Congruence between Parents' and Early Childhood Educators' Perspectives about Child Development using a Multi-dimensiona...

MISSING IMAGE

Material Information

Title:
Congruence between Parents' and Early Childhood Educators' Perspectives about Child Development using a Multi-dimensional Screening Measure
Physical Description:
1 online resource (172 p.)
Language:
english
Creator:
Pasia, Cathleen C
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
School Psychology, Special Education, School Psychology and Early Childhood Studies
Committee Chair:
Smith-Bonahue, Tina M
Committee Co-Chair:
Snyder, Patricia
Committee Members:
Waldron, Nancy L
Eyler, Fonda D
Pallas, Pamela J

Subjects

Subjects / Keywords:
congruence -- development -- parents -- screening -- teachers
Special Education, School Psychology and Early Childhood Studies -- Dissertations, Academic -- UF
Genre:
School Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Given the significant time many children spend with different caregivers, a need exists to identify efficient and effective strategies to support the reciprocal exchange of information about children’s abilities, needs, and preferences. Developmental screening instruments are being used with increased frequency in early education and care settings. While the purpose of developmental screening is to inform decisions about the need for additional assessment, supports, or services, screening might have an added value of serving as a method to gather and compare perspectives about child development from multiple caregivers who interact with the child in different contexts. Previous research suggests that parent-completed developmental screening instruments are a time and cost effective approach that improves decision-making, however, the methodology applied to examine the congruence between parent and professional perspectives using multi-dimensional developmental screening instruments limits the interpretations that can be made. The primary aim of the present study was to explore if parents and early childhood educators (ECEs) who spend significant time with young children shared similar perspectives about child development when they each completed the same multi-dimensional developmental screening instrument. A secondary aim was to examine whether or not congruence varied based on the developmental domain assessed, child age, or length of relationship between children and ECEs. Parent-ECE pairs (n = 126) were recruited from two university-based early education and care centers. Each member of the pair separately completed the same multi-dimensional developmental screening instrument for a child. Data obtained from parents and ECEs were analyzed descriptively by examining item agreement, classification agreement, and patterns of response selection. Exploratory analyses were conducted by examining intraclass correlation coefficients and beta regression. Parents and ECEs had highly similar perspectives as measured by overall item and classification agreement.  Parent-ECE item agreement varied across developmental domains and age cohort, but variation was not associated with significant differences across informants on referral decisions. Data gathered from comparing similarities and differences in ratings on developmental screening instruments suggest that the completion of routine developmental screenings may serve a dual-purpose of early identification and a foundation for rich discussion between significant caregivers.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Cathleen C Pasia.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Smith-Bonahue, Tina M.
Local:
Co-adviser: Snyder, Patricia.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-02-28

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045371:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 CONGRUENCE BETWEEN PARENT AND EARLY CHILDHOOD EDUCATOR PERSPECTIVES ABOUT CHILD DEVELOPMENT USING A MULTI DIMENSIONAL SCREENING INSTRUMENT By CATHLEEN C. PASIA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNI VERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Cathleen C. Pasia

