnaire to teachers and teachers completed questionnaire independently. Conducted direct observations Questionnaire, Interview, and ABC Questionnaires and interviews consistent with each other. Results from ABC were not consistent with questionnaire results N/A N/A Yarbrough & Carr (2000) -18.33 years Male, Autism & Profound M.R. -11.5 years. Female, Autism & Profound M.R. -12.4 years, Female DeLange Syndrome & Moderate M.R. Self-contatined classrooms, vocational and community placements Researchers administered adapted FAI to teachers and classroom assistants and conducted functional analysis FAI followed by a Likert scale rating the resulting hypothesized functions. FAI also generated high and low likelihood antecedents N/A Based on FAI information high and low likelihood antecedents manipulated in natural settings and one hypothesized function is implemented across antecedents Agreement between FAI and FAs consistent for high likelihood but not for low likelihood antecedents

PAGE 72

61Table 2-3. Continued Studies Participants: Age, Gender, Disability Setting Who completed the FBA? What descriptive assessments were used? Agreement between descriptive assessments How was the functional analysis completed? Agreement between FA and descriptive assessments Calloway & Simpson (1998) -3 years, Male, Developmental delays 4 years, Male, Speech & Cognitive delays -4 years, Male, Speech & Language, Chronic misbehavior Self-contained classrooms Researchers conducted FAs and administered descriptive assessment to teacher and paraeducators Informal ranking process of 4 possible functions of misbehavior, N/A Based on preliminary assessments each participant placed in 1-3 analog conditions. Analysis occurred in small work room or isolated classroom Minimal agreement between informal assessment and FA. Only one paraeducator for one participant agreed with the FA results for primary functions Arndorfer, Miltenberger, Woster, Rortvedt, & Gaffeney (1994) -3 years, Female, Speech delays -13 years, Male, Down syndrome -5 years, Male, Autism & Mild MR -2 years, Male, FAS -5 years, Male, Bipolar disorder ParticipantsÂ’ homes -Researchers administered MAS and FAI to parents. Researchers completed A-B-C direct observation assessments. Researchers directed parents in FA conditions MAS, FAI, ABC -MAS consistent with FAI for 2/5* -FAI and A-B-C consistent 5/5. MAS and A-B-C consistent 2/5 Each participant was placed in two analog conditions either escape or attention** Consistent agreement between ABC, FAI and FA. Inconsistent agreement for MAS and FA.

PAGE 73

62Table 2-3. Continued Studies Participants: Age, Gender, Disability Setting Who completed the FBA? What descriptive assessments were used? Agreement between descriptive assessments How was the functional analysis completed? Agreement between FA and descriptive assessments Sasso et al. (1992) -7.5 years, Female, Autism -13 years, Male, Autism Self-contained classrooms Teachers completed all aspects of FBA ABC direct observation N/A Researcher conducted formal FA. Each teacher conducted 5 analog conditions in classroom Consistent agreement between ABC results and FAs *For the other 3 participants, the second function on the MAS was the same as the primary function for the FAI. **One participant was placed in an interrupted play versus no n-interrupted play condition.

PAGE 74

63 The team scored the accuracy of the stat ements on a four-point Likert scale and then ranked the summary statements from leas t to most problematic for the student in the classroom. This final step of scoring summary statements for accuracy was developed in another study included in this review, Yarbrough and Carr (2000). Second, the first author conducted the same interview process with each individual st udent. Third, the first author and a paraprofessional conducted direct classroom observations using the Functional Assessment Observation Form (FAO F). Using the results from the teacher team interview as a guide, students were obs erved for 20-30 minutes, 1-3 times per week for three weeks.This resulted in a total range of 5-15 hours of direct observation. Interobserver agreement was conducted for 19% of the sessions and averag ed 80% agreement. Participating teachers also completed a soci al validity scale to assess the three-step process. There were four possible hypotheti cal functions (attenti on, escape, tangible and self-stimulation) that the three assessment methods could identify. However, tangible and self-stimulation were almost never id entified in teacher and student summary statements. Results of the study indicated agreement of a hypothesi zed function for all three methods 9/14 (64%) of the time when eith er escape or attention was identified. When there was not agreement across all thre e methods, the teacher team interview and the direct observations still agreed four out of the other five times. Escape from a task or activity was the most prevalent function identified. There are four conclusions that can be drawn from this study. First, these investigators achieved a moderate level of agreement across descriptive assessments. Second, although there was agreem ent across the results of different assessment methods (including a measure of confidence from par ticipating teachers and students), the validity

PAGE 75

64 of the results remains unconfirmed because no experimental procedure occurred. Thus, the results are correlational in nature. Without an experi mentally rigorous functional analysis, the consistency across instruments doe s not necessarily indicate validity of the function. In other words, all of the descrip tive assessment may agree with each other that the function is attention, but this may not actua lly be the function. This is especially true if the same individual is completing all of the descriptive assessments. Third, the nature of the participants limited the possible hypotheti cal functions that could be determined. It is unlikely that in this setti ng junior high school ag ed students with disabilities such as LD and EBD would engage in self-stimulation or have acquiring tangibles as the function of their behavior. In other words, this sa mple of students was unlikely to engage in stereotypy and also unlikely to argue over toys (although other ta ngibles could cause challenging behavior). By limiting possible f unctions to attention and escape, this increased the possibility for agreement acr oss measures. Fourth, the findings of the descriptive assessments were not kept separate in this study, the direct observations were informed by the teacher team interviews “the behaviors listed on the observation form came directly from those listed on the teacher interview form.” (p.10). This makes the level of agreement between the teacher in terviews and the direct observations less compelling. In Newcomer and Lewis (2004), three elemen tary-aged students (one classified as Other Health Impaired, two not classified with any specific disability) were first evaluated using five descriptive assessme nt methods followed by a brief functional analysis was conducted. The five assessment me thods were a teacher interview, a student interview, a teacher checklist, a scatter pl ot, and classroom observations. The functional

PAGE 76

65 analysis procedures presented a limited numbe r of conditions (atten tion v. no attention, preferred v. non-preferred peer groups, easy ta sk v. hard task). These conditions were guided by the descriptive assessment results a nd while the article refers to all three experimental procedures as functional analys es, the manipulations of peer group and task difficulty is more aptly described as a struct ural analysis, because it is a manipulation of antecedents rather than consequences. The five descriptive assessments were presented in the same order for all participants. First, the teachers were inte rviewed using an adapted version of the Functional Assessment Interview (OÂ’Neill et al., 1997). After a hypothesis regarding function was formulated using the interview re sults, teachers were asked to complete the checklist, the Problem Behavior Questionnaire (PBQ), and a scatter plot for a minimum of five days. It should be noted that unlik e the other assessment methods used in this study, the scatter plot does not lead to a hypot hetical function, it is a direct observational tool that identifies patterns between occurrences of the behavior and times of day and/ or activities (Touchette, et al., 1985). Targeted students were then interviewed using the student version of the FAI. Finally, the student was observed using the Antecedent Behavior Consequence (A-B-C) sheet. Th e observation periods were determined by results from the other descriptive asse ssments. The amount of time and number of observations for this descriptive assessmen t were not reported. These descriptive assessment results were assimilated to iden tify a primary hypothetical function that was tested through expe rimental analysis. The experimental analysis was constr ucted using a singlecase, alternating treatment design. The analyses occurred in natural classroom settings during both

PAGE 77

66 academic and low structure tasks and targeted individually identified challenging behaviors. Each child was exposed to 10-minute analog probes featuring the hypothesized function condition, ac cording to the descriptive assessments and a control or contra-indicated conditi on. There was a range of 13-22 sessions over 27 days. The results of the experimental an alysis showed clear differentia tion of rates of challenging behavior across conditions for 2 of the 3 pa rticipants, supporting the summative results of the descriptive assessments for these 2 partic ipants. There was only slight differentiation between the conditions for the third particip ant offering limited suppor t to the result of the descriptive assessment. The results of the descriptive assessment s were also compared with each other. Newcomer and Lewis state that all of the descriptive assessments were consistent with two exceptions. For one student, the teacher interview and the checklist indicated attention was the primary function, but al l of the other assessments including the experimental analysis indi cated that escape was the pr imary function. Similarly, for another participant, the checklist identified attention as the primary function when all other assessments supported escap e as the primary function. It is notable that neither of these 2 participants was the same individua l that had limited support from the functional analysis. Therefore, all 3 pa rticipants had some type of assessment result that was inconsistent with other assessment results. There are three conclusions th at can be drawn regarding th is research study. First, although there is some evidence to support th e reliability and valid ity of descriptive assessments, in that they had fairly cons istent results that were supported by an experimental procedure, the f unction of behavior in this ca se does appear contextually

PAGE 78

67 relative and somewhat fluid. Second, the use of descriptive assessm ents as a ‘package’ with a specific order and one assessment informing the next one does not allow for scrutiny of the individual assessments and doe s make the results vulnerable to an order effect. Third, by limiting the functional anal ysis conditions to those suggested by the descriptive assessments, no other functions were experimentally tested. This means that if additional function of the beha vior exists, but was not indi cated by the majority of the descriptive assessments, then it was not evaluated experimentally. In Cunningham and O’Neill (2000), three pr eschool aged children with autism were evaluated using four descriptive a ssessments and an experimental functional analysis. The four descriptive assessments were: an informal question to the teacher and classroom assistant (e.g. “Why do you think Dan bites?”), a teacher interview (the Functional Assessment Interview), a behavi or checklist (the Motivation Assessment Scale), and descriptive observa tions (the Functional Assessm ent Observation Form). The functional analysis was then conducted repl icating the procedure outlined in Iwata, Dorsey, Slifer, Bauman and Richman (1994/1982) presenting 5 analog conditions: demand (escape), attention, tangible, alone and free play. The procedure for this study consisted of conducting the indirect descriptive assessments (the informal question, the interv iew, and the behavior checklist) with both the teacher and the classroom a ssistant after class hours. These were conducted separately for each teacher/classroom assistant dyad to de termine if the two adults present in the classroom had similar perceptions of the challenging behavior and its antecedents and consequences. The MAS scores were then co mbined to give an average score for the primary function. The primary classroom teacher then completed the direct observations

PAGE 79

68 for two hours a day for 4-5 days, resulting in a total of 10-12 hours of direct observation for each participating student. Inter-observ er agreement was conducted for 20% of these sessions and ranged from 80%-100% agreemen t. The design for the experimental functional analysis was a single subject, multi-element design with 15-minute sessions being conducted in separate rooms that were made available to the researchers, e.g. a multi-purpose room adjacent to the gymnasium. The results for the various different descriptive assessments were presented as a rank ordering of hypothetical functions of the target behavior. For the descriptive assessments, there was perfect agreement for the informal question and the interview for all dyads of teachers and assistants. The othe r assessments also had fairly good agreement in which all the assessments varied only so mewhat, for example, the MAS listed escape as the primary function and tangible as seconda ry and the FAI listed tangible as primary and escape as secondary. For the functional anal ysis, 2 of the 3 participants had little differentiation between tangible and escape. Howe ver, this mixed result actually confirms the descriptive results that had similar mi xed primary functions across the different assessments. There are three conclusions regarding this research study. First, soliciting information from both the teacher and the classroom assistant determined that the two adults present in the room had similar per ceptions regarding behavi oral functions (i.e., inter-rater reliab ility). Second, the use of ranking the functions for each student and each assessment allowed for students whose challeng ing behavior may have dual functions but could have complex results. Subsequently comp lex interventions that may be difficult to implement in classrooms are the end product. Th ird, one aspect of the functional analysis

PAGE 80

69 that may make its results less compelling wa s the low rate of challenging behavior demonstrated across all analog conditions. But because the greatest rate for the students, means of 10%, 10% and 9%, was consistent with the descriptive assessment results, the functional analysis appears accurate. In Yarbrough and Carr (2000) 3 participan ts between the ages of 11 and 18 years old were evaluated. Two particip ants were classified autistic and the third participant was diagnosed with DeLange syndrome. There was only one descriptive assessment used. An adapted version of a Functiona l Assessment Interview was us ed with two professionals who had had up to two weeks of contact with the participating student. This descriptive assessment was followed by a brief functional analysis. The brief functional analysis manipulated both antecedents and consequen ces that had been identified by the descriptive assessment results. The FAI was adapted in three ways. First, after information had been gathered using the standard FAI questions, summary statements were generated for each participant that outlined setting events, ante cedents and consequences. Each professional who was interviewed then verified each st atement for accuracy and rated each summary statement on a 7-point Likert scale for likelihood of the situation occurring. Next, the second professional would then evaluate a nd agree or disagree with these hypothesized summary statements. Lastly, these same steps would be taken with the roles of the two professionals reversed. If statements were generated that the s econd professional could not confirm witnessing, then a third professiona l, such as another classroom assistant, would be asked to confirm the statemen t and its likelihood rating. Following this

PAGE 81

70 descriptive assessment proce ss, a brief functional analysis was conducted using the results. In the functional analysis, a reversal design was used in which the different hypothesized situations were compared with control conditions that regulated the presence of the hypothesized contributory vari ables. The 3 participan ts were placed in seven situations while one or two different hypothetical functions were tested in each condition. For example, the first participant wa s placed in a situation in which he had to clean a table. The hypothetical function was that his challenging beha vior of self-injury was maintained by escaping difficult tasks. A control condition consisted of assigning the participant to a task he found easier, obj ect identification. The frequency of the challenging behavior in these two conditions was then compared. Each participant was placed in two or three high and low-likeli hood situations with one or two hypothetical functions of the challenging behavior being tested for each participant. When the information gathered through th e descriptive assessments was tested using an experimental functional analysis, a cl ear pattern emerged. In those situations that had a high-likelihood of eliciting problem behaviors, the inform ants accurately identified the surrounding environmental variables. For the situations that had a low likelihood of eliciting challenging behaviors, the functional analysis did not validate the informantsÂ’ hypotheses of the antecedents a nd the functions of the challe nging behaviors. The authors offer four interrelated factors that may explain this pattern: (1) infrequent opportunities to observe the low-likelihood situations; (2) stim ulus complexity; (3) pe rceived strength of reinforcement; (4) the role of setting events These factors should be interpreted with caution, as they are not experime ntally validated. In terms of evaluating the efficacy and

PAGE 82

71 efficiency of the descriptive assessment method (i.e., the FAI adaptation), it can be concluded that this method is effective for identifying the function for certain frequent behavior problems, but limited in detecti ng more subtle variables in low-likelihood situations. Unfortunately, the authors did not conduct a treatment acceptability measure, so it cannot be determined if this benefit is worth the time and energy investment for the participating professionals. In Ellingson, et al., (2000) 3 participants, aged 19, 18 and 12, were evaluated using two different descriptive assessments. No e xperimental functional analysis procedures were used. The participants were all classified as either se verely or profoundly mentally retarded and one participant was also diagnos ed with AngelmanÂ’s syndrome. All of the assessments occurred in special education classrooms in public schools. A second phase of this study investigated f unction-based interventions, but th is phase of the research is outside the scope of th is literature review. The two descriptive assessments were each used in two different ways. The first descriptive assessment was an interview that was presented as both an interview with the first author asking follow-up questions and as a questionnaire that the teacher completed independently. The second descriptive assessmen t was a direct assessment that used the Antecedent-Behavior-Consequence (ABC) worksheet. These worksheets were simultaneously completed by both teachers a nd experts trained in Applied Behavior Analysis. The assessments were always presented in the same order: (1) the teacher completed the questionnaire independently, (2) the teacher was interviewed using the same questions as the questionnaire by the fi rst author except follo w-up questions were also asked, (3) using the information provide d by the interviews a nd questionnaires an

PAGE 83

72 ABC checklist was formulated and direct obs ervations with the ABC instrument were conducted at the same time by both the classr oom teacher and a trained expert. Two other trained behavior analysts conducted relia bility checks on the hypothetical function determined for the interview and the questionnaire. The results of the descriptive assessme nts were determined by the authorsÂ’ evaluations of the interview/questionnaire and the ABC checklist results. All of the hypothetical functions were determined to be either attention or escape. For one participant, all of the different methods of assessment, including the second rater who supplied reliability data, agreed that th e function of the challenging behavior was attention. For the second participant there was agreement th at the function of the behavior was attention for all of the assessments excep t for the second rater of the questionnaire. For the third participant there was agreement from the indirect assessments; the interview and questionnaire agreed that the function of the problem beha vior was both attention and escape. All of the direct obser vation results determined that the function of the problem behavior was attention. There are three conclusions regarding this research. First, the agreement between the questionnaire, the interview and the di rect observations are encouraging but in consideration for the amount of influen ce the questionnaire a nd interview had over structuring the direct observation, the agreem ent is less compelling. The construction of an ABC checklist using indirect assessments requires further testing, especially against a less structured ABC observation sheet and functional analysis. Second, the agreement between the questionnaire and the interview is a step toward making the process more efficient, (i.e., it only requires one person ra ther than two people); however, this result

PAGE 84

73 may be idiosyncratic to the skills of the interviewer. Some interviewers may elicit valuable information through follow-up que stions that a ques tionnaire completed independently could never draw out. Third, it is unclear wh ether the possible functions were in some way limited to the choices of attention or escape, or if this was simply a characteristic of these three specific participants. In Calloway and Simpson (1998), three pres chool aged children were evaluated using two descriptive assessments, an experi mental analysis and an informal question posed to teachers and para-educators. The re sult of the descriptive assessment and the experimental analysis were then compared with the responses to the informal question. The 3 participants were male, ages 4, 4 and 3 and three were in an inclusive community early childhood special education classroom Two participants were in the same classroom and one was in a different classroom; all assessments took place in the classroom setting. Two children had speech a nd language and cognitive delays and one participant was diagnosed with developmental delays and su spected autism. All three had been referred to this study for repeated demonstration of various problem behaviors. The procedure for this study occurred in th ree phases. The first phase consisted of descriptive classroom observations and co mpletion of the MAS. The researchers interpreted this information and generate d a hypothetical functi on. Manipulating through analog conditions tested this hypothesized f unction. A reversal design was used with a contra-indicated activity to compare behavioral fre quencies. These reversals took place every two or five minutes for either one 10 or 20 minute session. The second phase involved implementing an intervention direc tly related to this function. This was implemented not only to reduce the problem be haviors for the individual child, but also

PAGE 85

74 to provide further validation of the functi on of the behavior as determined by the descriptive assessment a nd functional analysis. In the third phase of the study, three classroom professi onals, either the teacher and/or the classroom para-educat ors, were asked to identify a hypothetical function of the problem behavior. They had no knowledge of the results of the formal analyses. The classroom professionals were asked to identify the behavi oral function using a ranking system of 1-4 for these choi ces: (a) avoidance, (b) attention, (c) se lf-expression, and (d) power/control. According to the formal analyses, on e student’s function was confirmed as attention, the second student’ s function was escape and th e third student’s results indicated that the problem behavior was maintained by multiple functions: attention, power/control and escape. The results of the ra nking system indicated that the classroom professional agreed with each other six out of nine opportunities (66%) in identifying the primary function. However when comparing the fo rmal and informal analysis results, the classroom professionals only agreed with th e formal analysis one out of nine times (11%). There are three salient aspects about this research study. First, the use of ‘selfexpression’ and ‘power/control’ as possible functions is divergent from most current literature regarding functions of problem behavior. Self -expression was the function chosen five out of nine times by the teach ers and paraprofessionals, and it was not confirmed by any of the formal analyses. Fu rther, the functions of self-expression and power/control are not even options on the MA S; therefore, a direct comparison between the interview and the MAS for some particip ants was not possible. Second, agreement

PAGE 86

75 between the analog conditions a nd the descriptive assessments is difficult to determine since the analogs directly a pplied the results of the descriptive assessments, making it impossible to separately evaluate them. Third, in regard to the results of the informal analysis, this point to a la rger consideration, namely th at consistency does not equal accuracy. The professionals were consistent with each other 66% of the time, but they were only accurate in the sense that they ag reed with the functional analysis 11% of the time. In Arndorfer, Miltenberger, Woster, Ro rvedt and Gaffeney (1994), five children ages 2 to 13 years old were evaluated usi ng three descriptive a ssessments and a brief experimental analysis. Unlike the other studies included in this section of the literature review, all of the assessments occurred in the participantsÂ’ homes. Parents served as the source of information for the indirect assessmen ts and also conducted direct assessments. The 5 participants had been diagnosed with different disabilities including: (1) Down syndrome, (2) fetal alcohol syndrome, (3) autism and mild ment al retardation, (4) developmental speech and cognitive delays and (5) bipolar disorder. All of the participants were referred to this study due to repeated demonstrations of problem behaviors. There were two phases of this research study. The first phase consisted of the parents completing the Motivation Assessment Scale (MAS) two times with one to two week delays between administrations. Repeated administrations were done to evaluate reliability over time. Before the MAS was scored, the first author conducted the FAI with the parents and the first and second authors independently reviewed the responses and generated hypothetical functions. The MAS was then scored to also determine a

PAGE 87

76 hypothetical function. Then, a third researcher and the parents of the participant, who were unaware of the indirect assessment results, conducted three to five 30-60 minute direct observations using the FAOF. The pa rents received a brief training session on conducting a direct observation using the FAOF A scatter-plot assessment determined specific times target behaviors were likely to occur and this guided the times the direct observations were conducted. Inter-observer agreement was conducted for 33% of all direct descriptive assessments. Using the results of direct obs ervations, hypothetical functions of the problem be havior were formulated. The second phase consisted of experime ntally manipulating antecedents and consequences that had been identified th rough the descriptive a ssessments. A reversal design consisting of high-likelihood anteceden ts and consequences and low-likelihood antecedents and consequences was conducted. For example, a participant whose descriptive assessments indicated that attent ion was the primary function of her behavior had a session where her mother looked at, talked to, or touc hed her at least twice every minute and all problem behaviors were ignore d. In the next condition, the mother looked at, talked to, or touched the participantÂ’s si blings at least twice every minute but gave contingent attention based on the particip ant exhibiting problem behaviors. These conditions were presented in 15-30 minute sessions on the same day for three of the children and over two to three days for the ot her two children. Only one session of each phase was conducted for each child in an effort to make the experimental analysis as brief, and socially valid, as possible. The result of the descriptive agreem ent indicated agreement 16 out of 20 opportunities (four descriptive asse ssment results for 5 participants). For three out of the