PAGE 3

3 ACKNOWLEDGEMENTS Over the past six years, I have received support and encouragement from a great number of people. Dr. Patricia Snyder has been an outstanding mentor, who always believed in me when I did not, and pushed me further than I ever thought I could grow. Her guidance has made this a thoughtful and rewarding journey. To Dr. Tina Smith Bonahue and Dr. Nancy Waldron, who not only helped to build the foundation for the practitioner I am today, but serve as models for connecting research and practice. This study would not have been possible without the support of Dr. Pamela Pallas. I also t hank Dr. James Algina and Dr. Cynthia Garvan for your valuable statistical advice and guidance and to Dr. Fonda Davis Eyler for supporting me without hesitation. I could not have accomplished the monumental feat of data collection without the assistance o f the Center for Excellence in Early Childhood Studies and the personnel affiliated would like to thank the early childhood educators and the parents for generousl y taking the time out of their busy days to participate my study. Also, thank you to Jamie Runyons for helping me with data entry. To Tara McLaughlin, Salih Rakap, and Crystal Crowe Bishop, I was fortunate to share this journey with all of you. I thank you for all the hours you spent listening to me talk about my research, reassuring me of my potential, and most of all, motivating me to work harder. I also could not have made it through this journey without my cohort working alongside me: Susan Craft, Jenny Heretick, Stacey Rice, Katrina Moore, Sally Moore, Suzie Long, and Angela Dobbins. I would like to thank Dieter Schaefer for always believing that I could do this. Finally, I have nothing but gratitude for my parents, who led by example through their hard work and dedication to their careers and family.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGEMENTS ................................ ................................ ................................ ............. 3 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF DEFINITIONS ................................ ................................ ................................ ............... 10 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 Statement of the Problem ................................ ................................ ................................ ........ 14 Conceptual Framework ................................ ................................ ................................ ........... 16 Definition of Developmental Screening ................................ ................................ ................. 19 Context for the Study ................................ ................................ ................................ .............. 22 Significance of the Study ................................ ................................ ................................ ........ 23 Length of Relationship between Professionals and Children ................................ .......... 24 Comparing Parent and Professional Responses ................................ .............................. 24 Directions for Research ................................ ................................ ................................ ... 26 Purpose of the Present Study ................................ ................................ ................................ .. 27 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 30 Identification of Articles for Review ................................ ................................ ...................... 30 General Procedure ................................ ................................ ................................ ........... 30 Topic One: Issues and Trends in the Developmental Screening Literatur e .................... 30 Topic Two: Multi Dimensional Developmental Screening Conducted by Parents and Professionals ................................ ................................ ................................ ......... 31 Topic Three: Empirical Studies on Ages and Stages Questionnaires ............................. 31 Scope and Sequence of Literature Review ................................ ................................ ...... 32 Historical Context for Developmental Screening of Children ................................ ............... 33 Early and Periodic Screening and Diagnosis Program ................................ .................... 34 Individuals with Disabilities Education Act ................................ ................................ .... 36 Parallel Themes in Medical and Educational Legislation ................................ ............... 38 Recommended Practices in Developmental Screening ................................ .......................... 38 Criteria for Selecting Developmental Screening Instruments ................................ ......... 39 Training ................................ ................................ ................................ ........................... 40 Time for Administration ................................ ................................ ................................ .. 40 Serial Use ................................ ................................ ................................ ......................... 42 Multi Source and Multi Method ................................ ................................ ..................... 42 Parent Involvement ................................ ................................ ................................ .......... 43 Historical and Contemporary Issues in Developmental Screening ................................ ........ 44 Multi Dimensional Developmental Screening Conducted by Parents and Professionals ...... 45

PAGE 5

5 Participants ................................ ................................ ................................ ...................... 46 Methods to Examine Parent Professional Relationship ................................ .................. 47 Measures of parent professional relationship ................................ .......................... 48 Analyses of parent professional relationship ................................ ........................... 49 Findings of Studies Primarily Investigating Congruence ................................ ................ 50 Findings of Studies with a Focus on Interrater Score Agreement or Reliability ............. 51 Discussion and Limitations of Current Research ................................ ............................ 54 Multi Dimensional Developmental Screening with the Ages and Stages Questionnaires ..... 57 Overview of ASQ Studies ................................ ................................ ............................... 58 Studies Examining the Technical Properties of the ASQ ................................ ................ 58 Standardization sample ................................ ................................ ............................ 59 Validity ................................ ................................ ................................ ..................... 61 Reliability ................................ ................................ ................................ ................. 64 Validated for use in other countries ................................ ................................ ......... 65 Studies that Examined the Utility of the ASQ within a Screening Process .................... 65 Settings ................................ ................................ ................................ ..................... 66 Detection of children in need of referral ................................ ................................ .. 66 Time and cost for implementation ................................ ................................ ........... 67 Benefits for children and families ................................ ................................ ............ 67 Implementation in other countries ................................ ................................ ............ 68 Support for the Use of the Ages and Stages as a Multi Dimensional Developmental Screening Instrument ................................ ................................ ................................ ... 69 Summary ................................ ................................ ................................ ................................ 69 3 METHODS ................................ ................................ ................................ ............................. 78 Design of Study ................................ ................................ ................................ ...................... 78 Setting ................................ ................................ ................................ ................................ ..... 79 Participants ................................ ................................ ................................ ............................. 8 0 Early Childhood Educators ................................ ................................ .............................. 80 Par ents and Children ................................ ................................ ................................ ........ 81 Measures ................................ ................................ ................................ ................................ 82 Demographic Forms ................................ ................................ ................................ ........ 82 Ages and Stages Questionnaire ................................ ................................ ....................... 82 Validity of ASQ scores ................................ ................................ ............................ 83 Reliability of ASQ scores ................................ ................................ ......................... 87 Procedures ................................ ................................ ................................ ............................... 89 Sampling and Recruitment ................................ ................................ .............................. 89 Instructions for ECE Practitioners and Parents ................................ ............................... 90 Data Collection ................................ ................................ ................................ ................ 90 Data Analyses ................................ ................................ ................................ ......................... 91 Congruence Examined using Percent Agree ment ................................ ........................... 93 Congruence on ASQ items ................................ ................................ ....................... 93 Congruence on classification decisions ................................ ................................ .... 94 Congruence on Developmental Domain Raw Scores ................................ ..................... 95 Item Re sponse Pattern of Parents and ECEs ................................ ................................ ... 96 Predictors of Congruence ................................ ................................ ................................ 97

PAGE 6

6 4 RESULTS ................................ ................................ ................................ ............................. 103 Overall Congruence Between Parent and ECE Ratings ................................ ....................... 104 Overall Item Level Agreement ................................ ................................ ...................... 104 Overall Classification Agreement ................................ ................................ ................. 104 Parent and ECE Congruence by Child Age Cohort ................................ .............................. 105 Overall Item Level Agreement by Age Cohort ................................ ............................. 106 Overall Classification Agreement by Age Cohort ................................ ......................... 106 Parent and ECE Congruence for Five Developmental Domains ................................ .......... 107 Item Level Agreement by Developmental Domain ................................ ...................... 107 Classification Agreement by Developmental Domain ................................ .................. 108 Item Level Agreement by Developmental Domain and Child Age Cohort .................. 108 Classification Agreement by Developmental Domain and Child Age Cohort .............. 110 Relationships b etween Parent and ECE Raw Scores on Developmental Domains ....... 111 Comparison of Frequencies and Percentages of Response Selection betwe en Types of Informant ................................ ................................ ................................ ........................... 112 Frequency and Percentage of Response Selection for All ASQ Items .......................... 113 Frequency and Percentage of Response Selection by Age Cohort ................................ 113 Frequency and Percentage of Response Selection by Developmental Domain ............ 114 Frequency and Percent age of Response Selection by Developmental Domain across Age Cohorts ................................ ................................ ................................ ............... 115 Child Age and Length of ECE Child Relationship as Predictors of Congruence ................ 116 Summary ................................ ................................ ................................ ............................... 117 5 DISCUSSION ................................ ................................ ................................ ....................... 124 Overall Congruence Between Pa rents and ECEs Using the ASQ 3 ................................ ..... 124 Overall Item Level Congruence ................................ ................................ .................... 124 Overall Classification Congruence ................................ ................................ ................ 127 Exploring Variation s in Congruence of Parent ECE Ratings on the ASQ 3 ....................... 130 Item Level Congruence by Developmental Domain ................................ ..................... 131 Classification Congruence by Developmental Domain ................................ ................ 132 Item Level Congruence by Age Cohort ................................ ................................ ........ 134 Classification Congruence by Age Cohort ................................ ................................ .... 139 Relationships between Parent and ECE Raw Scores on Developmental Domains ....... 142 Examination of Informant Response Patterns ................................ ............................... 143 Child Age and Length of ECE Child Relationship as Predictors of Congruence ......... 145 Limitations of the Present Study ................................ ................................ ........................... 146 Implications of Findi ngs ................................ ................................ ................................ ....... 148 APPENDIX: INSTRUCTIONS FOR PARTICIPANTS ................................ ............................. 155 Instructions for Early Childhood Educators ................................ ................................ ......... 155 Instructions for Parents ................................ ................................ ................................ ......... 157 LIST OF REFERENCES ................................ ................................ ................................ ............. 159

PAGE 7

7 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 172

PAGE 8

8 LIST OF TABLES Table pag e 2 1 Select technical properties and recommended criteria for screening instruments. ............ 72 2 2 Sample characteristics of s tudies examining the relationship between parent and professional informants using developmental screening instruments. .............................. 73 2 3 Summary of studies examining parent professional relationship on developmental screening instruments. ................................ ................................ ................................ ........ 74 2 4 Demographic information of participants descri bed in psychometric studies of Ages and Stages Q uestionnaires. ................................ ................................ ................................ 75 2 5 Results of validity studies of Ages and Stages Questionnaires (in percentages). .............. 76 2 6 Results of score reliability studies of Ages and Stages Questionnaires. ............................ 77 3 1 Comparison of ASQ 3 standardization sample distribution with the present sample distribution by age cohort. ................................ ................................ ................................ 99 3 2 Number of children who were eligible for inclusion in the present study, whose parents provided consent, and who were included in analyses. ................................ ....... 100 3 3 Summary of research aims by ASQ unit of analysis, variable of interest, and analysis procedure. ................................ ................................ ................................ ......................... 101 4 1 M ( SD ) of item level percent agreement by age cohort and developmental domain. ....... 119 4 2 M ( SD ) classification agreement by age cohort and developmental domain. ................... 120 4 3 Frequency of response selection by parents (P) and early childhood educators (ECE) on the Ages and St ages Questionnaire 3. ................................ ................................ ...... 121 4 4 Percentage of response selection by parents (P) and early childhood educators (ECE) on the Ages and Stages Questionnaire 3. ................................ ................................ ........ 122 4 5 Predictors of parent ECE item agreement. ................................ ................................ ...... 123