PAGE 88

77 four disagreements, the MAS was the instru ment with divergent results. Additionally, agreement between the two administrations of the MAS indicated agreement for individual questions ranging from 56% to 75% (M=64%). The experimental analysis supported the results of the interview and th e direct observations for all five of the participants (100%). There was differentiation in the frequency of the problem behaviors throughout the reversal of the conditions. There are three conclusions re garding this research study. First, this study bolsters the use of the FAI and the direct observations to generate accurate hypothetical functions of behavior, but draws the valid ity of the results of the MA S into question. Second, while the experimental analysis does offer support for the descriptive assessments, because it only tested one function, it is not possible to know if other functions of the behavior exist without testing multiple functions. Third, by only conducting sessions of each condition on one day for 3 of the 5 partic ipants, there are contextual f actors, and setting events that may not be accounted for and may impact the function of the problem behavior. It seems possible that this study may have sacrificed a degree of accuracy in an effort to increase efficiency and social validity of the experimental process. In Sasso et al. (1992), 2 participants, ages 7 and 13 years old, were evaluated using a conventional functional anal ysis, a direct observation de scriptive assessment and a teacher conducted functional analysis. Both par ticipants were classified as autistic and the descriptive assessment and teacher conducted functional analysis occurred in selfcontained special education cl assrooms. The conventional f unctional analysis occurred outside the classroom setting. The direct observation assessment was the ABC worksheet. However, it was completed duri ng specific activities that are similar to the conditions of

PAGE 89

78 the functional analysis conditions (i.e., pres entation of a difficult task). The functional analyses consisted of a multi-element desi gn based on Iwata et al (1982/1994) featuring four conditions: alone, escape, attention, a nd free play and also including a tangible condition. There were three phases of this study. Fi rst, researchers cond ucted a conventional functional analysis. Second, the first teach er was trained to complete an ABC observation. Then, the teacher conducted four 15-minute observations of the student. The environment was varied for the observations and included solitary play, a high demand, low teacher attention condition and a low de mand and high teacher attention condition. After the direct descriptive observations we re completed, a brief functional analysis within the classroom was conducted. The clas sroom teacher modified the environment to test the five analog conditions The same procedure occurred for the second participantthe only difference was that the first teacher, ra ther then the researcher, taught the second teacher to conduct the descriptive assessment and the classroom based functional analysis. The results of this study indicated that in all three methods of assessment, escape and tangible were the f unctions of the problem behavi or for one participant and escape was the function of the problem behavior for the other particip ant. Thus, there was 100% agreement for both participants. There are three conclusions th at can be drawn from this research study. First, this research article demonstrates flexibility in the functional analysis process, in that the same procedures can be app lied across different settings and produce similar results. Second, this study suggests that functional analysis procedures may have broader applicability for classroom teachers. This r uns counter to the concept that functional

PAGE 90

79 analysis procedures are too co mplex to have broad-based ap peal. Third, this idea that teachers can effectively train teachers sugge sts of the possibility of a much broader training model. There are two conclusions that can be dr awn from a summative evaluation of these eight studies that directly compared descrip tive assessments with ea ch other and/or some type of functional analysis. First, no desc riptive assessment met hod has emerged as the preeminent method to accurately determine a hypothetical function. It appears that for indirect assessments, teacher in terviews, both formal and informal, play an important role in identifying potential functions in seven out of eight of these studies. Their overall accuracy is further suggesti on of the conclusion drawn ear lier that research should continue to focus on the knowledge teachers al ready have about their studentsÂ’ behaviors and effectively and efficien tly hone methodologies to acce ss this knowledge. For direct descriptive assessments, ABC assessment sheets completed by both teachers and researchers do reveal accurate antecedents and functions. Ho wever, this conclusion is tainted by the second conc lusion about the overall state of this research. Namely, that all of the methods, descriptive and experimental were not conducted independently of each other with a high degree of control. This second conclusion, that the assimilati on of results for all of the methodologies muddled the final results, is pervasive throughou t all eight of the studies. This is true both for the descriptive assessm ents and for the link between the descriptive assessments and the functional analyses. Th e end result is that no conc lusions can be drawn about what single descriptive methodology is the most accurate. This occurred for a variety of reasons. First, some studies used a ranking system between raters that allowed for

PAGE 91

80 multiple functions to be enumerated within the results. The second and most influential reason is that the descriptive assessments were used to guide the functional analysis. By not maintaining separation of the descriptiv e assessments from the functional analysis, the functional analysis can only verify or refute the hypothetical function that the descriptive assessments rev eal. Conducted independently, functional analysis can establish the function of the challenging behavior (Iwata et al., 1982/1994). In future research, these results could then be take n back and compared with the independent descriptive assessments to determine whic h ones accurately arrive d at the function. In summary, the overall lack of comparat ive studies that evaluated descriptive assessments suggests that descriptive methods have not been validated by research. If descriptive assessments, such as rating scales, interviews, and hypothesis development using antecedent, behavior, consequence (ABC) analysis (Bijou, et al., 1968) indicate an inability to accurately determine the function of a young ch ild’s disruptive behavior, then their use is ill-advised. However, if descrip tive methods adequately arrive at the same function as experimental analysis, then th e use of these simpler, less time-consuming methods has been supported. Regardless of the outcome of the project, it is incumbent upon researchers to prepare school and early childhood center personne l with the best, experimentally validated procedures. Comprehensive Summary of Research Literature “Whether FBA can be effectively and e fficiently accomplished outside a research setting remains to be seen”(Reid & Nels on, 2002, p.23). As Reid and Nelson suggest, the use of FBA as a durable, highly effective strategy in widespre ad applications in applied settings remains to be seen. A review of 32 studies that used FBA, predominately with young children with high incidence disabiliti es in natural setti ngs, indicates three

PAGE 92

81 conclusions that can be drawn from the current st ate of the research. Firs t, in regard to the participants and settings, th e use of functional behavior assessment is still clearly evolving in terms of its use with large numb ers of young students with mild disabilities in classroom settings. Second, interventions ge nerated by FBA, while effective in the studies in this sample, remain idiosyncrati c and, at times, the relationship between the derived function and the function-based inte rvention could be described as tenuous at best. The third and the most important factor is that in terms of the FBA process the individual components that would make up a universal FBA protocol have not been established as valid and reliable for the wi despread application of functional behavior assessment. The first conclusion in regard to the research in this literature review concerns the participants and settings of FBA. Two considerations regard ing the practical limitations of all of these studies are the types of disabilities that are being addressed and their use of large numbers of students. A third ancillary cons ideration is the type of behaviors that are being addressed within the FBA literature. Wh at may be attributed to IDEA (1997) was this use of FBAs for children with disabiliti es other than developmental disabilities. In 1999, Nelson, Roberts, Mathur and Rutherford estimated that 80%-90% of functional assessment research had been conducted with children with devel opmental disabilities. The trend, however, to apply it to students with other disabilities or no disability at all is proliferating according to this sample of studi es. This is encouraging and indicates that FBA is a highly flexible process with multiple levels of applications. Combined with the problem of determining the ability to generate large numbers of FBAs, the problem of researcher influence is an important consider ation. The question of

PAGE 93

82 ‘how well do these studies indicate what will occur in generalized application after the study?’ remains. This threat to external va lidity points to an im portant area of future research. There is a need for further resear ch in this area that uses longitudinal methodology providing initial teac her training on the use of descriptive assessments, a series of temporally determined probes to evaluate implementation fidelity, and a final measure of effectiveness. Studies using a de sign such as this will allow researchers to minimize the impact that they are having on teacher implementation of the FBA process. This is a problematic shortcoming of the res earch that has not been answered to date. A third point regarding this fi rst conclusion is in regard to what types of behaviors research has addressed with FBA me thodology. Studies clearly addressed a predominance of externalizing behaviors. While this is not surprising, as teachers are more likely to identify externalizing behavi ors as problematic, it does point to another shortcoming in literature thus far. This is the question ‘can FBA procedures effectively address problematic internalizing behaviors?’ Internalizing behaviors, social withdrawal, isolation, and flat affect can be as problem atic for students as externalizing behaviors (Chandler & Dahlquist 2002). However, current resear ch has not addressed whether these behaviors can be effectively addre ssed through functional behavior assessment. The second conclusion is in regard to function-based interventions. While some interventions led to highly specific interventions directly related to the function of the challenging behavior, other interv entions were only indirectly related to th e function or were not related at all. Researchers and teachers implemented a number of effective interventions, manipulating both antecedents and consequences to reduce challenging behaviors; however, one pract ical and methodological limitation is evident from this

PAGE 94

83 sample of literature. In regard to methodology, even studies that were directly linked to the intervention received far less emphasis than the process that arrive d at the function of the behavior. Empirical research has not pr oduced a ‘menu’ of validated interventions that directly match functions. However, the overall lack of emphasi s on interventions is surprising as the ultimate goal of FBAs is th e formulation of effective intervention and essentially FBA methodology will be judged on its effectiveness. In regard to practical limitations, a limitation that this sample of studies did not address is the question of teachers being able to sustain prescribed, function-based interventions for a number of students over time. This will be an important area of future research as widespread application of function-based in terventions continues to broa den into different types of classrooms and different popul ations of students. The third and most important conclusion regarding this li terature is the lack of validity and reliability studies that directly evaluate the components that make up the FBA process. The reason this conclusion is th e most important is that before widespread application for many children in applied setti ngs with a variety of behaviors can occur, the process must be honed so that it is not only reliable and valid, but also accessible to minimally trained school personnel and highly efficient. Similarly, although effective function-based interventions are paramount to the process, it will be impossible to derive those without reliable and valid components. Un fortunately, even in the third section of this literature review, a section that was co mprised of studies that directly evaluated descriptive assessments, useful evaluation of the components did not occur. There was still a lack of separation of the findings be tween the individual descriptive assessments and the descriptive assessments and the functiona l analysis that was often used to validate

PAGE 95

84 the hypothetical functions. As a result in th ese research studies, it was very difficult to determine what components were most effective. This leads to an important area of resear ch namely, conducting a series of indirect and direct descriptive assessments and comparing the outcomes of these assessments to the outcomes of a functional analysis methodol ogy. However, to rectify limitations in the existing research, independence of the descri ptive assessment results and the functional analysis results is important to determine if the descriptive instruments are able to obtain valid functions independent of the functiona l analyses. Additionally, the results of the various direct and indirect de scriptive assessments would need to be kept separate and counterbalanced across participan ts to maintain control for carry over and order effects. These assessments should occur in natural se ttings with considera tion of total time of procedures and necessary levels of traini ng with the goal of maximizing efficiency without sacrificing va lidity. While this lite rature review has in dicated a number of problems with the widespread application of FB A in natural settings, it is vital that the methodologies be evaluated before issues of implementation, teacher training, and effective interventions are addressed. The re sults of this research are an empirically validated step towards establishing a universal protocol that can be applied consistently both in continuing FBA resear ch and in applied settings. Research Questions The following questions were addressed in this study: 1. How do the functions of target behavior s identified through various descriptive functional assessments compare to each other? 2. How do the functions of target behavior s identified through various descriptive functional assessments compare with f unctions identified through a functional analysis for young childrenÂ’s challenging behavior?

PAGE 96

85 CHAPTER 3 METHODS This chapter outlines the methodology used to conduct this study. The goals of this study were to compare the functions determin ed by three descriptive assessments with each other and with the functions determ ined by functional analyses (FA) for young children who demonstrate challenging behavior s. Teachers perceptions of the process and its results were also evaluated. This chapter presents the following information: (a) participants, (b) settings, (c) materials, (d ) measurement procedures including dependent measures, the independent variables, data collection and type of data recording, and interobserver agreement. This chapter then outlines the experimental procedures including recruitment and training, preexperimental assessments, descriptive assessments, and functional analysis. Finally, the methods of measuring social validity, procedural integrity, and the design and data analysis are presented. Participants Four teacher-child dyads participated in this study. In accordance with the policies set forth by the University of Florida In stitutional Review Board, the principal investigator obtained informed consent from the participating teachers and caregivers of participating children (see Appendix A for a pproved IRB forms and participant informed consent forms). All participating teachers a nd the child participants parents received a $25 dollar gift card. Child participants were f our children who met th e following criteria. Children were between the ages of 3 and 6 years old.

PAGE 97

86 They were currently enrolled in a Day Ca re Center, Pre-kinder garten, Kindergarten, or First Grade class (general education or special education) in North Central Florida. They had been identified through teache r nominations, in that teachers have reported that these children are having ch allenging behaviors that are interfering with their learning and this behavior warrants an FBA. They had a Total Problem Score of no less then 60 on the Child Behavior Checklist (CBCL-TRF 1.5-5 or 6-18 y ears old) (Achenbach, 1992). They had consistent attenda nce at school, defined as 80% attendance or 4 out of 5 days a week, documented through school records. Teacher reports indicated average cognitive functioning. Teacher participants were four indivi duals who met the following criteria: They taught children ages 3-6 years old. They had at least one child in the cl assroom who met the criteria above. They held at least an Associate of Arts degree, They served as the child participantÂ’s primary teacher for at least one month. The flow chart of the teacher and child pa rticipant recruiting process is outlined in Appendix B. The following is a brief descri ption of all four participant dyads. Ramon Ramon is a 4-year old Hispanic male. His primary teacher, Rose, is a Hispanic female. She has thirty years of e xperience in early chil dhood care settings; she has had 7 months of contact with the partic ipating child, and holds a Bachelor of Arts degree. His teacher reported average cogniti ve functioning. The teacher report form of the Child Behavior Checklist (CBCL-TRF 1.5-5) (Achenbach, 1992) indicated that Ramon has a total score on the Total Problem scale of 110 (>97 percentile) with subscores in the clinically significant ra nge for emotionally reactive behavior, anxious/depressed, withdrawn, at tention problems and aggr essive behavior. RamonÂ’s challenging behaviors were disruptive vocali zations defined as raising his voice above

PAGE 98

87 the classroom volume level, asking more than three questions in a 10-second interval or using profanity or “bathroom talk.” Anthony Anthony is a 5-year old African-A merican male. His teacher reported that Anthony has average cognitive functio ning. His primary teacher, Terry, is a white female. She has 4 years of experience in early childhood care se ttings, has known the participating child for 6 months and has co mpleted her Associate of Arts degree. The teacher report form of the Child Behavi or Checklist (CBCL-TRF 1.5-5) (Achenbach, 1992) indicated that Anthony has a total scor e on the Total Problem Scale of 88 (>97 percentile) with sub-scores in the clinical ly significant range for emotionally reactive behavior, sleep problems, and aggressive behavior. Anthony’s ch allenging behaviors were crying, yelling “Mom” or “Mama”, rolling around on the ground, raising his voice above the classroom volume level and leaving his assigned area (defined as being more than 6 feet away from his assigned area). Jimmy Jimmy is a 5-year old African-Am erican male. His teacher reported average cognitive functioning. His primary t eacher, Amy, is a white female. She has 25 years of experience in an ear ly childhood care setting, has known the participant for 7 months and holds a Bachelor of Arts degree. The teacher report form of the Child Behavior Checklist (CBCL-TRF 1.5-5 year s old) (Achenbach, 1992) indicated that Jimmy has a total score on the Total Problem S cale of 61 (87 percentile) with sub-scores in the borderline range for aggressive behavior. Jimmy’s challenging behaviors were disruption defined as verbal defiance screaming, teasing, sa ying ‘No’, crying or whining. Greg Greg is a 6-year old African-Ameri can male. Teacher reports indicated average cognitive functioning. His primary teach er, Alison, is a Hispanic female. She has

PAGE 99

88 1 year of experience in the classroom; sh e has had 7 months of contact with the participating child and has a Bachelor of Arts degree. The teacher report form of the Child Behavior Checklist (CBCL-TRF 6-18) (A chenbach, 1992) indicated that Greg has a total score on the Total Probl em Scale of 95 (>98 percentil e) with sub-scores in the borderline range for social problems and rule-b reaking behavior and in the clinical range for thought problems and aggressive behavi or. GregÂ’s challenging behaviors were disruptive noises defined as tapping his penc il, whistling and raising his voice above the classroom volume level. Child and teacher in formation are summarized in Tables 3-1 and 3-2. All CBCL sub-scores are reported in Appendix C. Table 3-1. Child Participant Information Child Participant Age Race Gender CBCL-TRF Total Score Ramon 4.9 Hispanic Male 110 (97th percentile) Anthony 5.0 African-American Male 88 (>97th percentile) Jimmy 5.5 African-American Male 61 (87th percentile) Greg 6.4 African-American Male 95 (>98th percentile) Table 3-2. Teacher Participant Information Teacher Participant Child participant Years of Experience Race Months of Contact with Child Current Level of Education Rose Ramon 30 years Hispanic 7 months Bachelor of Arts Terry Anthony 4 years White 6 months Associate of Arts Amy Jimmy 25 years White 7 months Bachelor of Arts Alison Greg 1 year Hispanic 7 months Bachelor of Arts

PAGE 100

89 Settings The settings for all of the assessment methods were the primary classrooms; a primary classroom setting was defined as the room in which the child participant spends the majority of his school day. For the firs t 3 participants listed (Ramon, Jimmy and Anthony), the setting was the 4 and 5-year-old classroom of a unive rsity-sponsored early childhood care center. The fourth participantÂ’ s (Greg) primary setting was at a public elementary school multi-grade, self-containe d, special education classroom. The preexperimental assessments and teacher training occurred in the primary classroom with the teacher during non-student contact times, (i .e., planning periods, nap times or after dismissal). The indirect descriptive met hods (interview and checklists were also conducted in the primary classroom of the individual child during non-student contact times, (i.e., planning periods, or naps). The direct descriptive observations occu rred during different classroom activities with the majority of the observations in th e child participantÂ’s pr imary classroom setting. For the three participants in the early childhood center, th e activities were primarily Centers Time and Free Play. For the student in Kindergarten participant (Greg), the activity was mainly independent work time. However, observations also occurred on the playground for all 4 participants and in the cafeteria for Greg. The functional analyses also occurred in the primary classroom setting. For the functional analysis, an area within the primary classroom was designated where othe r children were not located, e.g. an area in the corner of the room. Ta ble 3-3 lists summaries of th e phases, the assessments and the settings in which they took place.

PAGE 101

90 Table 3-3. Summary of Assessments and Their Settings For Each Phase Phases Assessments Settings (1) Pre-experimental Assessments Pre-experimental assessment interview, Child Behavior Checklist (CBCL-TRF) Primary classroom during non-student contact times Indirect Descriptive Assessments: Motivation Assessment Scale (MAS), and Functional Assessment Interview (FAI) Primary classroom during non-student contact times (2) Descriptive Assessments (Indirect Descriptive Methods & Direct Descriptive Observations Direct Descriptive Observation: Antecedent-Behavior-Consequence (ABC) Analysis Primary classroom, playground, and cafeteria (for Greg) (3)Functional Analyses Functional Anal ysis Conditions Primary classroom where other children were not located Materials During the pre-experimental assessmen t phase, the Child Behavior ChecklistTeacher Reporting Form (CBCL-TRF 1.5-5 y ears old or 6-18 years old) (Achenbach, 1992) was used as a preliminary screening a ssessment to evaluate the prevalence of challenging behaviors. Additionally, a pre-expe rimental assessment interview sheet was created to access information that provided information needed for the descriptive assessments and the functional analysis pha ses (the interview guide is included in Appendix DInstruments). In the functional analysis phase, additional materials were required. The tangible condition required a toy or an obj ect that is preferential to th e participant. This toy was individually determined by asking the primar y teacher “What is (the child’s) favorite thing to play with?” duri ng the pre-experimental assessment phase. Ramon’s teacher identified Play-Doh as his pr eferred item for the tangible condition. Anthony’s teacher identified Spider-Man action figures as hi s preferred item. Jimmy ’s teacher identified Legos as his preferred item. Greg ’s teacher identified a miniat ure football as his preferred

PAGE 102

91 item for the tangible condition. Additionally, materials were necessary for the task in the escape condition. The task was defined as a task that can be completed with approximately a 30 % success rate. A handwriting task (writing his name 10 times in a row) was used for the escape condition in the functional analysis fo r all 4 participants. This was determined by listing sample activit ies and asking the teacher to select an activity during the pre-experime ntal assessment phase. For the attention, free play and ignore conditions, a neutral activity, neither preferred nor non-preferred, was necessary. The primary teacher chose this from a list of examples of activities discussed during the pre-experimental assessment phase. All four teachers identified coloring with crayons in a coloring book as a neutral activity. Table 3-4 provides a summary of all materials necessary for functiona l analysis conditions. Table 3-4. Summary of Child ParticipantsÂ’ Preferred Items, Neutral Activities and Difficult Tasks Child Participant Preferred Item Neutral Activity Difficult Task Ramon Play-Doh Coloring with crayons in a coloring book Printing name 10 times in a row Jimmy Spider-Man Action Figures Coloring with crayons in a coloring book Printing name 10 times in a row Anthony Legos Coloring with crayons in a coloring book Printing name 10 times in a row Greg Miniature Football Coloring with crayons in a coloring book Printing name 10 times in a row Additional materials that were required for data collection in the functional analysis phase are: (a) digital stopwatc h to monitor the duration of the descriptive observation and functional analysis sessions, (b) a three-ring notebook that co ntained a laminated sign of

PAGE 103

92 each of the scripted verbal prompts and (c) a Panasonic video camcorder to record each session of the functional analyses. Measurement Procedures Dependent Measures The dependent measures were challenging be haviors that were defined individually for each participant. Similar to previous rese arch that has directly evaluated descriptive assessment methods, a number of different behaviors were colla psed under the label challenging behavior (Arndorfe r et al., 1994; Newcomer & Lewis, 2004; Sasso et al., 1992). Occurrences of any of the behaviors we re then treated as an occurrence of the challenging behavior and count ed together. The same individual dependent measures were evaluated across all assessment met hods. Ramon’s challenging behaviors were disruptive vocalizations defined as raising his voice above the classroom volume level, asking more than three questions in a 10-sec ond interval or using profanity or “bathroom talk.” Anthony’s challenging behaviors were crying, yelling “Mom” or “Mama,” rolling around on the ground, raising his voice above the classroom volume level and leaving his assigned area (defined as being more than 6 feet away from his assigned area). Jimmy’s challenging behaviors were di sruption defined as verbal defiance screaming, teasing, saying ‘No,’ crying or whining. Greg’s chal lenging behaviors were disruptive noises defined as tapping his pencil, whistling, a nd raising his voice above the classroom volume level. Each of these behaviors that was considered a ch allenging behavior was assessed using a frequency count. Independent Variables The independent variables in this study we re the different methods of assessing the function of behavior. Three different functi onal assessments instruments were used for

PAGE 104

93 the descriptive assessment procedure: the Functional Assessment Interview (FAI) (O’Neill et al., 1997, the Motiv ation Assessment Scale (MAS) (Durand & Crimmins, 1992), and an Antecedent Behavior Conseque nce worksheet (ABC) (Bijou, et al., 1968) (see Appendix D for copies of all instrument s not protected under copyright). A fourth indirect descriptive assessment, the Problem Behavior Questionnaire (PBQ) (Lewis, Scott & Sugai, 1994) was originally included as another 15-item Likert scale questionnaire used for determining the function of a ch allenging behavior. However, once data collection and analysis was completed, it becam e apparent that a comparison between the PBQ and the other descriptive assessments was not logical. The PBQ does not offer the same options for the hypothetical function of a challenging beha vior as the other descriptive assessments do because it lacks tangible and sensory as possible functions and only includes attention, escape and setting ev ents. This discordant scoring taxonomy made direct comparisons with the other me thods of assessment impossible; therefore, although the PBQ was administered to all partic ipants, the results ar e not compared with the other descriptive assessments or presented in Chapter 4. The Functional Assessment Inte rview (FAI) (O’Neill et al., 1997) is a multi-step structured interview form designed to iden tify the function of a child’s challenging behavior; and Gresham, Watson, and Skinne r (2001) describes it as “the most comprehensive, up-to-date treatment of this topic” (p. 162). It takes approximately 45 to 90 minutes to complete, and it is divided in to 11 sections. The fi rst section obtains descriptions of the behavi ors asking about the topogr aphy, frequency, duration and intensity. The first section also asks for predictable sequences or “chains” of behaviors. The second section elicits information about ecological events (set ting events) that may