PAGE 9

9 LIST OF FIGURES Figure page 3 1 The number of completed questionn aires for each ASQ 3 age interval. ......................... 102

PAGE 10

10 LIST OF DEFINITIONS Caregiver Any adult who spends extended time with a child. The adult may interact with the child within a home or child care setting. Congruence Agreement or correspondence between the judgments of two individuals as reflected in rating scale scores or observational codes (Dinnebeil & Rule, 1994; Snyder, Thompson, & Sexton, 1992). Convergence The process of synthesizing a wide base of information from multip le informants using the same or similar instrument in order to facilitate collaboration and teamwork (Bagnato 2007; Suen Logan, Neisworth, & Bagnato 1995). Developmental screening A form of assessment designed to be brief and used to help inform de cisions about a age peers as well as to identify if additional assessment, monitoring, or targeted support or intervention might be needed (Brassard & Boehm, 2007; Lichtenstein & Ireton, 1984; Nagle, 2007). Early education and care center A nonresidential facility that provides care and early learning experiences, often in classroom settings, for young children between the ages of 6 weeks through 5 years. Early childhood educator Adults who are employed b y an early education and care center to provide care and early learning experiences for young children. The adults may be trained, certified, or licensed to provide care and early learning experiences for young children. Examiner A person trained to adm inister an assessment instrument and who does not interact regularly with the child. Informant An adult who completes a developmental screening instrument.

PAGE 11

11 Microsystem The immediate setting(s) in which a child develops and grows and is compromised of 1979, p. 22). Mesosystem Formed by the interrelationships among microsystems (Bronfenbr enner, 1979 p. 22). Multi dimensional developmental screening instrument A type of assessment instrument designed to provide general information about child development in different domains, often including fine and gross motor, language and communication, problem solving/adaptive behavior, and personal social (Glascoe, 2010; Neiswor th & Bagnato, 2005). Parent Refers to an adult who is a natural, adoptive, or foster parent of the child or is designated as their guardian. The adult is legally live with the child or may provide assistance (Individuals with Disabilities Education Act [IDEA], 20 U.S.C. 1400 § 602(23), 2004). Professional A person who has expertise or training in the disciplines of medicine, psychology, education, or a related social sci ence discipline. Screening An activity or procedure involving a brief sampling of behavior for the purpose of providing general information about whether or not additional diagnosis or assessment is needed (Harrington, 1984; Lichtenstein & Ireton, 1984; Scott & Carran, 1989).

PAGE 12

12 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 CONGRUENCE BETWEEN PARENT AND EARLY CHI LDHOOD EDUCATOR PERSPECTIVES ABOUT CHILD DEVELOPMENT USING A MULTI DIMENSIONAL SCREENING INSTRUMENT By Cathleen Pasia August 2013 Chair: Tina Smith Bonahue Co chair: Patricia Snyder Major: School Psychology Given the significant time many children sp end with different caregivers, a need exists to identify efficient and effective strategies to support the reciprocal exchange of information about abilities, needs, and preferences. Developmental screening instrument s are being used with increa s ing frequency in early education and care settings. While the purpose of developmental screening is to inform decisions about the need for additional assessment, supports, or services, screening might have an added value of serving as a method to gather a nd compare perspectives about child development from multiple caregivers who interact with the child in different contexts. Previous research suggests that parent completed developmental screening instruments are a time and cost effective approach that i mproves decision making, however, the methodology applied to examine the congruence between parent and professional perspectives using multi dimensional developmental screening instruments limits the interpretations that can be made. The primary aim of the present study was to explore if parents and early childhood educators (ECE s ) who spend significant time with young children share d similar perspective s about child

PAGE 13

13 development when they each complete d the same multi dimensional deve lopmental screening ins trument. A secondary aim was to examine whether or not c ongruence varied based on the developmental domain assessed, child age, or length of relationship between children and ECEs. Parent ECE pairs ( n = 126) were recruited from two university based early education and care center s Each member of the pair separately completed the same multi dimensional developmental screening instrument for a child. Data obtained from parents and ECEs were analyzed descriptively by examining item agreement, classification agreement, and patterns of response selection. Exploratory analyses were conducted by examining intraclass correlation coefficients and beta regression. Parents and ECEs had highly similar perspectives as measured by overall item and classification agreeme nt. Parent ECE item agreement varied across developmental domains and age cohort, but variation was not associated with significant differences across informants on referral decisions. Data gathered from comparing similarities and differences in ratings o n developmental screening instruments suggest the potential for the completion of routine developmental screenings to serve a dual purpose of early identification and a foundation for rich discussion about child development between significant caregivers.