PAGE 105

94 predict the occurrence of the behaviors; this includes medical issues and a description of the daily schedule that indicates enjoyable and problematic activities. The third section of the interview asks about specific antecedents that immediately preced e the occurrence of the behaviors including times of day, people, activities, etc. The f ourth section of the interview identifies the outcomes or consequenc es that may be maintaining the behaviors (i.e., the function of the behavior). This sec tion includes questions a bout the behavior, the particular situation, what the individual gets and what the individual avoids as a result of the behavior. The fifth section asks the interv iewee to rate the overa ll efficiency of the problem behavior on a 5-point Li kert scale (1= low efficiency and 5= high efficiency) in terms of effort expended, frequency of reward as a result of the behavior and the time delay between the behavior and the rewar d. The sixth section of the FAI asks for alternative (replacement) behaviors that may already be in the part icipant’s repertoire. The seventh section inquires about communication ab ilities of the individual. The eighth section obtains ancillary information regard ing things one should do and should avoid when working with the individual. The ninth section asks for various items and activities that the participant enjoys or finds reinforc ing. The tenth section elicits information about past interventions that have been attempted and their results. The eleventh section has the interviewer develop su mmary st atements regarding distal setting events, immediate antecedents, the problem behavior and the maintaining consequence. The checklist employed was the Motivation Assessment Scale (MAS) (Durand & Crimmins, 1992). The MAS was designed for determining the hypothetical function of problem behaviors and is a frequently used checklist. “The MAS has been the most extensively evaluated psychometric instrume nt developed for functional assessment”

PAGE 106

95 (Sturmey, 1994, p.295). However, the studies of the MAS have revealed mixed results in regard to reliability and cont ent validity (Floyd et al., 2005). This instrument uses a 6point Likert scale with 16 items and asks the rater to evaluate the frequency of a specific behavior in various situa tions (0= Never, 1=Almost Never, 5= Almost Always, 6=Always). The direct observation method was the Antecedent, Behavior, Consequence (ABC) Analysis. The Antecedent Behavior Consequence (ABC) analysis is based on the procedure described by Bijou et al. (1968). In the ABC analysis, a running record of observed events is created. For example, if th e teacher asked the chil d participant to put away the toys, and the child participant yelled at the teacher and the teacher allowed the child to play with the toys for 5 more minutes, this would be coded as a request (antecedent), yelling (behavior) and tangibl e (consequence) (Newcomer & Lewis, 2004; Sasso et al., 1992). This process has been used in numerous studies (for a review, see Heckaman et al., 2000; Lane, Umbreit, Beeb e-Frankenberger, 1999; Scott et al, 2004). Data Collection Procedures After the recruitment of participants wa s completed, the principal investigator conducted individual pre-experi mental assessment interviews with each of the teacher participants. Following the pre-experimental assessment phase, a research assistant (Research Assistant 1) conducte d the FAI, the MAS and the PB Q. All research assistants were second and third year doc toral students with at least 2 years of previous experience as classroom teachers or school counselors; thei r training for this study is detailed in the recruitment and training section of experime ntal procedures. The indirect descriptive assessments were counterbalanced to control for an order effect. Research Assistant 1 completed the FAI with the primary teacher by asking each question and writing down all of the responses the teacher gave on the FA I interview form. Following completion of the

PAGE 107

96 FAI, the same research assistant dete rmined a hypothetical function based on the teacherÂ’s responses to the interview by writing down one of the four options of possible functions (attention, tangible, escape, or sens ory). The same research assistant also had the teacher complete the MAS and the PBQ. The research assistant did not read these items out loud to the teacher, but if the teach er had a question about any of the items then the research assistant clar ified the terminology of the questionnaire. Because the questionnaires could be completed before the FA I (due to counterbalancing), the research assistant did not tally or revi ew the questionnaire responses. This insured that the results of one instrument did not influence his judgment of hypothetical function based on the FAI. Regardless of the order in which they were presented, Research Assistant 1 always scored the FAI immediately after its comp letion due to its more subjective scoring method, and the MAS and PBQ were scored by th e principal investigat or after all other assessments were completed. A second research assistant, Research Assistant 2, conducted the ABC direct observation descriptive assessments for each pa rticipant for eight 15-minute intervals. In this process, the observer recorded the occu rrence of the challenging behavior (B), the antecedent (A) events that precede the beha vior and the events that followed the behavior, the consequences(C ). A running record of AB C direct observations was conducted by recording each occurrence of the identified problem behavior, its antecedents and its consequences. ABC r ecording occurred across each 15-minute assessment condition. The research assistant then chose a hypothetical function of the challenging behavior, from the four possible functions, base d on the observations he had done.

PAGE 108

97 The participating teachers conducted the functional analysis conditions with assistance from the principal i nvestigator. All sessions were videotaped. The primary investigator then reviewed the definition of the challenging behavior for the individual child and viewed each videotaped session. Th e rate of the challenging behavior per minute was calculated for each condition by watching the videotape, and using a paper and pencil, making a tally for each occurren ce of the challenging behavior for each condition. After all other assessments were complete, the teacher completed the social validity questionnaire and then was interviewed by th e principal investigat or using the social validity interview sheet. The principal investigator transc ribed the verbal responses of the teacher during the interview. The teacher completed the questionnaire without knowledge of the results of the descriptive assessment and functional analysis and then answered the first nine questions of the in terview before reviewing the re sults. This questionnaire and interview sheet can be found in Appendix E. Interobserver Agreement Interobserver agreement (IOA) was conducted for both the ABC direct observations and the functional analyses. For the direct observations (ABC), two trained observers watched 12/32 (38%), or 3 out of 8 of the same 15-minute intervals for all 4 participants. Both observers maintained a frequency count of occurrences of the challenging behavior. Agreement for occurren ce of behaviors was computed by dividing the smaller total occurrences of behavior by th e larger total occurren ces of behavior and multiplying by 100 (Sasso, et. al, 1992). For th e 4 participants, interobserver agreement ranged form 60%-100% with an average of 93%. For Ramon, agreement between the two observers ranged from 90%-100% with an average of 97%. For Anthony, agreement

PAGE 109

98 between the two observers ranged from 83%100% with an average of 94%. For Jimmy, agreement between the two observers ranged from 60%-100% with an average of 87%. For Greg interobserver agreement was 100%. Additionally, both observers were asked to separately formulate a hypothetical function of the challenging behavior based on their observations. They agreed on the hypothetical function of the behavior 3/4 (75%) times, the disagreement of hypotheti cal function was for Anthony in which the primary observer (Research Assistant 2) concl uded tangible was the primary function, and the reliability observer (Research Assistant 3) concluded attention was the primary function. Research Assistant 2 based the hypothe tical function on eight 15-mi nute observation sessions and Research Assistant 3 based the hypothetical function on three of the same 15-minute observations. For each participant, the two observers wrote down a hypothetical function after their last observation session and placed them in sealed envelopes. The principal investigator calculated the number of agreemen ts and disagreements after all of the other assessments were completed. For the functional analysis, the agreement percentages were computed on a sessionby-session basis for an average of 37% of the sessions using the following formuladividing the smaller total occu rrences of behavior by the la rger total occurrences of behavior and multiplying by 100 (Sasso et al., 1992). For Ramon, interobserver agreement was calculated for 8/21 (38%) cond itions. The interobserver agreement ranged from 75%-100% with an average of 95%. For Anthony, interobserver agreement was calculated for 8/19 (42%) condi tions. Interobserver agreem ent ranged from 80%-100% with an average of 98%. For Jimmy, inter observer agreement was calculated for 13/42 (31%) conditions. The inter observer agreement ranged from 66%-100% with an average

PAGE 110

99 of 97%. For Greg, interobser ver agreement was calculated for 7/18 (38%) sessions. The interobserver agreement ranged from 80%-100% with an average of 97%. Experimental Procedures Recruitment and Training Child participants were identified through teacher nominations. Teachers completed the age appropriate version of CBCL-TRF (complete sub-scale results are listed in Appendix C) on the child to determ ine eligibility for inclusion. Teachers also participated in an individual one-hour work shop to establish a baseline of requisite knowledge about FBA. Prior to training, all t eacher and parent consents were completed. All training sessions for particip ating teachers, research assi stants, and the IOA collector occurred prior to implementing any of the assessments. Participating teachers Each participating teacher took part in an individual onehour training session to revi ew principles of functional behavior assessment and functional analysis (See Appendix F for outline of teacher training session). This training session included watching a video of a sample f unctional analysis that was created by the principal investigator and a discussion with the principa l investigator of staffing adjustments that were necessary when the te acher was involved with functional analysis implementation. Research Assistants Research Assistant 1Â’s training on indirect descriptive assessments consisted of reviewing challe nging behaviors, revi ewing the order of presentation for the indirect assessments and reviewing protocol inst ructions and scoring that accompanied each of the commercially di stributed descriptive a ssessments. Research Assistants 2 and 3Â’s training consisted of reviewing the method for completing the ABC worksheet and reviewing operational defini tions of challenging behaviors. Practice

PAGE 111

100 sessions were also held for Research Assistan ts 2 and 3, in which they viewed three 15minute sessions of preliminary videotapes of the child participants until they reached 80% agreement in the completi on of the ABC worksheets. IOA data collector The IOA data collector is a cer tified special education teacher with five years of experience working with children with emotional and behavioral disorders. The IOA data collectorÂ’s traini ng consisted of a review of each of the conditions of the FA session and a discussion of each participantÂ’s operationally defined challenging behavior. The principal investig ator and the IOA data collector viewed videotaped sessions of the func tional analyses until they re ached 80% agreement for three consecutive sessions. Phase I--Pre-experimental Assessments After participants (i.e., children and teachers) had been identified, the following preliminary procedures were completed. Thes e preliminary procedures included: (1) a review of cumulative school records to obtain demographic information, (2) establishing an operational definition of the challenging be havior for each of the child participants, and (3) establishing a pr eferential toy for the tangible FA condition, a difficult task for the escape condition, and a neut ral activity for free play, atte ntion and ignore. The target behavior or challenging behavi or, as defined previously, we re the behaviors that the teacher identified as the most frequently o ccurring, problematic behaviors of the child with the exception of Jimmy. For Jimmy, the first behavior identified was aggression toward peers. However, ethics would prohibi t providing contingent reinforcement for the demonstration of aggression toward peers, such as in the functional analysis. Therefore, continued discussion with the teacher led to identification of other behaviors that were suggested to lead to aggre ssion including verbal defian ce, screaming, teasing, saying

PAGE 112

101 ‘No’, crying or whining. These behaviors were combined to become the operational definition of challenging behavior for Jimmy. To establish neutral and difficult activities, the principal investigator listed possible choi ces of activities and pr esented samples of a variety of worksheets that might be cons idered aversive to young children (i.e., a handwriting worksheet with the name listed 10 times, letter identification, multiple digit adding and subtracting). Phase II--Descriptive Assessments In this phase the identified primary teacher filled out the Motivation Assessment Scale, the Problem Behavior Questionnaire, and was interviewed using the Functional Assessment Interview by Research Assistant 1. Indirect assessments for each participant were conducted in a single 2-hour block. With in 2 weeks of the i ndirect assessments, Research Assistant 2 conducted eight, fifteen-m inute direct observati ons to complete the ABC observation form. All observations occu rred within the classroom, playground and cafeteria using the spec ifications previously outlined. The order of the indirect descriptive assessments were counterbalanced to cont rol for an order eff ect and the direct observations occurred before the functional analysis for Ramon and Jimmy and after the functional analysis for Anthony and Greg to co ntrol for any potential effect the functional analysis may have had on the occurrence of th e challenging behaviors in natural settings. Phase III--Functional Analysis In the next phase, the principal investigat or and the participating teacher completed the functional analysis (FA), without any know ledge of the results of the de scriptive assessments. The functional analysis occu rred in the primary classroom and the participant teacher acted as the therapist delivering the appropriate consequences. The principal investigator monitored the time in tervals, and provided prompts for the delivery

PAGE 113

102 of the appropriate consequence for each dem onstration of the challenging behavior. For the prompts, the principal investigator used laminated cards with the teacherÂ’s scripted verbal responses printed on them. Five conditions for the functional analysis phase of this proj ect were used: (1) tangible, (2) attention, (3) escape/ avoidance, (4) ignore, and (5) free play. There was a 1 to 2-minute break between each condition and the order of the conditions was randomly selected to control for order effects. For each condition, the principal investigator held up the laminated card with the verbal prompt for the teacher to say either upon every demonstration of the challenging behavior and/or every thirty seconds when noncontingent attention statements were requi red (i.e., escape and tangible). Before the attention condition of each se ssion, the principal investigator modeled the delivery of the attention condition statement (i.e., placing a ha nd on the childÂ’s back); and the principal investigator modeled appear ing busy with paperwork during the attention condition. Before the tangible condition, the principa l investigator modeled presenting and effectively removing the tangible item. For all sessions the principal investigator offered correctives between conditions when necessa ry and modeled the ot her non-contingent attention statements. See Appendix G for a description of the functional analysis procedures. With the exception of JimmyÂ’s FA, if othe r children in the cl assroom entered the FA setting, they were directed away by th e principal investigator. For Jimmy, it was necessary to include a peer confederate in the FA sessions as the first FA conducted without a peer confederate did not replicat e natural classroom conditions. The teacher indicated that other peers preci pitated his challenging behavi ors and that he was unlikely

PAGE 114

103 to engage in his target cha llenging behaviors (verbal defi ance, yelling, teasing, saying “No,” whining, crying) without a peer pres ent. For the other 3 participants, the participating teacher and the child were the on ly individuals included in the FA sessions. The FA in the current study was not hypot hesis driven but instead functioned independently of the descrip tive assessments, as conducted by Iwata et al., (1982/1994). All conditions were videotaped and data we re collected as previously described. The following is a description of each FA session. Free Play The free play condition was used as a control for the other four conditions. During free play, the participan t had access to preferred toys, neutral activities and the teacher provided positive at tention in a continuous, non-contingent rate, approximately 5 times per minute per free pl ay session. No instructional demands were placed on the participant and demonstrations of challenging behavior were ignored. Tangible During the tangible condition, the participant was presented with the preferred item. The item was removed after a set interval of 20 seconds. The teacher maintained a distance of three to six feet from the particip ant. Upon demonstration of a challenging behavior, the teacher presen ted the participant with the tangible reinforcement with the statement “Now it’s your turn with the toy.” After 30 seconds, the tangible item was removed again with the st atement “Now it’s my turn with the toy.” Demonstration of the behavior resulted in the item being returned. No other demands were placed on the child. All other behaviors that did not meet the response definition criteria were ignored, and the teacher provided non-contingent attention statements to the participant on a fixed 30-second interval (as prompted by the principal investigator).

PAGE 115

104 Attention In the attention condition, the teacher maintained a distance of 3-6 feet from the participant and pretended to be o ccupied with paperwork. When the participant engaged in the disruptive behavior, the teacher approached the partic ipant, placed a hand on the child’s back and gave a verbal reprim and that represents what might typically happen during demonstrations of this challe nging behavior (i.e., “Stop doing that. You shouldn’t do that because it is against the rules. You are not following directions when you do that”). Participant children had access to the neutral activity and no other demands were placed on them. When the behavior was not occurring, the teacher did not engage in social attention. Escape In the escape condition, the participan ts were assigned a predetermined task that was determined by the teacher to be challenging, but within their intellectual ability (i.e., copying their own name 10 times). Verbal instructions and modeling of the task were provided at the be ginning of the session. When th e participant engaged in the challenging behavior, the task was removed for 30 seconds and the teacher said, “Time to take a break.” Following a 30-second brea k, the task and demand was presented again. The neutral activity was not available. The teacher provi ded non-contingent attention statements to the participant on a fixed 30-second interval. Ignore During the ignore condition, the particip ant was directed to stay in the assigned area. The preferred tangible item and neutral activity materials were accessible but no direct interaction w ith the teacher occurred and no demands were place on the participant child. All demonstrations of the challenging behavior were ignored. All five conditions were presented for a minimum of three sessions. However, after the third session, the principal investigator a nd a researcher with extensive experience in

PAGE 116

105 single subject design reviewed all data and only conditions that c ontinued to elicit challenging behavior and a cont rol condition (either Free Play or Ignore) continued to be tested until stability in all the data occu rred (Northup et al., 1991). This was done to obtain at least a three-point trend and to minimize the intrusiven ess of conducting the functional analyses in natural settings. Social Validity Upon completion of all descri ptive assessments and the f unctional analysis phases, the teachers completed a 9-item Likert-type ra ting scale to evaluate the acceptability of each of the assessment procedures and to de termine how helpful and/or disruptive the assessments were to their classroom milieu. They were also given an opportunity to answer 14 open-ended questions regarding this procedure and the re sults of the various assessments in an interview with the principal investigator. The results were presented after the questionnaire a nd the first nine questions of th e social validity interview had been completed, as outlined in the data co llection procedures of this chapter. Procedural Integrity Procedural integrity data was collected for 21% of the FA conditions. Each FA condition had specific elements that were m onitored. Procedural integrity was based on the teacher providing the appropriate conse quence upon demonstration of the behavior for the escape, tangible and attention conditi ons. Procedural integrity was calculated by dividing the number of occurr ences of the challenging beha vior without the appropriate consequence/total occurrences of the challenging behavior X 100. Procedural integrity also monitored that continuous non-contingent attention was delivered for free play and that all challenging behavior was ignored for free play and ignore. For these conditions, ignore was evaluated by calculating the total number of challenging behaviors

PAGE 117

106 demonstrated that did not receive a conti ngency/ total number of behaviors X 100. Free play was monitored to insure that the partic ipant child had unrestri cted access to desired toys and continuous non-contingent adult a ttention at a minimum of one attention statement every 10 seconds. If it was dete rmined upon reviewing the videotape that access was not given, then this condition was conducted again. The procedural integrity for Ramon was evaluated for 5/24 (21%) conditions. The range of the procedural inte grity calculations was from 84%-100% with an average of 90.6%. The procedural integrity for Anthony was evaluated for 4/19 (21%) conditions. The range of procedural inte grity calculations was from 80%-100% with an average of 95%. The procedural integrity for Jimmy wa s 9/42 (21%) conditions. The range was from 75%-100% with an average of 97%. The proced ural integrity for Greg was evaluated for 4/18 (22%) conditions. The range was from 80%-100% with an average of 95%. Design and Data Analysis Design A multi-element design was used to evaluate the relative differences across functional analysis conditions. No experime ntal design occurred for the descriptive assessments because the outcomes were not experimental. Data Analysis This study was separated into three phases: preliminary assessments, descriptive assessment methods and functional analyses. For the descriptive assessments, data analysis for two of the inst ruments produced permanent products. For the checklist, the MAS, has a simple scoring rubric in whic h the highest numbered score indicates the primary function of the problem behavior. The principal investigator scored this instrument after all other assessments had been completed for the child participant.

PAGE 118

107 Scoring this instrument consis ted of tallying the responses to the different items and then assigning a hypothetical function ba sed on the highest score. Because the FAI and the ABC direct obs ervation sheet are both more subjective measures, both require some interpretation on the part of the scorer. Immediately after completing the Functional Assessment Interv iew, the research assistant (Research Assistant 1) who administered the indi rect descriptive assessments assigned a hypothetical function based on the responses to the interview questions. The ABC was scored by both research assi stants (2 & 3) who assigne d a primary function to the challenging behavior upon completion of the number of 15-minute intervals observed. It should be noted that all results for every desc riptive assessment were kept separate from every other instrumentsÂ’ result s and were also kept separate from the functional analysis in order to avoid influencing any other proc edures. For the descriptive assessments, all results were sealed in envel opes and were not discussed w ith other person nel involved with the research. The results were compared by the principal investigator after the final functional analysis session for each partic ipant had been completed. Since different individuals were collecting the assessment data, the results from each instrument remained separate from the others. For the functional analysis procedures, th e principal investigator used a frequency count that was translated into rate per mi nute, for each analog condition to evaluate the relative differences between conditions. The highest rate within a specific analog condition was used to determine the function of the behavior (i.e. if the behavior occurs most frequently during the atte ntion condition, then the functi on of the behavior would be identified as attention). All data was gra phically displayed and visually inspected for

PAGE 119

108 magnitude, trend, and stability to determine what analog condition elicited the highest rate of challenging behaviors and therefore represented the function of that behavior. Data was visually inspected using the criteria outlined in Hagopian et al., (1997) and also evaluated using criteria outlined in Tawney and Gast (1984). Each graph was evaluated for the highest number of data points across all sessions over all other conditions. Each data path was additionally evaluated for st ability across all sessions, percentage of overlapping data points between conditions and any discernible trend in the rate of the behavior. After all descriptive assessments and the functional analysis were individually evaluated to determine their results (i.e. wh at is the primary function according to the FAI, MAS, and ABC), the results for each pa rticipant were assimilated, using a form developed by the principal investigator. For comparing descriptive assessments, agreement results were calculated by creating ratios of agreements over the total possible agreements for the two other instruments for each participant (Agreements/2). This method of calculation of data is simila r to the methodology used by Murdock et al., (2005). The total number of agreements fo r each descriptive instrument was then calculated across four participants (Agreemen ts/Number of Other Instruments x Number of Participants). Thus, there were 8 po ssible comparisons (2 x 4). For the total agreements of the descriptive assessments with the functional analysis, the ratio was calculated for each descriptive assessment acro ss the four participants (Agreements / 4) and a total number of agreements for all descriptive assessment was calculated [Agreements / Participants x Number of Desc riptive Assessments (4 x 3)]. Because each assessment only determined a primary func tion, a final summary was completed noting

PAGE 120

109 agreement and disagreement between descriptive assessments and agreement and disagreement with the functional analysis Then a summative statement regarding the agreement between assessments and with the functional analysis was developed. The final scoring sheet can be found in Appendix H. The results of the social validity questionnaire were ca lculated by averaging the four teachersÂ’ responses per item. For the social validity interview, the principal investigator transcribed the responses and these transcri ptions were evaluated for commonalities, agreement with questionnaire responses and salient points regarding process components and their results.

PAGE 121

110 CHAPTER 4 RESULTS This chapter presents the results of th e current study. The purposes of this study were to compare the findings from three desc riptive methods for determining the function of behavior with each other and to compare the findings across methods, with the results of a functional analysis. Also, teacher percep tions of the usefulness of the process and procedures were evaluated. Research questi ons were addressed through three specific examinations. First, the functions of young ch ildrenÂ’s challenging be havior as identified through various descriptive functional a ssessments (e.g. the Motivation Assessment Scale, the Functional Assessment Interview and an Antecedent-Behavior-Consequence direct observation) were compared to assess co ncurrent validity across instruments. This examination evaluates the c onsistency of functions iden tified through commonly used descriptive assessments. Sec ond, outcomes of these same methods were compared with functions verified through expe rimental functional analyses. This examination determines the accuracy or construct validity of descri ptive methods. Third, social validity of the descriptive assessment and th e functional analysis procedures and outcomes were evaluated to determine how palatable t eachers found the various methods, how likely they were to use each in the future, and what impact they believed the results might have on future intervention strategies. Individual results are first presented for each participant child. Then, results comparing outcomes across descriptive me thods are delineated. Next, results of

PAGE 122

111 comparisons between descriptive methods and functional analysis are presented. Finally, the results from the social validity ques tionnaires and intervie ws are presented. As noted in chapter three, one deviati on from the original planned procedures occurred. Originally, the Problem Behavior Questionnaire (PBQ) (Lew is, Scott & Sugai, 1994) was administered for each participant, along with the other assessments mentioned above. However, due to an incompatible sc oring taxonomy, the results of this indirect assessment were not comparable with thes e assessments or the functional analyses. Therefore, these results will not be presente d in the results sec tion. The rationale for excluding this assessment was made post hoc. Th is decision was based on the instrument lacking two options for the function of a ch allenging behavior: tangible and sensory. Thus, it was impossible to determine if disagreements between the PBQ and other descriptive assessments, or func tional analyses results, were a result of these two options not being available or were a result of genuine disagreement across methodologies. Individual Child Participant Results In this section, results from the three descriptive assessments and the functional analysis are presented for each child partic ipant. Table 4-1 presents summative data regarding the results of all of the descriptive assessments and the functional analysis for each participant. The functional analysis column is presented in bold face as those results represent the function of the behavior as identified th rough experimental analyses as opposed to the hypothetical functions determin ed by the descriptive assessments. For the purposes of this study, the specific identified functions are presente d to illustrate the agreements and disagreements between the different assessment methodologies. Because the identified function is idiosyncratic to the participant being assessed, the actual function identified is incidental to evaluati ng the agreements versus disagreements.

PAGE 123

112 Table 4-1. Summary of the Functions of Challenging Behaviors Across Assessment Methods Participant MAS FAI ABC FA Ramon Tangible Tangible Attention Attention Anthony Tangible Escape Tangible Tangible Jimmy Tangible Attention Tangible Tangible Greg Sensory Attention Attention Attention Ramon Ramon’s challenging behaviors were disrup tive vocalizations defined as raising his voice above the classroom volume level, asking more than three questions in a 10-second interval or using profanity or “bathroom ta lk.” The MAS and the FAI indicated that the primary function was tangible. The ABC iden tified attention as the function of the challenging behavior. As seen in Figure 4-1, in Ramon’s functional analysis, in comparison to the other conditions, the attenti on condition elicited the highest rate of the challenging behavior ( M = 2.6 responses per minute with a range of 1.8 3.2 responses per minute). The other functional analysis cond itions elicited the cha llenging behavior at substantially lower rates, tangible ( M = 1.73), followed by escape ( M = 0.6), free play ( M = 0.3), and ignore ( M = 0.2). Although overlap between the tangible and attention conditions occurred for 50% of the se ssions (Tawney & Gast, 1984), the tangible condition demonstrated a deceler ating trend across sessions 4-6. Based on visual analysis of the magnitude of the differences between conditions and the trends of the data, these results suggest that Ramon’s challenging be havior is maintained by attention. This outcome only agrees with the ABC direct observation.

PAGE 124

113 0 0.5 1 1.5 2 2.5 3 3.5 123456 Sessions Attention Tangible Escape Free Play Alone Figure 4-1. Challenging Behaviors per Minute During Functional Analysis for Ramon. Anthony Anthony’s challenging behavior was disr uption defined as crying, yelling “Mom” or “Mama”, rolling around on the ground, rais ing his voice above the classroom volume level and leaving his assigned area (defined as being more than 6 feet away from his assigned area). For Anthony, the MAS and the ABC indicated that the primary function was tangible. The FAI concluded that the pr imary function of the challenging behavior was escape. For Anthony, the tangible condition elicited the highest rate of behavior ( M = 1.0 response per minute across all sessions). Th e other conditions elicited challenging behaviors at lower rates, attention ( M = 0.3), followed by escape ( M = 0.1), free play ( M =0.0) and ignore ( M = 0.0). See Figure 4-2 for Anthony’ s functional analysis graph. There was no overlap in the data between ta ngible and the other c onditions. These results

PAGE 125

114 suggest that Anthony’s challenging behavior is maintained by access to tangible items. This outcome for the functional analysis agreed with the MAS and the ABC. 0 0.2 0.4 0.6 0.8 1 1.2 1234 Sessions Attention Tangible Escape Free Play Alone Figure 4-2. Challenging Behaviors per Minu te During Functional Analysis for Anthony. Jimmy Jimmy’s challenging behavior was disr uption defined as verbal defiance, screaming, teasing, saying ‘No’, crying or wh ining. The MAS and the ABC indicated that tangible was the primary function. The FAI conc luded that attention was the function of the challenging behavior. The methodology of the functional analysis for Jimmy varied from the other participants. Tw o functional analyses (no peer confederate present v. peer confederate present) were conducted for Ji mmy. In the first FA, the conditions were identical to the other participants (i.e., no peer confederate was present). However, after conducting four sessions with al l five conditions, the rate of the challenging behavior was

PAGE 126

115 low: attention ( M = 0.15), tangible ( M = 0.10), escape ( M = 0.05), free play and ignore ( M = 0.00) and no apparent consistent differences between the conditions occurred. After these four sessions, all conditions le veled off to zero. After discussi ng the results of the initial FA with the teacher and observing Jimmy in the classroom, the researcher determined that the challenging behavior was likely to occur only in conditions where a peer was present. This discussion of the results w ith the teacher before completion of all assessments was atypical of the protocol us ed for all other participants. For other participants, results were only discussed af ter completion of all of the assessments. However, due to the rate of zero behavior s across all conditions, it was necessary to modify the protocol for Jimmy. Therefore, a second FA that included a peer confederate was conducted to assist in re plicating conditions that were naturally occurring in the classroom. The teacher selected the peer confederate based on observed patterns of behavior in the classroom and parental consent was attained for the peer to participate. As seen in Figure 4-3, during the second FA, the tangible condition had the highest mean rate of challenging behavior ( M = 0.34 responses per minute with a range of 0.2 0.8 responses per minute) and an accelerati ng 3-point data path for sessions 9 to 11. Three of the other conditions elicited respons es at considerably lo wer rates: attention ( M = 0.10) ignore ( M = 0.0) and free play (M= 0.0). One impor tant consideration is that the rates of behavior were lower for all conditions for this participant compared with the other three participants. The escape conditi on elicited the challenging behavior at a similar rate as tangible ( M = 0.31) and had 57% overlapping data points with the tangible condition. The escape condition initially demonstr ated a high rate of behavior, however beginning at session 8, the behavi or began a decelerating trend that leveled off to zero in

PAGE 127

116 the last three sessions. This suggests that Ji mmyÂ’s challenging behavior is maintained by access to tangible objects. This function agr ees with the outcomes of the MAS and the ABC. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1234567891011 Sessions Attention Tangible Escape Free Play Alone N o Peer ConfederatePeer Conferderate Figure 4-3. Challenging Behaviors per Minu te During Functional Analysis for Jimmy. Greg GregÂ’s challenging behaviors were disruptiv e noises defined as tapping his pencil, whistling, and raising his voi ce above the classroom volume level. For Greg, the MAS indicated that the primary function of challe nging behavior was sensory. The FAI and the ABC concluded the challenging behavior was ma intained by attention. For the functional analysis, the attention condition elicited the highest rate of cha llenging behaviors ( M = 1.5 responses per minute with a range of 1.0 1.8 responses per minute) and an accelerating 3-point data path across sessi ons 1 to 3. The next highest average rates of challenging behaviors per minute were tangible ( M = 1.2), ignore ( M = 0.3), escape ( M = 0.2), and free

PAGE 128

117 play ( M = 0.03). The tangible condition had 66% of overlapping data points with attention. Although the tangible condition elic ited the highest rate of challenging behavior for the first session, the data path demonstrated a decelerating trend for the next two sessions. Based on visual analysis of the magnitude of the differences between conditions and the trends of the data, atten tion was concluded to be the function of the challenging behavior. This outcome agrees with the FAI and ABC. See Figure 4-4 for GregÂ’s functional analysis graph. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 1234 Sessions Attention Tangible Escape Free Play Ignore Figure 4-4. Challenging Behaviors per Minu te During Functional Analysis for Greg.

PAGE 129

118 Comparison of Outcomes of Descriptive Assessments Agreements between descriptive assessmen ts were based solely on the primary function that was determined by each descrip tive assessment. For the MAS, the primary function was determined using the scoring ru bric in which the highest numbered (and ranked) score indicates the primary function of the challenging behavi or. For the FAI, the research assistant who administered the indirect descriptive assessments assigned one of four hypothetical functions (i.e., tangible, escape, attention, sensory) based on the responses to the interv iew questions. For the ABC, anothe r research assistant assigned a primary function from one of the four possi ble functions upon completion of the eight 15minute observation intervals. Table 4-2 presents data regarding th e total number of agreements for each descriptive assessment as it is compared with the other two descriptive assessments and with the functional analysis results. The agreement ratio for comparing descriptive assessments with each other was calculat ed by determining the total number of agreements over the total possible agreem ents with the two other instruments (Agreements / 2) (Murdock et al., 2005). Agreement for each descriptive assessment with the functional analysis is notated with an asterisk for each individual participantÂ’s descriptive assessment. The total number of agreements with other descriptive assessments was then calculated across all four participants (A greements/Number of Other Instruments x Number of Participants). Thus, there are eight possible comparisons (Agreements/2 x 4) for calculating the total agreements. For the total agreements of the descriptive assessments with the functional analysis, the ratio is calculated for each descriptive assessment across the f our participants (Agreements/4).

PAGE 130

119 Table 4-2. Summary of Agreements Between Descriptive Assessments Compared With Each Other and with a Functional Analysis Participants MAS FAI ABC Ramon 1/2 (50%) 1/2 (50%) 0/2 (0%)* Anthony 1/2 (50%)* 0/2 (0%) 1/2 (50%)* Jimmy 1/2 (50%)* 0/2 (50%) 1/2 (50%)* Greg 0/2 (0%) 1/2 (50%)* 1/2 (50%)* Total Agreements with Other Descriptive Assessments 3/8 (38%) 2/8(25%) 3/8 (38%) Total Agreements with Functional Analysis 2/4 (50%) 1/4 (25%) 4/4 (100%) *indicates agreement with function as determined by the functional analysis In general, the total consistency between the three descriptive assessments was low. When comparing across descriptive assessments the MAS findings were consistent with the ABC for total agreements of 3/8 (38%) a nd slightly greater than the FAI, 2/8 (25%) for total agreements. Across all participants the MAS agreed with the FAI only once and the MAS agreed with the ABC twice. Th e ABC agreed with the FAI only once. Calculating the number of agreements ove r the opportunities fo r agreement for all descriptive assessments [Other descriptive assessments x All descriptive assessments x Total number of participants) ( 2x3x4)] indicates that they agr eed with one another in 8/24 (33%) of all cases. For all 4 participants, a majority of descriptive assessments indicated the same hypothetical functions. In other words, of the three descri ptive assessments, for each participant, two of the assessments had the same resu lts. For example, for Jimmy, both the MAS and the ABC agreed that the hypothetical function of the challenging behavior was tangible. For 3 of the 4 part icipants, the hypothetical function that was

PAGE 131

120 determined by more than one instrument wa s confirmed by the functional analysis. Data evaluating the results of each descriptive a ssessment in comparison to the results of functional analysis are presen ted in the following section. The results of the three descriptive assessm ents were compared with the results of functional analyses conducted for all 4 pa rticipants. The MASÂ’s hypothesized function agreed with the FA results for 2/4 (50%) of the participants. The FAIÂ’s hypothesized functions agreed with the FA results for 1/4 (25%) of th e participants. Unlike the other descriptive assessments, the A BCÂ’s hypothesized functions agreed with the FA results for 4/4 (100%) of the participants. That is, for all 4 participants, both the ABC and FA were in agreement as to th e function of behavior. The overall agreement between the desc riptive assessments and the functional analysis was 7/12 (58%). This was calculat ed as the number of agreements/number of participants x total number of descriptive a ssessments (Agreements/4 x 3). However, the number of agreements was not evenly distri buted across descriptive assessments because the direct descriptive observations (ABC) ag reed with the FA conclusions for all 4/4 (100%) of the participants. T hus, the indirect descriptiv e assessments were only in agreement with the functional analysis in only 3/8 (38%) of cas es. Other important considerations were how pala table teachers found the vari ous methods, how likely they were to use each in the future, and what imp act they believed the results might have on future intervention strategies. This was evalua ted using a social validity questionnaire and interview.

PAGE 132

121 Social Validity Questionnaire and Interview Results Social Validity Questionnaire The social validity questionnaire was cr eated by the principal investigator and was comprised of eight questions on a 5-point Likert-type scale that was completed by each teacher participant after the results of al l descriptive assessments and the FA were determined. However, these results were not presented to the teachers until after they had completed the questionnaire and answered th e first nine questions of the structured interview. For the questionnaireÂ’s scale, 1 represented Not at All (helpful, useful, disruptiveÂ…) and 5 represented Ver y (helpful, useful, disruptiveÂ…). Items 1 through 6 asked how time consuming helpful or disr uptive various assessments were, and items seven and eight asked about future participa tion in this type of study and value of the overall process. It should be noted that this questionnaire is multi-directional, in that for some items a higher score represents positive perceptions of the process and for some items a higher score represents negative per ceptions of the process. For items #2, 3, 4, 7, & 8, a higher score represented a positive perception of this process and for items one, five, and six a higher score represents a nega tive perception of this process. Table 4-3 presents each item, the responses from all four teachers, and the aver age score for all of the teachers. Table 4-3. Results of the Soci al Validity Questionnaire Item Ms. Rose Ms. Terry Ms. Amy Ms. Alexis Mean Score (1) How time consuming were the checklists and interviews? 2.0 3.0 3.0 3.0 2.8 (2) How helpful was the training on Functional Behavior Assessment 5.0 5.0 4.0 4.0 4.5 (3) How helpful was the Functional Assessment Interview? 5.0 4.0 3.0 5.0 4.3

PAGE 133

122 Table 4-3. Continued Item Ms. Rose Ms. Terry Ms. Amy Ms. Alexis Mean Score (4) How helpful was the Motivation Assessment Scale? 4.0 4.0 3.0 5.0 4.0 (5) How disruptive was the direct observation descriptive assessment (ABC)? 1.0 1.0 1.0 3.0 1.5 (6) How disruptive was the functional analysis? 1.0 3.0 4.0 3.0 2.8 (7) Would you be willing to let another of your students participate in this study? 5.0 5.0 2.0 3.0 3.8 (8) How valuable was it to determine the function of the problem behavior? 5.0 5.0 3.0 5.0 4.5 The teachers rated the descriptive assessments as moderately low in terms of being time consuming ( M = 2.8) with a range of 2-3. For the ABC direct observation, the teachers felt that having the research assistan t observe natural clas sroom occurrences was not disruptive ( M = 1.5) with a range of 1-3. The functional analysis procedures were considered more disruptive ( M = 2.8) with a range of 1-4. Items 2,3,4 and 5 asked how helpful the training and various descriptiv e assessments were. The teachers rated the hour-long training session on functional behavior assessment and functional analysis, as helpful to very helpful ( M = 4.5) with a range of 4-5. The FAI was considered slightly more helpful ( M = 4.3) than the MAS ( M = 4.0). In response to whether they would be willing to participate in this study agai n, the teachersÂ’ responses were mixed ( M = 3.8) with a range of 2-5. According to the questi onnaire, teachers did beli eve it was valuable to determine the function of the challenging behavior, ( M = 4.5), with a range of 3-5.

PAGE 134

123 Social Validity Interview The social validity interview was created by the principal investigator and asked each participating teacher 14 ope n-ended questions regarding th eir experiences with this research project. The results of the descriptive assessment and the functional analysis were presented after the questi onnaire and the first 9 questions of the 14-question social validity interview had been co mpleted. According to intervie w responses, three out of the four teachers had previously conducted some type of FBA before. Amy, Jimmy’s teacher, had never before completed an FBA. The teachers’ defined the term function as ‘what the student gets out of doing it (the behavior)’, or ‘what factors in the environment cause the behavior.’ The teachers all responded that th ey liked all of the assessment methodologies. However, Terry, Anthony’s teacher, noted that the FAI had a number of questions that “didn’t seem to fit (this child) at all,” and that it was a little lengt hy. All of the teachers believed that the ABC was the best method of capturing the function of the behavior. Amy stated that the importance of the ABC was that “someone who is not involved” was observing the student because teachers who are within the situation “become immune” to a lot of the behaviors. All of the teachers’ respons es to the questions about th e use of functional analysis indicated that they believed the particip ating child had “gotten bored” with doing the same conditions repeatedly or that the ch ild had “figured it out” and was deliberately manipulating the conditions. Rose, Ramon’s teacher, noted that the best part of the FA was the use of the video camera because ot hers “saw him in actionthe videotape captured everything he did.” Before reviewing the results, the teachers were asked what they believed to be the function of the challenging behavior. All four teachers stated that they believed there

PAGE 135

124 were multiple functions and two teachers stat ed that the function could be constantly changing. For example, Terry, Anthony’s teac her, stated, “There are many functions. Sometimes it’s attention, sometimes he want s to get out of doing things…” When the teachers examined the results of the different assessments, all of them agreed with the results of the different assessments even wh en different assessments concluded different functions of the behavior. The teachers explai ned these differences by attributing them to the different contexts. For example, Terry sp ecifically identified that these different outcomes represented how the function of the behavior was fluid and often determined by a variety of environmental factors. She di scussed that in Circle Time, a large group passive activity, the function of his behavior is escape, but on the playground or during Centers Time, getting access to tangibles is the function of his behavior. All of the teachers stated that they would consider the function of the challenging behavior in the future but only Ms. Alison, Greg’s teacher, id entified a specific strategy that she could use with this information about the function of his behavior, “I’m going to use a timer to control his access to attention….” Summary Results of the study indicated the desc riptive assessments had relatively low consistency with each other 8/24 (33%) agreem ent, and that the indirect descriptive assessments (the FAI, and the MAS) also ha d low agreement with the functional analysis with 3/8 (38%) agreements. However, the di rect descriptive obser vations (ABC) agreed with the FA conclusions for all 4/4 (100%) of th e participants. Results of the social validity measures indicated that teachers fe lt all methods of assessment, including the functional analyses, were helpful, minimally disruptive, and that these results would impact their behavioral interventions. Ho wever, they identified the ABC direct

PAGE 136

125 observation as the least disruptive, most valuab le way to obtain information regarding the functions of challenging behaviors.

PAGE 137

126 CHAPTER 5 DISCUSSION The use of functional behavior assessme nts (FBAs) continues to increase as a result of increasing numbers of children engaging in problem behaviors (Chandler & Dahlquist, 2002), and a need to develop eff ective assessments and interventions that reduce these problem behaviors and increase adaptive behaviors. Unfortunately, the extensive use of FBAs in pr actice has outpaced its research base (Sasso et al., 2001). Failure to identify the most effective co mponents of the FBA process with valid outcomes will have adverse effects on childre n and educators, because ineffective FBAs may lead to interventions that will not decr ease challenging behavior or increase adaptive behaviors. There are a number of indirect and direct descriptive methodologies that can be used within the FBA process. However, limited research exists that compares the outcomes of these instruments with each ot her or validates th em through functional analysis procedures. Thus, there is a continui ng need to empirically compare and validate the outcomes of these descri ptive assessment methods. In this study, the outcomes of a seri es of indirect and direct descriptive assessments were compared with each other and compared with the outcomes of a functional analysis in order to answer two research questions First, how do the hypothetical functions of target behavior s identified through various descriptive functional assessments compare to each ot her for young childrenÂ’s challenging behavior? Second, how do the hypothetical functions of ta rget behaviors identified through various descriptive functional assessments compare with functi ons identified th rough a functional

PAGE 138

127 analysis? Additionally, teachersÂ’ beliefs about the process and the results were evaluated through social validity measures. Unlike prev ious research, (a) the functional analyses were conducted independent of the descri ptive assessment resu lts, (b) they were completed within the classroom setting, and (c) the teacher, with assistance from the researcher, served as the therapist for implementing the five functional analysis conditions. All assessments we re counterbalanced, and th e results of the various assessments were kept separate from each other until completion. Teachers also completed a social validity questionnaire and interview. The results indicated that the descrip tive assessments had low consistency with each other, with 8/24 (33 %) agreements. When comp aring across descriptive assessments, the MAS findings were consistent with the ABC for total agreements of 3/8 (38%) and greater than the FAI outcomes, 2/8 (25%), for total agreements. The MAS agreed with the FAI only once and the ABC twice across all partic ipants. Additionally, the ABC agreed with the FAI once. When the descriptive assessments were compared to the outcomes of the functional analyses, the fi ndings varied according to instruments. The indirect descriptive assessme nts (FAI, and MAS) had low agreement with the outcome of the functional analyses with 3/8 (38%) agr eements. However, the direct descriptive observations (ABC) agreed with the functional analyses conclusions for all 4/4 (100%) of the participants. Results of the social validity measures indicated th at teachers felt all methods of assessment, incl uding the functional analyses were helpful, minimally disruptive, and these results would imp act their behavior al interventions. Limitations of this study ar e presented in the first sect ion of this chapter. Then, interpretations of results are pr esented and compared with othe r studies that have directly

PAGE 139

128 evaluated descriptive assessment methods. Ne xt, extensions of this study and future research directions are discussed. Finally, th e implications of this studyÂ’s findings in regard to current practices in schools are presented. Limitations Although results of this study extend the literature regarding the reliability and validity of descriptive assessmen ts and functional analyses in the FBA process, there are notable limitations to this st udy. First, the limited sample si ze of 4 participants impacts the generality of these findings and replicatio ns of this study are necessary to improve the external validity. By only having 4 particip ants in this study, the ability to draw conclusions about these results in widesp read applications is severely limited. Unfortunately, this is a limitation that is prevalent among previous research studies within the literature that have conducted direct evaluati ons of different functional behavior assessment components. For the eight studies in the direct evaluation section of the literature review, the m ean number of participants is relatively the same in comparison to this study and only 3 participan ts each were evaluated in six of the eight studies. It is possible that the time-consumi ng, technically intensive nature of functional analysis makes using larger numbers of par ticipants difficult. For example, the one study that used 8 participants (i.e., Murdock et al., 2005) did not conduct functional analyses. Therefore, to extend these findings, this study mu st be replicated using more participants in the same setting and more participants of different ages and characteristics in different classroom settings. Second, the context in terms of settings (e.g. the playground) and situations (e.g., free play) for the occurrence of the behavior was variable across descriptive assessment instruments and may be a plausible explanation for the variability of the outcomes of the