PAGE 14

14 CHAPTER 1 INTRODUCTION Statement of the Problem Over the past several decades, the number of women in the workforce has been steadily increasing and current estimates are that 64.2% of women with children under the age of 6 years are employed (United S tates Department of Labor, 2011). As a result, children are spending an (2005) estimated 12 million infants, toddlers and preschoolers are cared for by adults other th an their parents, representing about half of all children in that age group. Of those 12 million children, approximately one fifth are reported to spend time in early education and care centers. Children can enter an early education and care center as youn g as 6 weeks of age, for as many as 40 hours per week. This translates to almost a quarter of the 168 hours in a week being spent with early childhood educators (ECEs). These data suggest that many children are having extended contact with at least two ty pes of caregivers: parents and ECEs. When primary preferences (Neiswor th & Bagnato, 2005). When caregivers from different settings have shared (Bronfenb renner, 1979). informal ways. It can be difficult for parents and ECEs to integrate regular communication about d preferences into their typical interaction patterns (Bailey & Wolery, 1992). One possible solution to this issue is to take

PAGE 15

15 advantage of existing practices in the early education and care program and incorporate strategies that actively include parents a learning, and behavior. Children are frequently screened throughout the early childhood years to determine the presence of early indications for delays in development and to identify children who might be in need of additional support (National Research Council, 2008). Although developmental screening has historically occurred in pediatric settings, screening is increasingly being conducted in early education and care programs (National Research Council 2008). Periodic screenings provide an opportunity to gather and share perspectives from all caregivers, Although early childhood professional organizations recommend early educatio n and care programs use multiple sources of information during the assessment process (Division for Early Childhood [DEC], 2007; National Association for the Education of Young Children [NAEYC], 2012), there is some flexibility for programs to determine ho w this is accomplished. Programs can gather information from parents using either informal or formal methods (NAEYC, 2012), leading to wide variability in the amount and quality of parental involvement. One common hesitation around incorporating parental perspectives about child development, learning, and behavior is the belief held by many professionals that parents are not reliable informants (Sheehan, 1988; Shelton, 1989). This belief persists despite promising evidence that shows when parents are provi learning, and behavior in the same ways as professionals using the same instruments there is not only convergence bet ween parents and professionals but often high levels of congruence (Bricker & Squires, 1989b; Bricker, Squires, Kaminski, & Mounts, 1988; Snyder, Thompson, & Sexton, 1992; Squires & Bricker, 1991; Squires, Bricker, & Potter, 1997). A needs exists to examine

PAGE 16

16 development, learning, and behavior, particularly because early education and care programs are routinely using multi dimensional development screening instruments as required by program policies and procedures (NAEYC, 2008, 2012) Given the significant time many young children spend in early care and education environments in addition to their home environments, it is perspectives abou uncover and discuss consistency or variation in perspectives would be useful for ensuring young children experience nurturing and responsive relationships to facilitate their devel opment and learning within and across contexts. The purpose of the present study was to compare the perspectives of two primary caregivers (i.e., parent and ECE) about the development of a young child by having each of these caregivers independently comp lete the same multi dimensional developmental screening instrument. In the remainder of this chapter, the conceptual framework and a definition for developmental screening are presented. In addition, the context for and significance of, the study is discus sed. In chapter two, a review of the developmental screening literature is conducted, including studies focused explicitly on examining the convergence and congruence of parental and professional perspectives about child development using multi dimensional developmental screening instruments. The methods used in the study are described in chapter three. In chapter four, findings are presented and chapter five includes discussion and implications of study findings for future research and practice. Conceptual Framework interplay between an individual and the immediate settings in which the developing person

PAGE 17

17 spends significant time (Bronfenbrenner, 1979). The immediate setting( s) in which a child primary microsystem, but children who spend si gnificant time in other settings, such as early education and care centers, experience additional microsystems. The transactional relationships between or among microsystems exert significant influences on the processes underlying human development and lea rning. Interrelationships among microsystems form a mesosystem. when there is more than one person who actively participates in both settings (Bronfenbrenner, 1979). On the contrary, mesosystems that are connected only by the presence of the child in both systems framework, mesosystems. The richness of a mesosystem is reflected in the number and quality of connections between the adults (i.e., parents and ECEs) from two micr osystems (i.e., home and early education and care centers; Garbarino, 1990). When strong connections are formed through information, advice, and experience relevant to o ne setting are made available to the other (Bronfenbrenner, 1979). Similarly, open communication supports a shared understanding about there is a compatibility of expectations and activities for children. Continuity across settings contributes to the development of mutual goals, including mutual goals related to development,