PAGE 140

129 instruments (Floyd, et al., 2005; Shriver, Anderson & Proctor, 2001). The teachers alluded to this in their op en-ended responses during the so cial validity interview when they said that different assessments were just measuring the function in different contexts. As Repp and Munk (1999), expl ain, by extending instruments beyond their “domain of interest,” (p.154) the instrument itself cannot be accurately assessed. In other words, if the function of the challenging be havior is assessed duri ng Circle Time, then that is the function of the behavior during Circle Time. The design of this study was not sensitive to the impact of different cont exts, but rather compared agreement or disagreement of function across contexts. Ques tionnaires and intervie ws were subject to the teachers’ memories and biases of behavior in different settings (Shores, et al., 1999). Additionally, direct descriptive observations captured short intervals across multiple settings. Finally, the functional analysis onl y measured independent (or in the case of Jimmy, one peer present), structured or se mi-structured activities. The difference in settings and situations is especially impor tant in considering the functional analysis outcomes, as this process is regarded in re search as the ‘gold st andard’ and the only way to determine the actual rather than th e hypothetical function (Lerman & Iwata, 1993). Third, related to the previous limitation are specific limitations of internal and external validity for the func tional analysis procedures. A lthough the functional analyses were conducted in situ, certain conditions could not be replicated in analogue assessments. For example, for Anthony’s teac her, specifically identified a large group, passive setting (e.g. Circle Time) that se rved as the specific antecedent to escapemaintained behavior. Unfortunately, this setti ng and activity could not be evaluated in the more individualized functional analysis settin gs because of practical limitations and

PAGE 141

130 threats to internal control that incorporat ing greater numbers of other children would entail. Similarly, in some cases, specific ta rget behaviors, identified by the teacher as problematic could also not be evaluated in the FA for legal and ethical reasons. Amy, Jimmy’s teacher, originally identified aggres sion toward peers as his most frequently occurring, most problematic behavior. Howe ver, contingently reinforcing aggression toward a peer was not possible for ethical reasons and it was nece ssary to find other challenging behaviors that may be precurs ors to aggression. This may provide an explanation for Jimmy’s extended functional an alysis, overall low rates of behavior and variability across descriptive instruments. One plausible explanation is that functiona l analysis as it was originally conceived may not be the ‘gold standard’ in natural sett ings with children at risk for EBD. This difference is feasible considering Iwata et al.’s (1982/1994) origin al description of functional analysis as a procedur e that is being used in highly controlled clinical settings to counteract the debilitating effects of self-injurious be havior and used in lieu of punishment in the form of aversive stimul ation. Sprague and Horn er (1999) detail the need to develop and refine methodologies to determine the functions of high intensity behaviors beyond clinical settings. They sugge st developing more sensitive systems of measurement (i.e., sequential analysis) and a ddressing lower intensit y behaviors that are part of the same response class. This may allow researchers to gather knowledge about the functions of high intensity behaviors without eliciting and/or reinforcing the dangerous behaviors. Evaluating the functional analyses data within this study al so indicates that limitations existed with the internal validity of these functional analyses. For three of the

PAGE 142

131 four participants, Ramon, Jimmy and Greg, th ere was overlap with a number of the data points including the condition that was conclu ded to be the function of the challenging behavior. This variab ility makes the functional analys es conclusions less persuasive. Gathering more data could have counteracted the variability for these three participants and made the functional analyses results more ro bust. This lack of a clear and stable trend in the data may be related to conducting th e functional analyses in the classroom and having the teachers play a primary role in conducting the functional analyses. Conducting additional functional analysis sessions and gathering more data was difficult due to logistical constraints. Finally, although the functional analyses we re conducted in natural settings with the teacher implementing the conditions, these conditions were still contrived and subject to researcher influence. In order to ma intain procedural integrity, the principal investigator guided the teacher through a ll conditions using prompts and cues. The principal investigator then code d the data and visually analyz ed the data to determine the function of the challenging behavior based on it This pronounced researcher influence is similar to the studies discussed in the second section of the literatur e review in chapter two. As the teachers also stated in the social validity interview, in some cases, the children appeared to habitu ate to the conditions and may have been active in manipulating the contingencies. For exampl e, after Ramon had received contingent attention for his challenging behavior several times in the attention condition, he would engage in a challenging behavi or, and repeat the teachers ve rbal prompt in unison with her (i.e., “Stop doing that…”) or at other times during the functional analysis conditions. While the data in this study do not reflect if the participants were deliberately

PAGE 143

132 manipulating the contingencies, it is a worthwhile consider ation, as functional analyses with populations other than developmental di sabilities is still not supported by a robust research base (Ervin et al ., 2001; Gresham, 2003: Sasso et al., 2001). Thus, researcher influence, high demand of time and personnel, and contrived settings continue to be problematic in assessing the implementation of functional analyses in natural settings (see Ervin et al., 2001 for a review). Fourth, although the strength of the design of this study wa s to use different people to conduct different parts of the assessment; th is may have contribute d to the variability in the data. Using different respondents a nd data collectors (res earch assistants and teachers) was necessary to both maintain separation of the findings between the instruments and explore the efficacy of teacher-implemented functional analysis. However, the use of these different people also poses the possibility of introducing variability into the conclusions for the vari ous assessments especi ally in the case of interviews. As different individuals complete the assessments, the differences may be a result of the person completing the assessmen t, as opposed to the assessment itself. An ancillary consideration is th e use of trained graduate students, in conducting the ABCs and concluding hypothetical functions based on the FAI responses. This also limits the external validity when considering the appl ication of FBA in early childhood settings. It is feasible that educators w ith less expertise and training are typically responsible for completing and interpreting these descriptiv e assessments; thus, the findings may differ based on level of experience (Conroy, Clark, Gable & Fox, 1999; Shellady & Stichter, 1999). It is also feasible that educators ma y have more expertise than the research assistants. The personnel completing the FBAs (i.e., behavior s upport team, classroom

PAGE 144

133 teachers, behavior specialists, school counselors) is idiosyncratic to different educational settings and thus far little empirical research has been conducted on the effects of these staffing decisions (Scott, McInty re, Liaupsin, Nelson & Conroy, 2004). Fifth, evaluating only the primary function according to these assessments does not provide information regarding other aspects of the functional assessment process that may influence function, such as antecedent or setting events. For example, the FAI also gives information on specific antecedents or setting events, like medical or communication problems that may contribute to the occurrence of ch allenging behavior. Similarly, the MAS provides a rank order of hypot hetical functions that may be valuable; thus, indicating that behaviors may be maintained by multiple functions. By only focusing on a single, primary function determ ined by each assessment, the overall value of the assessment method is underestimated. Although this is a limitation, there were two reasons for focusing only on the primary func tion determined by each assessment. First, for clarity these other functions were disregarded in order to calculate agreements or disagreements between assessment methods. This is unlike previous research identified (Calloway & Simpson, 1998; Cunningham & OÂ’ Neill, 2000; Murdock et al., 2005) that used either a rank order or matched pr imary and secondary functions. Second, in response to calls for a more efficient, e ffective FBA methodology that translates to efficient and effective behavioral intervention plans in educational settings, addressing multiple functions in a multi-component behavi or intervention plan may not be feasible. Evaluating multiple functions within a behavior intervention plan for multiple children may be unrealistic (Johnston & OÂ’Neill, 2001; Sasso et al., 2001; Scott et al., 2004).

PAGE 145

134 Sixth, another limitation that may have contributed to the variability in the data are the broad definitions of the target behavior s for each participant. Specifically, these behaviors, defined under the term ‘challengi ng behaviors,’ addresse d a wide range of actions across multiple topographies. Although these broad definitions are similar to those used in other studies (Murdock, et al., 2005; Newcomer & Lewis, 2004; Sasso, et al., 1992), this does suggest a w eakness within the overall rese arch literature. By defining behaviors more narrowly, the individual func tions of specific challenging behaviors may be determined more accurately. The dr awback to narrowly defining challenging behaviors is the possibility of low rates of occurrence of the behavior making the determination of the function more difficult. Seventh, in regard to the social validity measures, the lack of anonymity for teachers completing the questionnaire and interview may have impacted their answers. Within this methodology, teachers completed the questionnaire and interview in a faceto-face meeting with the principal investigat or. Because teachers may have felt pressure to answer questions in a specific way (i.e. sa ying that the functional analysis procedures were not invasive), their respons es may not be representative of their actual perceptions. In the future, these questionnaires and interv iew sheets could be completed and returned anonymously, and this may give teachers gr eater comfort in discussing their true perceptions of the FBA process. Eighth, because no intervention phase was initiated within this study, it is impossible to determine if any of the result s would translate to effective function-based interventions for the students. Because the end goal of all FBA methodologies is to develop effective function-based interventions, th is study only represents an initial step in

PAGE 146

135 FBA research. Future research must focu s on the impact of conducting accurate FBAs and then using them in effective function-based interventions. It is vi tal that functions can be accurately identified, but these results then must be translated to viable function-based interventions that can be implemented in wi despread applications in natural settings. Interpretations of Results Comparison of Descriptive Assessments One interesting outcome of the current st udy is that the overall low levels of agreement between descriptive assessments disagree with other studies that have conducted comparisons between descriptiv e assessments (Arndorfer et al., 1994; Cunningham & OÂ’Neill, 2000; Ellingson et al., 2000; Murdock et al., 2005; Newcomer and Lewis, 2004). For example, Murdock et al. (2005) reported a higher level of agreements across all three descriptive assessment met hods and a higher level of agreement between teacher in terviews and direct observa tions. The current study only reported agreement between the FAI and the ABC for 1/4 (25%) participants. However, this discrepancy may be a result of compar ing different instruments across studies. Two studies (Cunningham & OÂ’Neill, 2000; Mur dock et al., 2005) used the Functional Assessment Observation Form (FAOF) (OÂ’N eill et al., 1997) as opposed to the ABC form for direct descriptive observations. Other studies used questionnaires created explicitly for that study or simply asked t eachers to name what they believed was the function of the target behavior (Cunningha m & OÂ’Neill, 2000; Ellingson et al., 2000). There are two other studies that used the Motivation Assessment Scale and then compared the results with other descriptiv e assessments. Compared with this studyÂ’s results of 3/8 (38%) agreements, Cunningham and OÂ’Neill (2000) used the MAS and agreed with other descriptive assessments for 8/9 (88%) opportunities. Arndorfer et al.

PAGE 147

136 (1994) also used the MAS and had 4/10 (40% ) agreements. Arndorfer et al., (1994) does point out that elevations on se veral subscales made the MAS difficult to interpret and that it was the secondary functions identified by the MAS that ha d high levels of agreement with other descriptive assessmen ts. However, while these results are divergent from this study, this study is similar to st udies that tested the reliabil ity of the MAS using a testretest format (Conroy, Fox, Bucklin & G ood, 1996; Sigafoos, Kerr & Roberts, 1994; Zarcone et al., 1991) or compared reliab ility with other instruments and analog assessment with individuals demonstrating self-injurious or st ereotypic behaviors (Crawford et al., 1992; Paclaw skyj et al., 2001). One possibl e explanation for the lower levels of agreement with this study may be related to the amount of independence between assessments that is maintained by having different personnel conducting the assessments and keeping the results of the assessment separate from each other. In evaluating the aforementioned studies, the greater the separation of completing the assessments, the lower the agreement between measures. Four other studies compared the FAIÂ’s re sults with other descriptive assessments (Arndorfer et al., 1994; Cunningham & OÂ’Neill 2000; Murdock et al., 2005; Newcomer and Lewis, 2004). All four studies had higher le vels of agreement when compared to this study. No studies to date have evaluated the te st-retest reliability or stability of the FAI (Floyd et al., 2005) although one study compar ed teacher and student FAI responses and found a high level of agreement between teac her and student responders (Reed, Thomas, Sprague & Horner, 1997). One consideration regarding the lower level of agreement on the FAI is the role of respons e bias for the interviews, and the fact that the responder may be answering with only partial or incomplete knowledge (i.e., only considering specific

PAGE 148

137 antecedents or unaware of antecedents that are shaping behavioral functions) (Gresham, et al., 2001). For example, for Anthony, whos e FAI indicated escape as the maintaining consequence, the majority of the teacher responses centered on occurrences of the challenging behavior during Circle Time, a large group, passive activity. According to the transcript, the teacher did not discuss occurrences of the behavior across other contexts. This shortcoming of indirect asse ssments has been comprehensively discussed in the literature (Gresham, 2003; Johnst on & OÂ’Neill, 2001; Johnston & Pennypacker, 1993; Sasso, Peck, & Garrison-Harrell, 1998). An other consideration is that the way the data is summarized does not lend itself to summarizing a function a nd that the structure of the summary of information is prohibiti ve to concluding a primary function. This suggests that restructuring the interview form to be more direct in concluding a primary function may be a way to access this information more conclusively. Five other studies used direct observati ons and compared them to other indirect descriptive assessments (Ar ndorfer et al., 1994; Cunningham & OÂ’Neill, 2000; Ellingson et al., 2000; Murdock et al., 2005; Newcomer a nd Lewis, 2004). All five studies reported higher levels of agreement with the direct a nd indirect descriptive observations. Again, a possible explanation for higher rates of agreement across desc riptive measures in other studies is the same lack of independence between assessment procedures and results making a confluence of results more feasib le. For example, for Newcomer and Lewis (2004) and Murdock et al. (2005) the first author conducted th e teacher interview and the direct observations generating hypothetical func tions based on each of these assessments. This makes it possible that the results of the first assessment informed the results of the second assessment. Additionally, in studies that used the FAOF (Cunningham & OÂ’Neill,

PAGE 149

138 2000; Murdock et al., 2005), by virtue of the FAOF procedure, the variables considered in the FAOF are directly informed by re sponses from the FAI. In the FAOF, O’Neill indicates in two items “… Lists behaviors to be monitored” and “… List any additional possible functions of behavior, if necessary, in the Perceived Functions section” (O’Neill et al., 1997, p. 40). For studies th at used the FAOF, all of this information is gathered from the FAI that was previously conducted. These results could reflect the inconsistent findings across commonly used descriptive as sessments in this study. The next section compares these same outcomes with functions verified through e xperimental functional analyses. This examination helped to dete rmine the accuracy or construct validity of descriptive assessment methods. Comparison of Descriptive Assess ments with Functional Analyses The overall agreement between the desc riptive assessments and the functional analysis was 7/12 (56%) agreements, but ther e was a clear difference between the number of agreements for direct and indirect assessments. The direct descriptive observations (ABC) agreed with the functional analysis conclusions for all 4/4 (100%) of the participants and the indirect descriptive a ssessments were only in agreement 3/8 (38%) times. This finding means that the two most pr evalent indirect descriptive assessments in research studies, the MAS and the FAI (F loyd et al., 2005), onl y agreed with the functional analysis results, 2/4 (50%) times and 1/4 (25%) times respectively. This low level of agreement for the indirect assessmen ts may be an important factor in guiding future research and practice and supports Sc ott et al.’s (2004) stat ement that “we have little or no evidence that altern ative methods such as interv iew, questionnaire or rating scales protocols can be va lid stand-alone methods of behavior analysis” (p.197).

PAGE 150

139 A comparison of levels of agreement betw een indirect descrip tive assessments and functional analyses indicates that the indirect descriptive assessments were lower in this study than in previous res earch that has compared de scriptive assessments with functional analysis results (Arndorfe r et al., 1994; Cunningham & OÂ’Neill, 2000; Newcomer and Lewis, 2004; Yarbrough & Carr, 2000). However, Calloway and Simpson (1998) had similarly low levels of agreement when they compared teacher interviews to functional analysis procedures. Again, this disparity between the aforementioned studies and this study may primarily be due to the use of the descriptive assessment results in previous research to guide the functional analysis conditions. In these studies, the conditions in the functional analysis were in some way informed by all descriptive assessments, often serving as confirmation, as opposed to independe ntly arriving at the function of the challenging behavior. For example, in Newcomer and Lewis (2004) hypothesis statements regarding antecedents fo r the challenging behavior were developed using descriptive assessments, then the expe rimental analysis cons isted of a reversal design, with and without the hypothesized an tecedents. Higher ra tes of challenging behavior in the antecedent condition conf irmed the descriptive assessments. Other possibilities for the low level of agreement be tween indirect descriptive assessments and the functional analyses relate to issues di scussed in the limitations section, that is, teachersÂ’ memories and biases of behavior in different settings (Shores, et al., 1999) and the limited knowledge regarding the effects of conducting functional analyses in natural settings with children with high incide nce disabilities (Ervin et al. 2001). The high level of agreement for the direct observations and the functional analysis is similar to the other studies that compar ed naturalistic dire ct observations with

PAGE 151

140 functional analyses (Arndorfer et al., 1994; Ellingson et al ., 2000; Murdock et al., 2005; Newcomer & Lewis, 2004 Sasso et al., 1992). Ho wever, the higher level of agreement in this study is more compelling considering the independence of assessments that was maintained between the ABC assessment and the functional analysis. For example, in Sasso et al. (1992) the direct observation assessment (ABC worksheet) was completed during specific activities that were similar to the conditions of the functional analysis conditions (i.e., presentation of a difficult ta sk). Thus, the ABC was designed to mimic functional analysis conditions making an i ndependent agreement be tween their results less compelling (Lerman & Iwata, 1993). The high level of agreement between the ABC and the FA for this study may be related to the overall higher levels of obser ved target behaviors and consequences that were concluded to be the function of the ta rget behavior. A review of the ABC direct observations indicates that for all 4 participants, the target behavior occurred between 1025 times, and the consequence of the behavior that was concluded to be the hypothetical function occurred an average of 14.8 times across the eight 15-minute observation periods. This high rate of observed behavior is related to the type of high frequency challenging behaviors that were addressed concurrently fo r each child (e.g. yelling, asking three questions in a 10-second interv al, leaving the assigned area). Addressing these types of behaviors may not be applicable for all children and certainly not for the low frequency, high intensity behaviors that often require an FBA according to federal mandates (Gresham, 2003). As Scott et al. ( 2004) point out, a behavi or that happens one time for a 1-minute interval per day requires 3.5 hours of observation time to have a 50% chance of observing the behavior. Thus, increa sing the efficiency and effectiveness of

PAGE 152

141 direct observation is an important considerati on for future research. To this end, a scatter plot could be used in conjunction with other direct descriptive asse ssments to make them more effective (Symons, et al., 1998). There are two final considerations in rega rd to the overall cont ribution this studyÂ’s results make to the research literature. The firs t consideration is the issue of the age of the participants. The use of FBA with young children, 8 years and younger, continues to garner researchersÂ’ attention as demonstrated in the first two sections of the literature review. However, only two studies in th e third section (Calloway & Simpson, 1998; Cunningham & OÂ’Neill, 2000) specifically focused on young children in their investigations. An evaluation of these three studies does not indicate that the methodology of FBAs with young children func tions any differently than the other studies in this direct evalua tion section. However, comparis ons across studies are difficult due to very different descriptive methodologies The results of this study add to this body of literature, but certainly more focus on FBA methodologies with young children in varying settings (i.e., preschool ve rsus kindergarten) is necessary. The second consideration in re gard to these results is th e issue of efficiency in conducting FBAs and the use of different met hodologies in light of this consideration. The results of this study s uggest that more efficient me thodologies may not lead to accurate results in terms of eith er concurrent or construct validity. Rather, in this study a negative relationship appears to exist between ef ficiency and valid resultsthe greater the efficiency (i.e., the MAS and FAI) the less valid the results. This consideration is elaborated further in the implications for practice section. Another important and related consideration is the social va lidity of this study. The social validity results measured the

PAGE 153

142 overall acceptability of the FBA process and its components, and the perceived value of their outcomes and are discu ssed in the next section. Social Validity Social validity of FBA processes an d outcomes remains an under-investigated aspect of the research literature (Ervin et al. 2001; Floyd et al ., 2005; Gresham, 2003). The social validity interview and questionnaire results in this study in dicated that teachers understood the principles of f unctional behavior assessment and found the training to be helpful. The measures also determined that the teachers found the de scriptive assessments helpful and minimally invasive. These social va lidity results are similar to those found in the studies that conducted dire ct evaluations (Murdock et al., 2005; Newcomer & Lewis, 2004; Sasso et al., 1992). The respondents indicated th at the functional analysis was more invasive and too contrived to be worthwhile. The ABC dire ct observations cond ucted by an outside observer (i.e., research assistan t) were judged to be the most valuable and acceptable way of determining the function of challenging behaviors. However, the sentiment expressed most frequently was that, in regard to these studentsÂ’ problematic behaviors, the respondents were willing to take extensive m easures to reduce the behaviors. One final aspect that is worth noting is that Alison, th e only participant teacher in a public school with multiple grades, had corresponding scores to the teachers in the universitysponsored setting. This finding is encourag ing because unlike a university sponsored early childcare facility, equivale nt social validity scores and interview responses suggest that these procedures could be used in more widespread ap plications in public schools. An area of future research will be to evaluate what effect withdrawing researcher influence, thereby putting the onus of assessment and implementation on school

PAGE 154

143 personnel, would have on perceptions of the FBA process and outcomes. There are a number of other important areas of future re search that can extend the findings of this study to continue to improve FBA processes and outcomes. Future Research Clearly, this study is only a beginning. Future research base d on comparisons of descriptive assessments with each other and with a functional analysis is necessary. Additionally, continuing to hone the methodology of functional analysis in more natural settings with increased teacher participati on will be another important extension of the existing research base. There are four areas of future research that could extend the results of this study and conti nue to evaluate the effectiven ess and efficiency of the FBA process in its widespread application within school settings. The four areas that will be discussed in this section are: (a) replicati ng this study to increase the total number of participants across sites and then systema tically manipulating ot her variables, (b) evaluating the characteristics of the direct descriptive observations (ABC), (c) linking results of different methodologi es to function-based interven tions, (d) refining functional analysis methodology by reducing researcher in fluence and measuring the effects of the process over time. As mentioned in the limitations section, a replication of this study is important to increase internal and external validity. Th e replication would add child and teacher participants making the results more robust. Because this study drew 75% of its sample from an early childhood center and 25% from a public school, it will be important to extend this study both within other early ch ildhood centers and in public schools. The replication could also use diffe rent types of schools, grade le vels, and classroom types to conduct the indirect and direct descriptive assessments and functional analyses. Also by