PAGE 18

18 learning, and behavior, thus increasing the frequency of opportunities for a child to exp erience developmentally enhancing learning opportunities across different contexts and people. Mesosystems are created by the sharing of perspectives across settings, which can take a variety of forms and vary in the direction of flow (Bronfenbrenner, 197 9). Three general types of & Barfoot, 2000). Parents and ECEs can share perspectives through the verbal exchange of information at arrival or pick up, written exchanges through daily journals or behavior report cards, or visual exchanges through sample materials or weekly photographs (Bailey & Wolery, 1992). Developmental scr eening has potential to serve as a method to actively share information centers, this information is already being gathered as part of ongoing decision making abou t behavior and experiences in other settings (Owen et al., 2000). Likewise, ECEs can provide similar information to parents. This reciprocal flow of communication can either form or enhance a multiply linked mesosystem. Studies f ocused on examining the congruence of parental and early childhood educator perspectives about child development can help inform how information should be gathered, shared, and used to form or enhance a multiply linked mesosystem. Although convergence and congruence are not always necessary or expected, knowing the similarities and differences in perspectives can serve as a starting point for conversations about why children might exhibit

PAGE 19

19 different developmental skills in different contexts or ways all care givers can support development, learning, and behavior across contexts. Developmental screening can serve as one method for strengthening mesosystems by systematically gathering, comparing, and sharing De finition of Developmental Screening Broadly defined, screening is an activity or procedure involving a brief sampling of behavior (Harrington, 1984; Lichtenstein & Ireton, 1984; Scott & Carran, 1989). Other authors specify the procedure or activity as se parating a population into two groups: those at risk for later problems and those who are not (Barnett, 1984; Brooks Gunn & Lewis, 1983; Glascoe, 2005; Green & Scriven, 1980; Kochanek, 1993; Meisels & Wasik, 1990; Sheldrick & Perrin, 2009). Both definitio ns focus on the procedure of screening, but more contemporary definitions of screening specify its role as an entry point into a larger decision making process. In the latter type of definition, screening is described as a form of assessment to determine a developmental status relative to peers of the same age and to identify if additional assessment, monitoring, or targeted support or intervention might be needed (Brassard & Boehm, 2007; Harrington, 1984; Lichtenstein & Ireton, 1984; McLean & McCor mick, 1993; Nagle, 2007; Taylor, 1993; Zehrbach, 1975). Screening can occur at any age to aid in the detection or prevention of many types of illnesses or conditions (Lichenstein & Ireton, 1984). Developmental screening refers to the administration of a standardized assessment instrument or tool focused on sampling key developmental milestones or skills for the purposes of aiding in the identification of children who are in need of additional supports, evaluations, or services (American Academy of Pediatr ics [AAP], 2001; Council on Children with Disabilities, Section on Developmental Behavioral Pediatrics, Bright Futures Steering Committee, & Medical Home Initiatives for

PAGE 20

20 Children with Special Needs Project Advisory Committee, 2006). By definition, developm ental screening alone should not result in a diagnosis or treatment plan. It is intended to identify areas age peers and should only be used to guide decisions as to whether a child may benefit from further evaluation, closer monitoring, or additional support or intervention. Children are likely to encounter developmental screening multiple times before the age of 6 years as a result of recognition by professional organizatio ns of the importance of early identification (Council on Children with Disabilities et al., 2006; NAEYC, 2012; National Association of School Psychologists [NASP], 2009; Neisworth & Bagnato, 2005). Consequently, developmental screening can serve as a metho d for gathering and exchanging information from contexts (i.e., microsystems) and to form or enhance mesosystems. In addition to its applicability as a multi infor mant tool, developmental screening instruments can serve as a multi dimensional assessment. Practitioners might choose to use a multi dimensional developmental screening instrument or combine multiple instruments to screen across different developmental do mains. For the purpose of the present study, the focus was multi dimensional developmental screening instruments because of legislative requirements and policy recommendations that children be assessed in five developmental areas. For example, the AAP and Division for Early Childhood (DEC) recommend that developmental screening instruments should be broad and examine all domains of development: fine and gross motor skills, language and communication, problem solving/adaptive behavior, and personal social sk ills (Council on Children with Disabilities et al., 2006; Glascoe, 2010 ; Neisworth & Bagnato, 2005 ).