PAGE 155

144 having different individuals complete th e direct observations (i.e. teachers, paraprofessionals) this modifi cation could help determine if the high level of agreement between the ABC and the functional analyses was a result of this specific instrumentationÂ’s value or the result of an experienced, graduate-level trained researcher accurately identifying behavior functions. Th ese observations could be compared with the functional analyses and the level of agr eement could be compared with the current study. There are other variables that could be manipulated after c onducting replications across sites that add participants and contro l for the impact of specific personnel. The replication could also use more public school settings with more school staff involvement to determine the influence of conducting resear ch outside a university sponsored setting. This includes continuing examination of staffi ng responsibilities and teacher training. As indicated in Scott et al., (2004b) additional research into the role of collaborative teams is necessary as their value in producing reliable and valid FB As has not been established. The next step in replicating this study would be to systematically modify what descriptive assessments could be incorporated into the de scriptive assessment phase of the study including asking teachers what they be lieve the function is in more systematic ways (Calloway & Simpson, 1998; Cunningha m & OÂ’Neill, 2001; Yarbrough & Carr, 2000) or other descriptive assessments that ma y be in use in school settings. A scatter plot may be an important tool to use in fu ture research. A scatter plot is a direct observational tool that tracks the occurrence of behaviors according to times of day and/ or activities (Touchette, MacDonald, & Lange r, 1985). While it doe s not arrive at a hypothetical function, and therefor e was inappropriate for use in this study, it could be used in conjunction with other descriptive assessments to make them more effective

PAGE 156

145 (Symons, McDonald & Wehby, 1998). This future research would be a way to build on the apparent value of direct observation methodologies that th is studyÂ’s results indicate. Additionally, measuring the effects of levels of training in FBA methodology will be an important area of future research (Cone, 1997; Ervin et al., 2001). In consideration of evaluating different descriptive assessments, although the MAS and the FAI represent th e most widely used indirect desc riptive assessments (Floyd et al., 2005), educational settings may still employ a wide range of commercial instruments. This necessitates continuous empirical comparisons and validation as these instruments may have varying degrees of accuracy. This is relevant in consideration of the indirect descriptive assessment, the Problem Behavior Questionnaire (PBQ) that was ultimately excluded from these comparisons. While the discordant scoring taxonomy (i.e., no tangible or sensory options as possible functions) made one-to-one comparisons impossible for this study, specific practical co nsiderations also exis t. Individuals using the PBQ would be unable to assess if access to tangible items or sensory reinforcement served as the function of th e problem behavior. Both of these conditions have been empirically validated as possible functions of childrenÂ’s challenging behavior (Carr, 1977; Iwata et al., 1990; Repp & Munk, 1999). The next extension of this study is th e continued examination of the direct observation instrument (ABC). Building on its high level of agreement with the functional analyses in this study, future resear ch could focus on what attributes make this method effective and what modifications can be done to make it more efficient. The research assistant conducted eight 15-minute observations se ssions, for a total of two hours. Future research should determine what the effect of systematically reducing the

PAGE 157

146 total amount of time of th e observations has on the accu racy of the assessment procedures. This is important because in th e widespread application of FBA, assessment methods have to be as efficient as possibl e (Scott et al., 2004). Another extension of future research could examine the impact of lower rates of behavior and consequences on the overall accuracy of the assessment met hod. Finally, future research should further explore what other factors impact the high le vel of agreement between direct descriptive observations and the functional analysis outcomes including usi ng different direct observation instruments (i.e., the Functi onal Analysis Observation Form) (FAOF) (OÂ’Neill et al., 1997). Another extension of this research would be to examine the link between these assessment results and function-based interventi ons. Future research should examine how well these results translate into practical inte rventions in school se ttings. This research should monitor what effects focusing on pr imary functions have on reducing overall challenging behavior. It will also help determin e if the different assessment results could translate to different interventions across cont exts. In other words, if the FAI indicated that in specific settings (i.e., Circle Time ), challenging behavior is escape-maintained, could a different intervention be implemented in that setting as opposed to other settings where behavior is maintained by attention? This future area of research could focus on a decision-making system, (i.e., a flow chart or decision-making tree) that evaluates the comprehensive results of the assessment and li nks these results with empirically tested components of behavior intervention plans. Th ese interventions coul d then be evaluated for overall effectiveness, durability and social validity.

PAGE 158

147 Because functional analysis represents the “gold standard” for determining the function of challenging behavior it is vital to continue to make this procedure more amenable to school personnel in natural settin gs (Sasso et al., 2001). Although this study conducted functional analyses in natural sett ings with teachers acting as therapists, procedural integrity necessitated pronounced re searcher influence. Future research should focus on increasing the number of setti ngs that these assessments occur in, documentation of specific antecedents and setting events, and a greater degree of experimental control of settings and ant ecedents while measuring their impact on the consistency of assessments by different instru ments. Additionally, future research should focus on the process of training teachers to cond uct functional analyses in natural settings with students with high incidence disabilities or at risk for the development of behavioral disorders. Researchers should also monitor pr ocedural integrity, bu t not directly conduct the conditions. Measures of social validity should be used to monitor teachers’ perceptions of the process. Finally, probes could be used to evaluate the durability of this methodology within school settings and analys es of rates of challenging behavior and desired behaviors could determine the overall effect of functional analyses in school settings. Implications for Practice There are a number of impli cations of the result of th is study in terms of the application of FBA in schools and early chil dhood centers. First, as mentioned in the limitations section, because th ese assessment measures produce more information than just a primary function (e.g. setting events, secondary functions, and so forth), their practical applications exte nd beyond primary functions for schools. The main goal of school personnel using FBA methodology s hould be to produce a comprehensive

PAGE 159

148 behavior intervention plan (BIP), thus, ot her information from these assessments is relevant. However, this study indicates that if educators were to use multiple methods to evaluate function, these multiple methods coul d produce conflicting results. This could have a variety of effects. For example, educators could discard the results of one assessment over another. School personnel could attempt to formulate more complex intervention plans based on multiple functions across different contexts. Finally, school personnel could discard the lack of consistency across measures as indicative of a lack of validity and discount the w hole process entirely as being nothing more than a meaningless, bureaucratic chore. In this inst ance, it is reasonable to assume that schools may choose to use the most efficient a ssessment methods, (i.e., questionnaires and interview forms), in order to satisfy legal or school policy reform mandates (Sasso et al., 2001; Scott et al., 2004). Base d on the current study, the e nd result of this possible implication is more conflicting results becau se the more efficient methods, namely the questionnaire and interview, were more in consistent than the more time-intensive methods, namely the direct descriptive observations. Considering this possible implication in terms of the comparisons with the functional analysis results indicates lower levels of accurately determined functi ons; thus continuing th e potential cycle of FBA worthlessness. This is regrettable as co rrectly matching behavior al interventions to behavioral function is still viewed as the most promisi ng method for decreasing undesired behaviors (Iwata et al., 1 982/1994; Gresham, 2003; OÂ’Neil l et al., 1997; Sasso et al., 2001) The results of the comparisons between descriptive assessment outcomes and the functional analysis indicate that their use of indirect descriptive assessments to determine

PAGE 160

149 the primary function of challenging behavior may be imprudent. Because the indirect descriptive assessments were only accurate 5/12 (42%) of the time, educators would be more likely to determine the incorrect functi on than the correct function when using any single indirect desc riptive assessment. However, this study also underscores the importance of direct descriptive obser vations. Because the ABC had 4/4 (100%) agreements with the functional analysis results, consistently integrating direct descriptive assessments into FBA protocols appe ars to be a recommended practice. In regard to the use of functional analys is in natural setti ngs, this study further demonstrated the difficulty of balancing st rong internal validity, the hallmark of functional analysis (Sasso et al., 1992), and strong external validity, the generalizability of this methodology in applied settings. In this study, strong internal validity required continuous researcher involveme nt throughout the functional analysis process. Thus, it is difficult to conclude that these procedures could be effectively implemented in applied settings without researcher support or more exte nsive teacher training. However, this study did extend the existing FA literature by conducting the functional anal ysis in the classroom settings (unlike the majority of previous studies that conducted functional analyses in separate rooms away from the cl assroom). Of the eight studies that directly compared descriptive assessments with functional analysis, only Sasso et al. (1992) used the classroom setting to conduc t the functional analyses. Finally, there are three suggestions for indi viduals who are faced with continuing to produce FBAs, but want to increase the validity of the process. First, using direct observation methods, particular ly ABC forms for high inci dence behaviors may be a valuable way to increase the validity of the assessment. These ABC forms may be

PAGE 161

150 supplemented with the use of other direct observation methodologies (i.e., scatter plots). Second, educators should use caution with indire ct measures and use them in conjunction with these direct observation methods to eval uate antecedents and other possibly relevant information. Third, the use of functional anal ysis, or other experimental analysis, in widespread application does not appear advisable based on existing research. The limitations, technical demand and cost-benefit ratio do not support widespread use of this methodology at this time. Summary This study extends the continued investig ation of reliability and validity of the descriptive components of th e FBA process. By maintaining separation between the descriptive components, counter balancing their order, and conducting functional analyses within the classroom setting, the purpose of this inves tigation was to individually compare the results of three commonly used descriptive assessment methods with each other to determine reliability and with a functional analysis to determine construct validity. Descriptive assessmen ts including one checklist, a direct observation and an interview were completed for four young children. Teachers and researchers then completed functional analyses within the cla ssroom setting of the same participants. All results of the assessments were kept separa te from each other until completion. Teachers also completed a social validity questionnaire and interview. This study indicated that descriptive a ssessments had low consistency with each other, 12/36 (33%) agreements, and that the interview and checklist also had low agreement with the functional analyses with 3/8 (38%) agreements. However, the direct descriptive observations agreed with the functional analys es conclusions for all 4/4 (100%) of the participants. Results from social validity measures indicated that teachers

PAGE 162

151 felt methods of assessment were helpful, mini mally disruptive, and results would impact their behavioral interventi ons. The results va lidate direct desc riptive observation assessment methods and extend the literature by examining the use of functional analyses conducted within natural settings.

PAGE 163

152 APPENDIX A APPROVED IRB FORMS AND PARTICIPANT INFORMED CONSENT FORMS University of Florida Institutional Review Board 1. TITLE OF PROTOCOL: Improving Functional Behavior Assessments: Comparing Descriptive Methodologies w ith Functional Analysis 2. PRINCIPAL INVESTIGATOR(s): Peter Alter, Doctoral Candidate, Department of Special Education, P.O. Box 117050, G-315 Norman Hall, Gainesville, Florida 326117050, 392-0701 3. SUPERVISOR (IF PI IS STUDENT): Maureen Conroy, Associate Professor Department of Special Education, P.O. Box 117050, G-315 Norman Hall, Gainesville, Florida 32611-7050, 392-0701, mconroy@coe.ufl.edu 4. DATES OF PROPOSED PROTOCOL: From 12/01/05 to 12/01/06 5. SOURCE OF FUNDING FOR THE PROTOCOL: Unfunded 6. SCIENTIFIC PURPOSE OF THE INVESTIGATION: The purpose of this investigation is to examin ethe Functional Behavior Assessment (FBA) process by comparing four commonly used descriptiv e assessment results with each other to determine consistency and with a functional an alysis to determine construct validity. 7. DESCRIBE THE RESEARCH METHODOLOGY IN NON-TECHNICAL LANGUAGE Each participating teacher will nomi nate a child in his/her classroom who is exhibiting problem behavior s that interfere with their ability to learn. Descriptive assessments: Problem Behavior Questionnaire, Motivation Assessment Scale, Functional Assessment Interview and an Antecedent, Be havior, Consequence direct observation sheet (see attached examples) will be comple ted for four young children (ages 3-6) by the researcher with information supplied by the primary teacher as part of a functional behavior assessment. Researchers will then complete functional analyses of the same participants. The results of these two types of assessments will be compared to determine the validity of descriptive assessment methods. The children will be incl uded if they meet the following criteria: Children are between the ages of 3 and 6 years old. They are currently enrolled in a Day Ca re Center, Pre-kinderg arten, Kindergarten, or First Grade class (general education or special education) in north-central Florida.

PAGE 164

153 They have been identifie d through teacher nominations in that teachers have reported that these children are having probl em behaviors that ar e interfering with their learning and this problem warrants an FBA. A Total Problem Score of no less then 60 on the Child Behavior Checklist Teacher Reporting Form (CBCL-TRF) (Achenbach, 1992) Students will not be classified mentally retarded as determined by a cumulative file review. Consistent attendance at school, defined as 80% attendance or 4 out of 5 days, must be documented either thr ough school or teacher records. Recruitment of child and teacher participants will occur by approaching schools that are already participating in a Positive Be havior Support (PBS) program, which describes functional behavior assessment as a cornerstone to its program. The initial contact will be made to the principal or childcare director of the PBS school. Principals or childcare directors will be asked to recommend teachers who teach children with problem behaviors that necessitate a functional be havior assessment. Teachers will then be approached to participate in this research study. After teach ers sign teacher participation consent forms, they will make initial contact with parents/ caregivers about participating in this research study. If parent s/ caregivers would like to have their child to participate, then the teacher will contact the researcher. The researcher will then contact the parent and gain the necessary consent for th e childÂ’s particip ation in the study. After recruitment of four teacher and stude nt participant dyads has been completed, there are three phases of this investigation. In the first pha se, pre-assessment procedures will be conducted. Participating teachers will be asked to attend a one-hour training session that will include explanations of ope rational definitions of problem behavior, functions of behavior, Functional Behavior Assessment, functional analysis and the descriptive assessment instruments that will be used. The teacher will be asked to identify and operationally define the most fre quently occurring, problematic behavior demonstrated by the child, identify times or activ ities this behavior is likely to occur, and to identify a preferred classroom toy for the tangible condi tion of the functional analysis and a task that can be completed with an 85% success rate for the es cape condition of the functional analysis. Additionally, the invest igator will conduct a review of pertinent school records to determine consistent at tendance and average cognitive ability. The second phase of the procedure will cons ist of the completion of the descriptive assessments. The teacher will be asked to fill out the Problem Behavior Questionnaire, the Motivation Assessment Scale. A trained gr aduate assistant will interview teachers using the Functional Assessment Interview. Th e principal investigator will then conduct eight fifteen-minute direct observations of the student, during those times when problem behaviors are most likely to occur, in or der to complete the Antecedent-BehaviorConsequence instrument. In the third phase the principal investigator will conduct a functional analysis for each participating child. In this phase, we will manipulate five specific types of situations that may be maintaining the problem behavior in the classroom environment (1) attention (2) esca pe (3) tangible (4) alone (5) free play (control). During phase three, data will be collected using a real time, continuous recording system that calculates the frequenc y and duration of the problem behavior and the condition in which that behavior occurred and translates this data into rate of behavior per session. The condition in which the problem behavior occurs most frequently will be

PAGE 165

154 determined to be the primary function of the problem behavior. For both the direct descriptive observations and the functional analysis, obser vations will occur in the classroom during naturally occurring activities and the researchers will not identify to the class the target child included in the obser vation. Once all of the phases have been completed the different assessments will be compared for agreement or disagreement, and findings will be used for the princi pal investigatorÂ’s doctoral dissertation. 8. POTENTIAL BENEFITS AN D ANTICIPATED RISK: The goal of this study is to improve the efficacy of the Functional Behavior Assessment process by evaluating frequently used descriptive assessments for consistency and validity. There are no anticipated risks. Teachers will be instru cted to address problem behaviors through typical classroom intervention procedures. In the unlikely event that any type of the assessment procedures is judged to be elic iting excessive occurre nces of the problem behavior then the session will be terminated and teachers will be asked to respond to the target childÂ’s problem behaviors usi ng their typical classroom practices. 9. DESCRIBE HOW PARTICIPANT (S) WILL BE RECRUITED, THE NUMBER AND AGE OF THE PARTICIPANTS, A ND PROPOSED COMPENSATION (if any): Four teacher participants will be appro ached by the principal investigator via the school principal and will be given an explan ation of the investigation. The teachers will then be asked to nominate students whose pr oblem behavior is interfering with their learning. After the student participants have b een identified, the participantsÂ’ family will first be approached through th e teacher, if the parents/ caregiver is interested in participating, then the teacher will contact the principal investigator and the principal investigator will then meet w ith the parents/ caregiver to describe the study and obtain consent for participation A total number of 4 students will be sele cted and the ages of the participants will range from 3-6 years old. A $25 dollar gift card will be provided for all teacher and student particip ants who complete the project. 10. DESCRIBE THE INFORMED CONSENT PROCESS. INCLUDE A COPY OF THE INFORMED CONSENT DOCUMENT (if applicable). See Attached. Please use attachments sparingly. __________________________ Principal Investigator's Signature _________________________ Supervisor's Signature I approve this protocol for submission to the UFIRB:

PAGE 166

155 ____________________________ Dept. Chair/Center Director Date

PAGE 167

156 University of Florida Department of Sp ecial Education P.O. Box 117050/G-315 Norman Hall Gainesville, Florida 32611-7050 Caregiver/ Parent Form Consent Form for Improving Functional Behavior Assessments: Comparing Descriptive Assessment Methodologies with Func tional Analysis for Young Children The research study described below is being co nducted by University of Florida graduate students/staff under the supervision of Dr. Ma ureen Conroy (Associate Professor, Special Education Department). The purpose of this st udy is to evaluate different descriptive assessments that are frequently used in the completion of a Functional Behavior Assessment by comparing the resu lts of the different assessm ents with each other and with a functional analysis. Your childÂ’s teacher has indicated that you may be interested in having your child participate in this study. If you agree to have your child participate in the study, all three phases described below will be conducted in your childÂ’s typical classroom settings, which for your child may be the childcare cent er, preschool, or clas sroom during typical activities (recess, lunch, free time, center time) Your childÂ’s typical routine will not be disrupted and your child will not be removed from their classroom at any time during the study. The research study will be implemente d according to phases (as described below). Phase 1: The first phase consists of pr e-assessment procedures. We will review your childÂ’s cumulative school record to obt ain additional demographic information for research. We will conduct an interview with your childÂ’s teacher/ early childcare provider to determine what behavior is most problem atic and is interfering with your childÂ’s learning. In addition, we will ask your childÂ’s teacher what toys and activities your child prefers and what toys/activities/ or material s are difficult for your child. By identifying these materials, we can best assess your childÂ’s behavior. Phase 2: The second phase consists of an inte rview, a direct observa tion process, and the completion or two behavior checklists w ith your childÂ’s teacher or early childhood provider. We will conduct the interview and behavior checklists during times when the children are not in th e room, i.e., planning times, and directly observe your child in natural settings during typical activities. The interview and behavior checklists with the childcare provider/teacher will take approximately two hours and we will observe your child for approximately two hours. Your chil d will not be removed from the natural setting nor will his/her schedule be changed. Phase 3: The third phase involves an assessmen t technique called functional analysis. We will create five specific types of situations that may be maintaining your childÂ’s behavior in the classroom environment (1) attention (2 ) escape (3) tangible (4 ) alone (5) free play. This will help us to determine the possibl e reason for your childÂ’s behavior. In rare

PAGE 168

157 occasions, conducting functional analyses ma y increase your childÂ’s behavior. Although this is highly unlikely, if your childÂ’s beha vior or another childÂ’s behavior becomes threatening or violent at anytime during th is assessment period, we will terminate the session immediately and the teacher will be asked to follow typical intervention procedures for that setting. We will also ask the teacher or early chil dhood provider to complete acceptability rating forms after all phases of assessment in order to give his/her opinion on how he/she viewed the use of these procedures. We will be videotaping the sessions in the third phase in your childÂ’s classroom setting. This will be done to record the behaviors that occur during our sessions. These videotapes will be shown only under our direct ion to persons collecting and scoring the data, to you or the teacher or early childhood prov ider if), to students or other research at the University of Florida, and at professiona l presentations outside of the University for research or training purposes only. Dr. Conroy will strictly monitor the confidentiality of these videotapes. However, you should be aw are that the showing of these videotapes might result in others being able to re cognize your child/childÂ’s teacher or early childhood provider. In addition, data on your ch ildÂ’s participation will be used for training or as part of a publication, but your childÂ’s name will not be used for data publication or presentation and records will be kept confidential to the extent provided by law. As required by the Alachua County School Board, the final res earch report will be submitted to the Board for review (e.g., conferen ce abstract or research publication) and your childÂ’s name will not be used. Each phase of the study will take approximate ly 2 weeks. Observations will take place approximately 2 to 3 days per week for approximately one hour each time. We will be providing some preliminary recommendations at the conclusion of the final phase of assessment. Participation or nonparticipation in this study will not affect your childÂ’s placement in any educational programs. You and your child ha ve the right to wit hdraw consent at any time without consequence. There are no known risks for participati on in this study; however, your child may continue to demonstrate some of his/her existing challenging behaviors during the assessment phases of the study. These behavior s may be the same behaviors that your child has been demonstrating in the natural setting and the childcare provider/teacher may implement typical natural setting procedures if the behaviors occur. Possible benefits of participation include identific ation of current reasons fo r your childÂ’s inappropriate behavior, improvement in your childÂ’s social behavior, and your childÂ’s teacher or early childhood provider. A one-time compensation in the form of a $25 gift card will be awarded to you upon completion of the study. For your information, if your child is injured during assessment or treatment and the investigator is at fault, the University of Florida Board of Trus tees, and the State of

PAGE 169

158 Florida shall be liable only as provided by law. Appropriate compensation for injury may be sought by contacting the Insurance Coordi nator at 107 Tigert Hall, University of Florida, Gainesville, FL 32611, Tel: (352) 392-1325. Questions or concerns about research participantÂ’s rights may be directed to the University of Florida Institutional Review Board, Box 112250, the University of Florida, Gainesville, FL 32611, Tel: (352) 392-0433. ************************************************************************ I have read the procedures described above for this project. I have re ceived a copy of this statement and I agree to allow my child to pa rticipate. I am free to ask questions or to express any concerns that I ha ve about the project. I am fr ee to withdraw consent at any time, and this will have no effect on other se rvices provided to my child. I understand that my childÂ’s teacher or early childhood provider will sign a separate form. ___________________________________ ______________________________ Signature 2nd Caregiver/Witness ___________________________________ ______________________________ Date Date ___________________________________ Relationship to Participant I give consent for my child, ___________________________________ to be videotaped during sessions to serve as a re cord of his/her responses and fo r the purposes stated in this consent. ___________________________________ ___________________________________ Signature 2nd Caregiver/Witness ___________________________________ ___________________________________ Date Date ___________________________________ Relationship to Participant ************************************************************************ Questions about this research project may be directed to: Dr. Maureen Conroy, (352) 392-0701 ext. 245.

PAGE 170

159 University of Florida Department of Sp ecial Education P.O. Box 117050/G-315 Norman Hall Gainesville, Florida 32611-7050 Teacher/ Early Childhood Care Provider Form Consent Form for Improving Functional Be havior Assessments: Comparing Descriptive Assessment Methodologies with Func tional Analysis for Young Children The research study listed below is being c onducted by University of Florida graduate students/staff under the supervision of Dr. Ma ureen Conroy (Associate Professor, Special Education Department). The purpose of this st udy is to evaluate different descriptive assessments that are frequently used in the completion of a Functional Behavior Assessment by comparing the resu lts of the different assessm ents with each other and with a functional analysis. This child’s parent will also need to provide consent for this child to participate in this research project. If you agree to participat e with this child in this projec t, all three phases described below will be conducted in your classroom (e.g., childcare center or preschool classroom) during typical activities (recess, lunch, fr ee time, center time) without removing or disrupting this child from his/her typical routine. Phase 1: The first phase consists of pre-as sessment procedures. We will ask you to attend a one-hour one-on-one training session to review basic principles of this study and complete a Child Behavior Checklist in regard to this student. We will review this child’s cumulative school record to obtain additional information for research (specifically—IEP or IFSP, cognitive assessment, medical, beha vior, and attendance information). We will schedule and conduct a intervie w with you, at your convenience, to determine what behavior is most problematic in interfering with learni ng, to determine when problem behaviors are most likely to occur, to es tablish a preferential toy for the tangible condition of the third phase, and determine an a ppropriate task for the escape condition of the third phase. Phase 2: The second phase consists of an inte rview, a direct observa tion process, and the completion or two behavior checklists w ith you. We will conduct the interview and behavior checklists during times that are convenient to you wh en the students are not in the room, i.e., planning times and directly observe your child in natural settings during typical activities. The interview and behavior checklists with you will take approximately two hours and we will observe the child for a pproximately two hours. The child will not be removed from the natural setting nor will his/her schedule be changed. Phase 3: The third phase involves an assessmen t technique called functional analysis. We will create five specific types of situations that may be maintaining the problem behavior in the classroom environment (1) attention (2 ) escape (3) tangible (4) alone (5) free play (control). This will be done in your classroom at times that you identify as problematic. This will help us to determine the possi ble reason for the child’s behavior. Although

PAGE 171

160 highly unlikely, if the childÂ’s behavior or a nother childÂ’s behavior becomes threatening or violent at anytime during this assess ment period, we will terminate the session immediately and you will be asked to follow t ypical intervention proce dures for that child and setting. These sessions will be videotap ed in order to determine when problem behaviors occur most frequently. Finally, we will also ask you to complete an acceptability rating form after all phases of assessment in order to give your opinion on how you viewed the use of these procedures in your classroom settings. We will be videotaping the sessions in the th ird phase in the classroom. This will be done to record the behaviors that occur during our sessions Th ese videotapes will be shown only under our direction to persons collecting an d scoring the data, to you or the parents if requested (e.g., as a training tape for treatment), to student s or other research at the University of Florida, and at professional presentations outside of the University for research or training purposes only. Dr. Conroy will strictly monitor the confidentiality of these videotapes. However, you should be aw are that the showing of these videotapes might result in others being ab le to recognize you. In additi on, data on the ta rget childÂ’s participation will be used for training or as part of a publication, but the name will not be used for data publication or presentation and records will be kept confidential to the extent provided by law. As required by th e Alachua County School Board, the final research report will be submitted to the Board for review (e.g., conference abstract or research publication) and your childÂ’s name will not be used. Each phase of the study will take approximate ly 2 weeks. Observations will take place approximately 2 to 3 days per week for approximately one hour each time. Participation or nonparticipation in this study wi ll not affect the child Â’s placement in any educational programs. You have the right to withdraw consent at any time without consequence. There are no known risks for pa rticipation in this study; how ever, the target child may continue to demonstrate some of his/her existing challenging during the assessment phases of the study. These behaviors may be th e same behaviors that the child has been demonstrating in the classroom and you shoul d implement typical procedures if the behaviors occur. Possible benefits of part icipation include identification of current reasons for the childÂ’s inappropriate behavi or, and you will receive behavioral assessment and intervention training individualiz ed for the target childÂ’s needs. Results of this study will be shared with you and at the end of the study. No information will be shared with anyone other than you and the parent unless you authorize this to occur. However, if the parent requests inform ation on the childÂ’s progress in the project, this information will be provided wit hout additional consent from you. You cannot otherwise request the release of other data on this child without additional permission from the parent.

PAGE 172

161 For your information, if you are injured during as sessment and the investigator is at fault, the University of Florida Board of Trustees, and the State of Florid a shall be liable only as provided by law. Appropriate compensation for injury may be sought by contacting the Insurance Coordinator at 107 Tigert Hall, Un iversity of Florida, Gainesville, FL 32611, Tel: (352) 392-1325. Questions or concerns abou t this research participantÂ’s rights may be directed to the University of Flor ida Institutional Revi ew Board, Box 112250, the University of Florida, Gain esville, FL 32611, Tel: (352) 392-0433. ************************************************************************ I have read the procedures described above for this project. I have re ceived a copy of this statement and I agree to particip ate in this project. I am free to ask questions or to express any concerns that I have about the project. I am free to wit hdraw consent at any time, and this will have no effect on other services pr ovided to this child. I understand that the childÂ’s parent has signed a separate consent form. __________________________________ _____________________________________ Signature 2nd Early Childhood Provider/Witness __________________________________ ____________________________________ Date Date __________________________________ Relationship to Participant I understand that his childÂ’s parent has given consent for __________________________ to be videotaped during sessions and to serve as a record of his/he r responses and for the purposes stated in this consent. I unders tand that established school guidelines for obtaining video consent for the classroom will be followed. __________________________________ __________________________________ Signature 2nd Early Childhood Provider/Witness __________________________________ __________________________________ Date Date __________________________________ Relationship to Participant ************************************************************************ Questions about this research project may be directed to: Dr. Maureen Conroy, (352) 392-0701 ext. 245.

PAGE 173

162 APPENDIX B FLOW CHART OF TEACHER AND CHILD PARTICIPANT RECRUITING PROCESS IF NO IF NO IF NO SCHOOL OR EARLY CHILDCARE CENTER RESEARCHER CONTACTS PARENTS FOR CONSENT FOR PARTICIPATION TEACHER CONTACTS RESEARCHER TO CONTACT PARENTS TEACHERS CONTACT PARENTS TO PARTICIPATE IN RESEARCH STUDY RESEARCHER CONTACTS PRINCIPAL/ CHILDCARE DIRECTOR TO RECOMMEND TEACHERS RESEARCHER CONTACTS TEACHER TO PARTICIPATE IN RESEARCH STUDY CHILD PARTICIPANT SCREENING PROCESS TEACHER TRAINING

PAGE 174

163 APPENDIX C STUDENT PARTICIPANT CHILD BEHAVIOR CHECKLIST (CBCL) SCORES Student Total Scores and Pe rcentile for CBCL (1.5-5) Participants Emotionally Reactive Anxious/ Depressed Somatic Complaints WithdrawnSleep Problems Attention Problems Aggressive Behavior Ramon 15 >97 percentile 13 >97 percentile 3 79 percentile 7 >97 percentile 4 73 percentile 7 >97 percentile 32 >97 percentile Anthony 10 >97 percentile 7 95 percentile 1 <50 percentile 5 96 percentile 9 >97 percentile 5 89 percentile 33 >97 percentile Jimmy 4 90 percentile 2 62 percentile 0 <50 percentile 2 <50 percentile 0 >50 percentile 3 58 percentile 26 96 percentile Student Scores for CBCL (6-18) Participant Anxious/ Depressed Withdrawn Depressed Somatic Complaints Social Problems Thought Problems Attention Problems RuleBreaking Behavior Aggressive Behavior Greg 4 76 percentile 4 90 percentile 2 89 percentile 7 97 percentile 7 >97 percentile 31 93 percentile 6 95 percentile 32 >97 percentile

PAGE 175

164 APPENDIX D INSTRUMENTS Pre-experimental Assessment Checklist 1) Target behavior operationally defined ________________________________________________________________ __________________________________________________________________ 2) Preferred toy _________________________________________________________________ 3) Task that can be completed 30% of the time __________________________________________________________________ 4) Record Review Completed ________ 5) CBCL Score________________ 6) Times or Activities Problem Behavior Is Likely To Occur __________________________________________________________________ __________________________________________________________________

PAGE 176

165 APPENDIX E SOCIAL VALIDITY MEASURES Social Validity Questionnaire Teacher: _________________________________ Date: ________________________________ School/ Center: __________________________________ Age level of child: 3 years 4 years 5 years 6 years (Circle one) Please complete the items below by circ ling the number under the question that best fits how you feel about th e assessment techniques conducted. How time-consuming were the checklists and interview? Not at all Very 1 2 3 4 5 How helpful was the training on F unctional Behavior Assessment? Not at all Very 1 2 3 4 5

PAGE 177

166 166 How helpful was the Functional Assessment Interview? Not at all Very 1 2 3 4 5 How helpful was the Motivation Assessment Scale? Not at all Very 1 2 3 4 5 How disruptive was the direct observation descriptive assessment (ABC)? Not at all Very 1 2 3 4 5 How disruptive was the functional analysis? Not at all Very 1 2 3 4 5 Would you be willing to let another of your students pa rticipate in this study? Not at all Very 1 2 3 4 5

PAGE 178

167 167 How valuable was it to determine the function of the problem behavior? Not at all Very 1 2 3 4 5 Any additional comments: Social Validity Interview 1. HAVE YOU EVER COMPLETED A FUNCTIONAL BEHAVIOR ASSESSMENT BEFORE? 2. HOW WOULD YOU DEFINE F UNCTION (OF A BEHAVIOR)? 3. WHAT DID YOU THINK OF THE PROBLEM BEHAVIOR QUESTIONNAIRE? WHAT WOULD YO U HAVE CHANGED ABOUT IT? 4. WHAT DID YOU THINK OF THE MO TIVATION ASSESSMENT SCALE? WHAT WOULD YOU HAVE CHANGED ABOUT IT? 5. WHAT DID YOU THINK OF THE FUNCTIONAL ASSESSMENT INTERVIEW? WHAT WOULD YOU HAVE CHANGED ABOUT IT? 6. WHAT DID YOU THINK OF THE CO MPLETED DIRECT OBSERVATION WORKSHEETS? WERE THEY ACCU RATE OR INACCURATE? WHY? 7. WHAT DID YOU THINK OF TH E FUNCTIONAL ANALYSIS?

PAGE 179

168 168 8. WHAT DO YOU BELIEVE THE FUNC TION OF (THE PARTICIPANT CHILD) THE BEHAVIOR IS? REVIEW RESULTS 9. DO YOU AGREE WITH THE RE SULTS OF THE DIFFERENT ASSESSMENTS? 10. DO YOU FEEL LIKE YOU ALREADY KNEW THIS INFORMATION? IF SO HOW DID YOU USE TH E INFORMATION? 11. CONSIDERING (1) THE AMOUNT OF TIME EACH ASSESSMENT TOOK, (2) THE AMOUNT OF EFFORT EACH ASSESSMENT REQUIRED AND (3) THE FUNCTION EACH ASSESSMENT CO NCLUDEDWHAT ASSESSMENT METHOD DO YOU BELIEVE IS THE BEST? WHY? 12. WILL THIS INFORMATION CHA NGE HOW YOU STRUCTURE YOUR BEHAVIOR MANAGEMENT SYSTEM? HOW? 13. WILL YOU USE INFORMATION REGARDING FUNCTION TO CHANGE INTERVENTION STRATEGIES? HOW? 14. IS THERE ANYTHING ELSE YO U WOULD LIKE TO ADD?

PAGE 180

169 APPENDIX F TEACHER TRAINING OUTLINE What is a problem behavior? In not so many words a problem behavior is a behavior that is a problem. It may be a problem either for the indi vidual demonstrating the behavi or or for the people around the individual, in this case the teacher and/ or the other students. Problem behavior might include any of the following results: cause phys ical or emotional harm to the child or another person cause damage to property di srupt classroom learning either for the individual or others Common problem behaviors incl ude but are not limited to: 1. Disruptionincluding yelling, talking out of turn, crying, and noise making 2. Off-taskrefusing to complete work or participate in activities 3. Noncompliancerefusing to follow di rection, intentionally slow responses, ignoring adult directives What is an operational definition? Operational definitions of problem be haviors must have three components: 1. precisethe behaviors must described be so that different people can read a description of the behavior and interp ret it the same way (the Stranger Test) 2. positiveThe Dead ManÂ’s Testif a d ead man can do the behavior that is being described then it does not pass the Dead ManÂ’s test

PAGE 181

170 170 3. distinct(have a clear beginning and e nding) so that the behavior can be counted (Kaplan, 1995) What is the function of a behavior? 1. The nature of the consequence that reinfor ces a target behavior is considered that behavior’s function In other words, if the reinforc ing consequence of a behavior is gaining attention from the teacher, then the function of the behavior is teacher attention. 2. Researchers typically view function in te rms of one of three classes: positive reinforcement either through a tangible objec t or attention; negative reinforcement through escaping or avoiding an unpleasant s ituation; or intrinsic reinforcement wherein the behavior is hypot hesized to stimulate an in trinsic event (Alberto & Troutman, 2003). 3. For practitioners, the function is “a statement explaini ng why the student engages in this behavior.” 4. The difference between these two interpreta tions is subtle but critical. Function refers to the “purpose” of the behavior or how it helps the students to meet his or her needs, in terms of one of the three supporting consequences. 5. While when a practitioner answers the que stion ‘why does the student engage in this behavior?’ there are a host of responses. For exampl e, in a practitioner might identify ‘unmet basic needshungry, tired’ as a function of behavior, however this explanation provides a reason why but not a f unction. If the statement is correct then the function might be to access sleep or food, or to escape activities that compete with

PAGE 182

171 171 a search for food or sleep. The function of a behavior is often defined as ‘why the behavior occurs’ What is a Functional Behavior Assessment? 1. .The FBA is a process in which informa tion is gathered regarding a specific behavior or set of behaviors in order to determine whether that behavior has any predictable relationship with the en vironment (O’Neill et al., 1997). 2. The environment is comprised of all factors that came before (antecedents) and after (consequences) the behavior. 3. A relationship exists when these en vironmental factors both predict the behavior and explain why it is maintaine d. Identification of these relationships helps us to understand why behavior occu rs, or how behavior is reinforced and maintained by the specific environmenta l consequences (i.e., functions) they elicit. 4. The purpose of an FBA is to understa nd behavior so th at individualized interventions may be designed to replace undesirable behavior with appropriate alternatives that efficiently obtain th e functions of the problem behavior. What is a functional analysis? A functional analysis is a procedure used to evaluate possible functions of a behavior by manipulating what occurs after the problem behavior takes place. This is different then the other assessments that schools often use, be cause we are actively changing the environment to test what cons equences support this behavior. For this research study, we have five possible conse quences that will be used to determine the function of the problem behavior (1) attenti on (2) escape (3) tangible object (4) alone (5)

PAGE 183

172 172 free play. We will systematically apply each of these conditions after the behavior takes place and the condition that causes the behavior to occur most frequently is the function of the behavior. We will be doing this during regular classroom times but try not to disrupt the classroom environment. I can s upply you with additional resources if this procedure is still confusing. Watch FA Example Video What are the PBQ, MAS, FAI and ABC? The Problem Behavior Questionnaire (PBQ ) and the Motivation Assessment Scale (MAS) are behavior checklists that we are going to ask you to fill out. They are fairly briefthe PBQ has 15 questions and the MAS ha s16 questions. Both checklists ask you to circle a number between 0-6. These assessments will be scored to determine a hypothetical function of the behavior. Here is what these forms actually look like (show forms). The Functional Assessment Interview (FAI) is an interview that takes anywhere between 45-90 minutes. It has 11 sections and asks a variety of questions about behaviors and environmental factors before and after the behaviors. The Antecedent-Behavior-Consequence (ABC) direct observation sheet provides a method to directly observe behavior in natu ral environments. A graduate assistant will simply watch the target child and record what happened before and after the target behavior. Again, we will be doing this duri ng regular classroom times but try not to disrupt the classroom environment.

PAGE 184

173 APPENDIX G FUNCTIONAL ANALYSIS PROCEDURES Preliminary Steps Prior to each FA session the follo wing procedures will occur: Before going to the school/ daycare site the principal investigator will test the video camera and review FA conditions and procedures Review operational definiti on of target behavior Review order of conditions (the order of conditions are counterbalanced and will varythe task analysis shoul d be adjusted accordingly) Review specific procedure for each condition Obtain materials for all conditions: free play, attention, escap e and tangible condition. Arrange room so that the area for the f unctional analysis is defined. It is not completely separate, but peer and target child movement and access can be controlled and/ or restricted. Set up video camera in unobtrusive location to capture as much of this area as possible Teacher and principal investigator review audible cue for non-contingent attention and end of session.

PAGE 185

174 174 Review non-contingent attentionthe teacher will make statements directed toward participant child (not praise or redirection) “I see you’re using th e red crayon.” “You’re wearing a blue shirt, today.” Set up free play condition with neutral activity e.g. coloring in a coloring book. Free play condition Target child and peers ar e moved to defined area. Stopwatch and video camera are turned on Teacher says “You can play with whatev er you want. I’ll tell you when to stop.” Teacher is in defined area 3-6 feet from child and offers non-contingent attention. Occurrences of the problem behavior will be ignored. If the target child attempts to access more attention or leave assi gned area they will be gently redirected back to the activities. After five minutes have elapsed, the video camera is turned off and the teacher will take a 2 minute break: o Review next condition procedure (attention) o Clean up and set neutral activity o Put magazine or catalog in area for teacher to “look bus y” during attention condition Attention condition Target child and peers are in defined area Stopwatch and video camera are turned on Teacher says “You can ______(neutral activity, e.g. color in the coloring book). I’ll tell you when to stop.”

PAGE 186

175 175 Non-contingent attention is not provided. Teacher looks busy reading magazine 3-6 feet away. Attention is provided contingent on the de monstration of the problem behavior. At that time the teacher says “Stop doing that You shouldn’t do that because it is against the rules. You are not followi ng directions when you do that.” These verbal statements are paired with the teacher coming close to the student and putting a hand on their back. This same response is repeated upon each demonstration of the problem behavior After five minutes have elapsed the video camera is turned off and the teacher will take a 2 minute break: o Review next condition procedure (tangible) o Clean up o Put tangible item in a place where it is readily accessible Tangible condition Target child and peers are in defined area Stopwatch and video camera are turned on Teacher says “You can play with (prefe rred item). I’ll tell you when to stop.” Teacher is in defined area 3-6 feet from child. Principal investigator signals teacher ever y thirty seconds to offer non-contingent attention. The tangible item or activity is provided initially for thirty seconds and then presented upon demonstration of the problem behavior. The teacher says “ You can play with _____for little while.” After thir ty seconds, the item is removed. This same response is repeated upon each demonstration of the problem behavior

PAGE 187

176 176 Principal investigator signals teacher ever y thirty seconds to offer non-contingent attention. Teacher is in defi ned area 3-6 feet from child. After five minutes have elapsed the video camera is turned off and the teacher will take a 2 minute break: o Review next condition procedure (escape) o Clean up put difficult task in place Escape condition Target child and peers are in defined area Stopwatch and video camera are turned on Teacher says “I would like you to work on _______. I know it may be difficult but I want you to try your best. If you n eed a break, I will give you a break“ Principal investigator signals teacher ever y thirty seconds to offer non-contingent attention. Teacher is in defi ned area 3-6 feet from child. Upon demonstration of the problem behavior, the teacher will say “ Time to take a break.” At this point, the ta sk will be removed for thirty seconds and then returned. This same response is repeated upon each demonstration of the problem behavior After five minutes have elapsed the video camera is turned off and the teacher will take a 2 minute break: o Review next condition procedure (alone) o Clean up and remove activity Alone condition Target child is in defined area (no peer access) Stopwatch and video camera are turned on

PAGE 188

177 177 Teacher does not offer any directions and does not stay in the area Preferred item and neutral activity are available. Problem behavior is ignored. It may be necessary to shorten th is session (Peck, Sasso, & Stambaugh, 1998) Post Session Steps Classroom is returned to typical set-up Brief interview with teacher for ge neral feedback on FA procedures Monitor target child for increases in probl em behavior as a result of FA conditions

PAGE 189

178 APPENDIX H FINAL SCORING SUMMARY CHART Ramon Instrument Function identified Agreement with Other Descriptive Assessments Agreement with Functional Analysis MAS Tangible FAI No FAI Tangible MAS No ABC Attention None Yes FA Attention ABC Anthony Instrument Function identified Agreement with Other Descriptive Assessments Agreement with Functional Analysis MAS Tangible ABC Yes FAI Escape None No ABC Tangible MAS Yes FA Tangible MAS, ABC Jimmy Instrument Function identified Agreement with Other Descriptive Assessments Agreement with Functional Analysis MAS Tangible ABC Yes FAI Attention None No ABC Tangible MAS Yes FA Tangible MAS, ABC Greg Instrument Function identified Agreement with Other Descriptive Assessments Agreement with Functional Analysis MAS Sensory None No FAI Attention ABC Yes ABC Attention FAI Yes FA Attention FAI, ABC

PAGE 190

179 LIST OF REFERENCES Achenbach, T.M. (1992). Manual for the Child Behavior Checklist/1.5-5, 6-12 and 1992 Profile Burlington, VT: University of Ve rmont Department of Psychiatry. Arndorfer, R.E., & Miltenberger, R.G. (1993). Functional assessment and the treatment of challenging behavior: A review w ith implications for early childhood. Topics in Early Childhood Special Education, 13 82-105. Arndorfer, R.E., Miltenberger, R.G., Wost er, S.H., Rortvedt, A.K., & Gaffaney, T. (1994). Home-based descriptive and experi mental analysis of problem behaviors in children. Topics in Early Childhood Special Education, 14 64-88. Baer, D.M., Wolf, M.M., & Risley, T.R. ( 1968). Some current dimensions of applied behavior analysis Journal of Applied Behavior Analysis,1 91-97. Belfiore, P.J., Browder, D.M. & Lin, C.H. (1993). Using descriptive and experimental analyses in the treatment of self-injurious behavior. Education and Training in Mental Retardation and Devel opmental Disabilities, 28 57-65. Bijou, S.W., Peterson, R.F., Ault, M.H. (1968) A method to integrate descriptive and experimental field studies at the leve l of data and empirical concepts. Journal of Applied Behavior Analysis, 1 175-191. Blair, K.S., Umbreit, J., & Bos, C.S., (1999) Using functional assessment and childrenÂ’s preferences to improve the behavior of young children with behavioral disorders. Behavioral Disorders, 24 151-166. Blair, K.S., Umbreit, J. & Eck, S. (2000). Analysis of multiple variables related to a young childÂ’s aggressive behavior. Journal of Positive Be havior Interventions, 2 33-39. Boyajian, A.E., DuPaul, G.J., Handler, M.W., Eckert, T.L. & McGoey, K.E. (2001). The use of classroom-based brief functional analysis with preschoolers at-risk for Attention Deficit Hyperactivity Disorder. School Psychology Review, 30 278294. Broussard, C.D., & Northrup, J. (1995). An approach to functional assessment and analysis of disruptive behavior in regular education classrooms. School Psychology Quarterly, 10 151-164.