PAGE 21

21 For most developmental screening instruments two methods are used to gather information: direct elicitation or parent/caregiver report (Boan, Aydlett, & M ultunas, 2007; National Research Council, 2008). Direct elicitation involves the observation of child actions in response to the presentation of stimuli. When developmental screening instruments are administered through direct elicitation, examiners should be trained to administer the instrument reliably (Bradley Johnson & Johnson, 2007; Brassard & Boehm, 2007; Kenny & Culbertson, 1993; Paget & Nagle, 1986). The second method is to gather information from a parent/caregiver er through a questionnaire or interview format (Boan et al., 2007). The second method can be used with any caregiver, including an ECE, provided the caregiver has spent enough time with the child to be able to answer all of the items. Most screening instru ments specify approximately how much time a caregiver should spend with a child before the caregiver completes a questionnaire or participates in an interview. Multi dimensional developmental screening instruments can be administered as part of various as sessment frameworks. Two common frameworks are developmental surveillance and multi tiered screening. Developmental surveillance is most likely to occur in the pediatric setting nuous, and developmental history, and risk and protective factors and integrate them with observations, physical examinations, and laboratory assessments. (Cou ncil on Children with Disabilities et al., 2006, p. 407). In an alternative framework, screening occurs within a multi level or multi tiered approach (Brooks Gunn & Lewis, 1983; Dunbar & Reed, 1999; Jackson & Needleman, 2007; Ikeda, Neessen, & Witt, 2008; Teska & Stoneburner, 1980; Thurlow & Gilman, 1999; Zehrbach, 1975). One variant of the multi tiered framework includes a pre screening procedure intended to

PAGE 22

22 limit the number of children for whom a multi dimensional developmental screening is administered. In a second variant of the multi tiered framework, developmental screening could be administered to all children as a universal screener and followed by targeted screeners or additional supports (Ikeda et al., 2008). Of these two frameworks, a multi tiere d framework for screening may be more appropriate for implementation of a multi dimensional developmental screening instrument in an early education and care center. Context for the Study Three major pieces of legislation have mandated screening as a me thod of early identification of children in need of intervention. In the medical field, the Early and Periodic Screening, Diagnosis and Treatment (EPSDT) program was established as a part of a 1967 provision to the Medicaid Act to target children experienc ing environmental risk due to low income backgrounds (Frankenburg & North, 1974). In the field of education, Child Find was initially established in 1974 through the Elementary and Secondary Act (ESEA) as an effort to identify children with disabilities fr om birth through 21 years (Paget & Nagle, 1986). As Child Find and other legislative mandates have undergone successive reauthorizations other organizations have disseminated additional position statements and recommendations supporting the use of developm ental screening: American Academy of Pediatrics (AAP, 2001; Council on Children with Disabilities et al., 2006), National Association for the Education of Young Children (2012), and National Association of School Psychologists (2009). Additionally, the Hea d Start Act (Improving Head Start for School Readiness Act, 42 U.S.C. § 1304.20 2007 ) mandates developmental screening upon 45 days of entry into programs and NAEYC (2008, 2012) requires programs to conduct ongoing screening and assessment of child progre ss to obtain accreditation status.

PAGE 23

23 Although developmental screening typically occurs within the pediatric setting because it serves as a primary context for professional contact with infants and toddlers (National Research Council, 2008), the increase in the number of children entering early education and care settings education and care programs, regular and systematic developmental screening is being implemented to meet accreditation requirements (NAEYC, 2008, 2012). When conducting assessment in early childhood, recommended practices include involving parents as active team members in screening and subsequent assessment pr ocesses and procedures (Bagnato, 2007; Br assard & Boehm, 2007; Kochanek, 1993; Ireton, 1990; Moore, 1978; Neisworth & Bagnato, 2005; Snyder, McLaughlin, & McLean, in press; Thurlow & Gilman, 1999). As developmental screening becomes a universal practice across different contexts and settings it fulfills multiple functions by helping to (a) identify children at risk (b) (c) incorporat e parent input, and (d) support sharing of perspectives about child development across caregivers and settings Significance of the St udy Reviews conducted by Snyder, Thompson, and Sexton (1992) and Dinnebeil and Rule developmental status. These reviews only included studies that examined relationships between parent report (typically elicited through a questionnaire or interview) with a comprehensive, developmental assessment completed by a professional. In both of these reviews, the term correspondenc e between parent and professional reports of child developmental status. To date, there has not been a published review of studies examining congruence when parents and early childhood educators complete multi dimensional developmental screening assessment instruments. Through the use of systematic