PAGE 191

180 Broussard, C. D., & Northrup, J. (1997). The use of functiona l analysis to develop peer interventions of disruptiv e classroom behavior. School Psychology Quarterly, 12 65-76. Calloway, C.J., & Simpson, R.L. (1998). Deci sions regarding func tions of behavior: Scientific versus informal analysis. Focus on Autism & Other Developmental Disabilities, 13 167-176. Carr, E.G. (1977). The motivation of self-i njurious behavior: A review of some hypotheses. Psychological Bulletin, 84 800-816. Chandler, L.K., & Dahlquist, C.M. (2002). Functional assessment: St rategies to prevent and remediate challenging behavior in school settings. Upper Saddle River, NJ: Merrill Prentice Hall. Chandler, L.K., Dahlquist, C.M., Repp, A.C., & Feltz, C. (1999). The effects of teambased functional assessment on the behavior of students in classroom settings. Exceptional Children, 66 101-129. Clarke, S., Dunlap, G., Foster-Johnson, L., Ch ilds, K. E., Wilson, D., White, R., & Vera, A. (1995). Improving the conduct of stude nts with behavioral disorders by incorporating student interests into curricular activities. Behavioral Disorders, 20 221-237. Cone, J.D. (1997). Invited essay: Issues in f unctional analysis in behavioral assessment. Behaviour Research and Therapy, 35 259-275. Conroy, M.A., Clark, D., Gable, R.A., & Fox, J. J. (1999). Building competence in the use of functional behavior assessment. Preventing School Failure,43 (4), 140-144. Conroy, M.A., Dunlap, G., Clarke, S. & Alter, P.J. (2005) A descri ptive analysis of positive behavioral intervention research with young children. Topics in Early Childhood Special Education, 25 157-166. Conroy, M.A., Fox, J.J., Bucklin, A., & Good, W. (1996). An analysis of the reliability and stability of the Motiva tion Assessment Scale in assessing the challenging behaviors of persons with developmental disabilities. Education and Training in Mental Retardation and Devel opmental Disabilities, 31 243-250. Conroy, M.A., Fox, J.J., Crain, L., Jenkins, A. & Belcher, K. (1996). Evaluating the social and ecological validity of analog assessment procedures for challenging behavior in young children. Education and Treatment of Children, 19 233-256.. Conroy, M.A., & Stichter, J.P. (2003). The a pplication of anteceden ts in the functional assessment process. The Journal of Special Education, 37 15-26.

PAGE 192

181 Cooper, L.J., & Harding, J. (1993). Exte nding functional analysis procedures to outpatient and classroom settings for children with mild disabilities. In J. Reichle & D.P. Wacker (Eds.), Communication alternatives to challenging behavior: Integrating functiona l assessment and intervention strategies (pp.41-63). Baltimore: Paul H. Brookes. Crawford, J., Brockel, B., Schauss, S., & Miltenberger, R.G. (1992). A comparison of methods for the functional assessm ent of stereotypic behavior. Journal of the Assosciation for Persons with Severe Handicaps, 17 77-86. Cunningham, E., & OÂ’Neill, R.E. (2000). A comparison of results of functional assessment and analysis methods with young children with autism. Education and Training in Mental Retardation and Developmental Disabilities, 35 406-414. Doggett, R.A., Edwards, R.P., Moore, J.W., Tingstrom, D.H., & Wilczinski, S.M. (2001). An approach to functional assessment in general education settings. The School Psychology Review, 30 313-328. Dunlap, G., dePerczel, M.,Clarke, S., W ilson, D.,Wright, S., White, R.,& Gomez, A. (1994). Choice making to promote adaptive behavior for studen ts with emotional or behavioral challenges. Journal of Applied Behavior Analysis, 27 505-518. Dunlap, G., & Kern, L. (1993). Assessment and intervention for children within the instructional curriculum. In J. Reichle & D.P. Wacker (Eds.), Communication alternatives to challenging behavior: Integrating functional assessment and intervention strategies (pp. 177-204). Baltimore: Paul H. Brookes. Dunlap, G., Kern-Dunlap, L., Clarke, S., & Robbins, F.R. (1991). Functional assessment, curricular revision, and severe behavior problems. Journal of Applied Behavior Analysis, 24 387-397. Durand, V.M., & Crimmins, D.B. (1992). The Motivation Assessment Scale (MAS) administration guide. Topeka, KS: Monaco. Ellingson, S.A., Miltenberger, R.G., Stricker, J., Galensky, T.L. & Garlinghouse, M. (2000). Functional assessment and interv ention for challenging behaviors in the classroom by general classroom teachers. Journal of Positive Behavior Interventions, 2 85-98. Ellis, J., & Magee, S.K. (1999). Determinati on of environmental correlates of disruptive classroom behavior: Integration of func tional analysis into the public school assessment process. Education and Treatment of Children, 22 291-317. Ervin, R.A. Ehrhardt, K.E., & Poling, A. ( 2001). Functional assessment: Old wine in new bottles. School Psychology Review, 30 173-179.

PAGE 193

182 Ervin, R.A., Radford, P.M., Bertsch, K., Pipe r, A.L., Ehrhardt, K.E. & Poling, A. (2001). A descriptive analysis and critique of the empirical literature on school-based functional assessment. School Psychology Review, 30 193-211. Floyd, R.G., Phaneuf, R.L., & Wilczynski, S.M. (2005). Measurement properties of indirect assessment methods for functiona l behavioral assessment: A review of research. School Psychology Review, 34 58-73. Fox, J.J., Conroy, M.A., & Heckaman, K. (1998). Research issues in functional assessment of the challenging behavior s of students with emotional and behavioral disorders. Behavioral Disorders, 24 26-33. Fox, L., Dunlap, G., Powell, D. (2002). Young children with challenging behavior: Issues and considerations for behavior support. Journal of Positive Behavior Interventions, 4 (4), 208-217. Grandy, S.E., & Peck, S.M. (1997). The use of functional assessment and selfmanagement with a first-grader. Child and Family Behavior Therapy, 19 (2), 2943. Gresham, F.M., Watson, S.T., Skinner, C.H. (2001). Functional Behavioral Assessment: Principles, Procedures, and Future Directions. School Psychology Review, 30 156-173. Gresham, F.M. (2003). Establishing the te chnical adequacy of functional behavior assessment: Conceptual and measurement challenges. Behavioral Disorders, 28 282-298. Gresham, F.M., McIntyre, L.L. Olson-Tinker, H., Dolstra, L. McLaughlin, V., Van, M. (2004). Relevance of functional behavioral assessment research for school-based interventions and positiv e behavioral support. Research in Developmental Disabilities, 25 19-37. Hagopian, L.P., Fisher, W.W., Thompson, R.H. Owen-DeSchryver, J., Iwata, B.A., & Wacker, D.P. (1997). Toward the devel opment of structured criteria for interpretation of functional analysis data. Journal of Applied Behavior Analysis, 30 313-326. Hanley, G.P., Iwata, B.A., & McCord, B.E. (2003). Functional analysis of problem behavior: A review. Journal of Applied Behavior Analysis, 36 147-185. Harding, J., Wacker, D.P., Cooper, L.J., As mus, J., Jensen-Kovalan, P., & Grisolano, L.A., (1999). Combining descriptive and experimental analyses of young children with behavior problems in preschool settings. Behavior Modification, 23 316333.

PAGE 194

183 Heckaman, K., Conroy, M.A., Fox, J. & Ch ait, A. (2000). Functional assessment intervention research on students with or at risk for emotional and behavioral disorders in school settings. Behavioral Disorders, 25 196-210. Individuals with Disabilities Educa tion Act Amendments of 1997, 20 U.S.C. 1415 (1997). Individuals with Disabiliti es Education Act Amendments of 2004, 11 Stat. 37 U.S.C. Section 1401 (2004). Iwata, B.A., Dorsey, M.F., Slifer, K.J., Ba uman, K.E., & Richman, G.S. (1982). Toward a functional analysis of self-injury. Analysis and Interven tion in Developmental Disabilities 2 3-30. Iwata, B.A., Pace, G., Kalsher, M., Cowd ery, G., & Cataldo, M. (1990). Experimental analysis and extinction of se lf-injurious escape behavior. Journal of Applied Behavior Analysis, 23 11-27. Johnston, S.S., & OÂ’Neill R.E. (2001). Search ing for effectiveness and efficiency in conducting functional assessments: A revi ew and proposed process for teachers and other practitioners. Focus on Autism & Other Developmental Disabilities,16, 205-215. Johnston, J., & Pennypacker, H. (1993). Strategies and tactics of behavioral research (2nd ed.). Hillsdale, NJ: Erlbaum. Kazdin, A.E. (1982). Single-case research desi gns: methods for clinical and applied settings. New York : Oxford University Press. Kaiser, A.P., Kai, X., Hancock, T.B., & Fost er, M.E. (2002). Teacher reported behavior problems and language delays in boys a nd girls enrolled in Head Start. Behavioral Disorders 28 23-29. Kamps, D.M., Elllis, C., Mancina, C., Wybl e, J., Greene, L., & Harvey, D. (1995). Case studies using functional analysis for young children with behavior risks. Education and the Treatment of Children, 18 243-260. Kennedy, C.H., Long, T., Jolivette, K., Cox, J., Tang, J.C. Thompson, T. (2001). Facilitating general e ducation participation for stude nts with behavior problems by linking positive behavior supports and person-centered planning. Journal of Emotional & Behavioral Disorders, 9 161-172. Kinch, C., Lewis-Palmer, T., Hagan-Burke, S., & Sugai, G. (2001). A comparison of teacher and student functional beha vior assessment interview information from low-risk and high-risk classrooms. Education & Treatment of Children, 24 480-495.

PAGE 195

184 Lalli, J.S., & Goh, H.L. (1993). Naturalistic ob servations in community settings. In J. Reichle & D.P. Wacker (Eds.), Communication alternatives to challenging behavior: Integrating fu nctional assessment and in tervention strategies (pp.1140). Baltimore: Paul H. Brookes. Lane, K.L., Umbreit, J., & Beebe-Frankenbe rger, M.E. (1999). Functional assessment research on students with or at ri sk for EBD: 1990 to the present. Journal of Positive Behavior Interventions 1(2), 101-111. Lerman, D.C. & Iwata, B.A. (1993). Descriptiv e and experimental an alyses of variables maintaining self-injurious behavior. Journal of Applied Behavior Analysis, 26 293-319. Lewis, T.J., Scott, T.M., & Sugai, G. ( 1994). The problem behavi or questionnaire: A teacher-based instrument to develop functional hypothesis of problem behavior in general education classrooms. Diagnostique, 19 (2-3), 103-115. Lewis, T.J., & Sugai, G. (1996). Descriptive and experimental analysis of teacher and peer attention and the use of assessment-based interven tion to improve pro-social behavior. Journal of Behavioral Education, 6 (1), 7-24. Mace, F.C., & Lalli, J.S. (1991). Linking de scriptive and experimental analyses in the treatment of bizarre speech. Journal of Applied Behavior Analysis, 24 553-562. Mace, F.C., & Roberts, M.L. (1993). Fact ors affecting select ion of behavioral interventions. In J. Reichle & D.P. Wacker (Eds.), Communication alternatives to challenging behavior: Integrating fu nctional assessment and intervention strategies (pp.113-134). Baltimore: Paul H. Brookes. McKerchar, P.M., & Thompson, R.H. (2004) A descriptive anal ysis of potential reinforcement contingencies in the preschool classroom. Journal of Applied Behavior Analysis, 37 431-444. Murdock, S.G., OÂ’Neill, R.E., & Cunningham, E. (2005). A comparison of functional behavioral assessment procedures with a group of middle school students with emotional/ behavioral disorders. Journal of Behavioral Education, 14 5-18. National Institute of Health. (1989). NIH consensus on the treatment of destructive behaviors in persons with developmenta l disabilities. Bethesda, MD: Author. Neilsen, S.L., & McEvoy, M.A. (2004). F unctional behavior assessment in early education settings. Journal of Early Intervention, 26 115-131.

PAGE 196

185 Nelson, J.R., Roberts, M.L., Mathur, S.R.& Rutherford, R.B. (1999). Has public policy exceeded our knowledge base? A review of the functional behavior assessment literature. Behavioral Disorders, 24 169-179. Newcomer, L.L., & Lewis, T.J. (2004). Functional behavioral assessment: An investigation of assessment reliability and effectiveness of function-based interventions. Journal of Emotional and Behavioral Disorders, 12 168-181. OÂ’Neill, R.E., Horner, R.H., Albin, R.W., Sprague, J.R., Storey, K., & Newton, J.S. (1997). Functional assessment and program d evelopment for problem behavior: A practical handbook Pacific Grove, CA: Brooks/ Cole. Packenham, M., Shute, R. & Reid, R. (2004). A truncated functional behavioral assessment procedure for children with disruptive classroom behaviors. Education and Treatment of Children, 27 9-25. Paclawskyj, T.R., Matson, J.L., Rush, K.S., Smalls, Y., & Vollmer, T.R. (2001). Assessment of the converg ent validity of the Ques tions About Behavioral Function scale with analogue functional an alysis and the Motivation Assessment Scale. Journal of Intellectual Disability Research, 45 484-494. Reed, H., Thomas, E., Sprague, J., & Horner R. (1997). The student-guided functional assessment interview: An analysis of student and teacher agreement. Journal of Behavioral Education, 7 33-49. Reid, R. (2000). Functional behavior assessme nt for children with behavior disorders. Unpublished manuscript, University of Nebraska-Lincoln. Reid, R. & Nelson, J.R. (2002) The utility, a cceptability and practic ality of functional behavioral assessment for students with high-incidence problem behaviors. Remedial & Special Education, 23 15-24. Reid, J.B., & Patterson, G.R. (1991). Ea rly prevention and inte rvention with conduct problems: A social interaction model for th e integration of research and practice. In G. Stoner, M.R. Shinn, & H.M. Walker (Eds.), Interventions for achievement and behavior problems (pp.715-739). Silver Springs MD: National Association of School Psychologists. Repp, A.C., Felce, D. & Barton, L.E. (1988). Basing the treatment of stereotypic and self-injurious behaviors on hypotheses of their causes. Journal of Applied Behavior Analysis, 21 281-289. Repp, A.C., & Munk, D.D. (1999). Threats to inte rnal validity and exte rnal validity of three functional assessment procedures. In A.C. Repp & Horner, R.H. (Eds.), Functional analysis of problem behavior: From effective asse ssment to effective support (pp. 147-167). Belmont, CA: Wadsworth.

PAGE 197

186 Roberts, M.L., Marshall, J., Nelson, J.R ., Albers, C.A. (2001). Curriculum-based assessment procedures embedded within functional behavioral assessments: Identifying escape-motivated behaviors in a general education classroom. School Psychology Review, 30 264-278. Romaniuk, C., Miltenberger, R.G., Conyers, C ., Jenner, N., Jergens, M., Ringenberg, C. (2002). The influence of activity choice on problem behaviors maintained by escape versus attention. Journal of Applied Behavior Analysis, 35 349-362. Sailor, J. Guess, J. & Rutherford, J. (1968) Functional explanation of tantrum behavior of a nine-year old girl. Journal of Applied Behavior Analysis, 1 230-240. Sasso, G.M., Conroy, M.A., Stichter, J. & Fox, J.J. (2001). Slowing down the bandwagon: The misapplication of func tional assessment for students with emotional or behavioral disorders. Behavioral Disorders, 26 282-296. Sasso, G.M., Peck, J., Garrison-Harrell, L. (1998). Social inter action setting events: Experimental analysis of contextual variables. Behavioral Disorders, 24 34-43. Sasso, G.M., Reimers, T.M., Cooper, L.J., W acker, D., Berg, W., Steege, M., Kelly, L., & Allaire, A. (1992). Use of descriptive a nd experimental analysis to identify the functional properties of aberrant behavior in school settings. Journal of Applied Behavior Analysis, 25 809-821. Schill, M.T., Kratochowill, T.R., & Elliott, S.N. (1998). Functional assessment in behavioral consultation: A treatment utility study. School Psychology Quarterly, 13 116-140. Scott, T.M., Bucalos, A., Liaupsin, Nelson, C. M., Jolivette, K., Deshea, L. (2004). Using functional behavior assessment in genera l education settings: Making a case for effectiveness and efficiency. Behavioral Disorders, 29 300-323. Scott, T.M., McIntyre, J., Liaupsin, C ., Nelson, C.M., & Conroy, M.A. (2004). An examination of functional behavioral assessment in public school settings: Collaborative teams, experts and methodology. Behavioral Disorders, 29 384395. Shellady, S., & Stichter, J.P. (1999). Traini ng preservice and inservice educators to conduct functional assessments: Initial issues and implications. Preventing School Failure, 43 (4), 154-159. Shonkoff, J.P. & Phillips, D.A. (Eds.). (2000). From neurons to neighborhoods: The science of early childhood development. Washington, DC: National Academy Press.

PAGE 198

187 Shores, R.E., Wehby, J.H., & Jack, S.L. ( 1999). Analyzing behavior disorders in classrooms. In A.C. Repp & Horner, R.H. (Eds.), Functional analysis of problem behavior: From effective a ssessment to effective support (pp. 219-237). Belmont, CA: Wadsworth. Shriver, M.D., Anderson, C.M., & Proctor, B. (2001). Evaluating the validity of functional behavior assessment. School Psychology Review, 30 180-193. Sigafoos, J., Kerr, M., Roberts, D. (1994). Interrater reliability of the Motivation Assessment Scale: Failure to rep licate with aggressive behavior. Research in Developmental Disabilities, 15 333-342. Skinner, B.F. (1953). Science and Human Behavior New York: Free Press. Sprague, J.R. & Horner, R.H. (1999). Low fr equency high intensity problem behavior: Toward an applied technology of functiona l assessment and intervention. In A.C. Repp & Horner, R.H. (Eds.), Functional analysis of problem behavior: From effective assessment to effective support (pp. 98-116). Belmont, CA: Wadsworth. Stichter, J.P., Sasso G., Jolivette, K. (2004). Structural analysis and intervention in a school setting: Effects on problem behavior for a student with an emotional/behavioral disorder. Journal of Positive Behavior Interventions, 6, 166-178. Storey, K., Lawry, J.R., Ashworth, R., Danko, C.D., & Strain, P.S. (1994). Functional analysis and intervention for disruptiv e behaviors of a kindergarten student. Journal of Educational Research 87, 361-370. Sturmey, P. (1994). Assessing the functions of aberrant behaviors: A review of psychometric instruments. Journal of Autism and Developmental Disorders, 24 293-304. Sugai, G., Horner, R.H. & Sprague, J.R. ( 1999). Functional assessment-based behavior support planning: Research to practice to research. Behavioral Disorders, 24 253-257. Symons, F.J., McDonald, L.M., Wehby, J.M. (1998). Functional a ssessment and teacher collected data. Education and Treatment of Children, 21 135-160. Tawney, J.W., & Gast, D.L. (1984). Single subject research in special education. Columbus: Merrill. Touchette, P.E., MacDonald, R.F., & Langer, S. N. (1985). A scatter plot for identifying stimulus control of problem behavior. Journal of Applied Behavior Analysis, 18 343-351.

PAGE 199

188 Umbreit, J. (1996). Functional analysis of disr uptive behavior in an in clusive classroom. Journal of Early Intervention, 20 18-29. Umbreit, J., & Blair, K.S. (1997). Using stru ctural analysis to f acilitate treatment of aggression and noncomplisnce in a young child at-risk for behavioral disorders. Behavioral Disorders, 22 75-86. VanDerHeyden, A.M., Witt, J.C., & Gatti, S. (2001). Descriptive assessment method to reduce overall disruptive behavior in a preschool classroom. School Psychology Review, 30 548-567. Webster-Stratton, C. (1997). Early interven tion for families of preschool children with conduct problems. In M. J. Guralnick (Ed.), The effectiveness of early intervention. (pp.429-454). Baltimore: Brookes Publishing. Webster-Stratton, C. (2000). Oppositional-de fiant and conduct disord ered children. In M. Hersen & R.T. Ammerman (Eds.). Advanced abnormal child psychology (2nd ed.) pp. 387-412. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Publishers. Yarbrough, S.C., & Carr, E.G. (2000). Some re lationships between informant assessment and functional analysis of problem behavior. American Journal of Mental Retardation, 105 130-151. Zarcone, J.R., Rodgers, T.A., Iwata, B. A., Rourke, D., & Dorsey, M.F. (1991). Reliability analysis of the Motivation A ssessment Scale: A failure to replicate. Research in Developmental Disabilities, 12 349-360.

PAGE 200

189 BIOGRAPHICAL SKETCH My professional goal to make a differ ence in childrenÂ’s lives has remained unchanged since I began working with child ren in 1994. I initially entered the field through an internship while pursuing my b achelorÂ’s degree in psychology at Furman. I worked as a residential counselor for children with emotional and behavioral disorders at the Marshall Pickens ChildrenÂ’s Program in Greenville, South Carolina. I gained practical experience in cognitive behavioris m and worked under skilled professionals who taught me invaluable skills for working with children. I gained a tremendous respect for the field and knew working with children would be a lifelong goal. I use skills and knowledge from this first work experience often. It was also this experience that developed my interest in teaching at a university. In 1997, I began teaching children with seve re emotional disturbance in an inner city elementary school in Jacksonville, Florida. This was my first j ob as a public educator as I was hired without a degree in edu cation. I began teaching under an emergency temporary certificate with the agreement that I would begin my ma sterÂ’s in education immediately. Although typical of a first teach ing job, this experience teaching in the inner city made me aware of the state of public education and has been the source for inspiration and practice for teaching many unive rsity-level courses. During that initial year, being a teacher and a student offered a uni que perspective. I saw the impact a single teacher could have on a group of students, and I experienced the impact education professors could have on a group of teachers. As I moved back and forth from one role to

PAGE 201

190 another, I began to recognize the importance and relevance of research in education. I had the opportunity to apply what I learned to how I taught. I sa w a real connection between the training of teachers and th e immediate benefit of that training to students. I enjoy teaching all ages, and I am continually excited to translate evidence-ba sed practices into a learning environment that is structured, individualized, and, most of all, fun. My passion for learning and love for the classroom led me to the Ph.D. program at the University of Florida where I coul d realize my goal in becoming an education professor at a university. During this experien ce of pursuing my doctorate, my skills have been both sharpened and broadened. Teaching co urses and supervising teacher interns at the University of Florida has allowed me to put theory into practice. I have enjoyed incorporating my higher educat ion knowledge with my practi cal classroom experience to improve tomorrowÂ’s classrooms. My abilities to analyze and synthesize research and apply new strategies for the classroom have been valuable assets. I am pragmatic and empathetic towards teachersÂ’ goals for themselves and their students. Additionally, my research activities have allowed me to gain experience in writing grants, presenting at national conferences, collecting and translating data collection, exploring multicultural aspects of learning, collaborating with othe r professionals and working closely with families of young children. Education is my passion