PAGE 24

24 search procedures described in detail in Chapter 2, 10 studies were identified that compared parent and professional perspectives about child development using multi dimensional developmental screening instrument s. Findings from these studies that are significant for the present study are summarized below. Length of Relationship between Professional s and Child ren dimensional developmental screening instrume nts to those of professionals who did not typically interact with the child. The dimensional developmental Frankenburg, Fandal, &Thornton, 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976). Across all ten studies, no descriptions were provided about the length of the time that the professional or day care provider knew or interacted with the child prior t o completing the developmental screening instrument. Lack of congruence might be due to the limited opportunities that professionals have to observe or elicit skills or behaviors. A need exists for developmental status using a multi dimensional developmental screening instrument when completed by two individuals who interact often with a child in the home or early education and care microsystems. Comparing Parent and Professional Responses Across all 10 studies, moderate to high levels of congruence were found between parent and professional responses on multi dimensional developmental screening instruments. Often reported as interrater score reliability, congruence wa s examined using the following methods: (a) agreement on the classification of children related to those who require additional evaluation and those who do not; (b) item level agreement; or (c) statistical evaluation of score

PAGE 25

25 agreements/differences or rela tionships (e.g., Pearson product moment correlation coefficients, F tests, and kappa statistics). overall classification of children (e.g., not at risk, at risk ). Alt hough this method provides a gross estimate of congruence, it is limited with respect to similarities or differences in perspectives across the two respondents on specific items. Likewise, calculating agreement by using an overall average masks potential d ifferences by developmental domain. Chiu and DiMarco (2010) and Coghlan et al. (2003) demonstrated that parent and professional responses might differ depending on the skill being addressed. Both of these studies showed high congruence across respondents f or gross motor skills and low agreement across the same respondents for language skills. It is plausible that agreement or score congruence on items included on multi dimensional s age or the amount of time the professional has known the child. None of the reviewed studies, however, explored these two variables, suggesting a need for additional studies focused on examining correlates of congruence. In seven studies, parent and pro fessional perspectives were gathered by having both informants use the same instruments. When developing the Denver Prescreening Developmental Questionnaire (PDQ) and its revised edition, the R PDQ, authors compared parent perspectives to those of day care providers and examiners, but only day care providers completed the same screening instrument as parents (Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976). Use of the same instrument to gather and quantify perspectives eliminate s potential measurement bias that might result from using two different instruments. A need exists for more studies that gather perspectives of parents and early childhood educators about child

PAGE 26

26 development using the same multi dimensional developmental scr eening instrument. In addition, individuals who regularly interact with the child should complete these instruments. Employing these procedures would eliminate at least one form of measurement bias, potentially strengthen the inferences that could be made about congruence, and adhere to recommended practices in early childhood assessment (Bagnato, 2007; Neisworth & Bagnato, 2005). Developmental screening instruments used in empirical investigations of congruence should meet standard psychometric criteria. The primary focus in all but two of the 10 studies was an examination of interrater score reliability for a new or revised developmental screening instrument. Interrater score reliability was calculated by comparing scores of parents and professionals. The se studies provided some information about similarities and differences in reliability is viewed by some researchers as a different construct than congruence. Conventional statistical evaluations of interrater score reliability operate under the assumption that informants need to be interchangeable and are randomly selected (Suen, Logan, Neisworth, & Bagnato, 1995). Interrater score reliability is a reflection o f the stability of a score across informants and demonstrates that similar scores would be obtained independent of the informant used. When informants are not assumed to be equivalent, perhaps as in the case of parents and professionals, assumptions relate d to interrater score reliability might not apply. Directions for Research A typical misconception of the goal of research focused on congruence is to provide evidence for parental estimates of child ability being similar to professionals (see discussi ons by Dinnebeil & Rule, 1994; Snyder et al., 1992). These authors and other proponents for multi informant approaches to the assessment process recognize the need to incorporate information from both practitioners and parents. Striving for exact agreemen t in judgments or ratings would

PAGE 27

27 negate the benefits of including multiple perspectives in assessment (Suen et al., 1995). Likewise, discussing parent overestimations or timations (Sheehan, 1988). Parents and professionals bring different sets of frameworks and expertise to the assessment process and each perspective is essential to making decisions as a team. Thus, while the studies re viewed were interested in levels of agreement, it may be more appropriate to investigate congruence similarities in estimations that preserve the richness of each perspective by removing the expectation of absolute agreement (Dinnebeil & Rule, 1994; Snyde r et al., 1992; Neisworth & Fewell, 1990 ). The interest in investigating congruence is to understand the relationship between the judgments or ratings of parents and professionals, specifically one who also knows the child (e.g., ECE), and how this is chan ges with developmental domain measured, child age, and length of relationship between ECE and child. Exploring these relationships may provide valuable information for the formation of multiply linked mesosystems. Ultimately, if congruence is explored acro ss multiple microsystems so that several adults (e.g., parents, ECE providers, and other professionals) are sharing information, then convergence might occur. Convergence is the process of synthesizing a wide base of information from multiple informants us ing the same instrument in order to facilitate collaboration and teamwork (Bagnato 2007; Suen et al., 1995). Understanding the nature of the relationship between different perspectives could inform the way that information is gathered and shared across set tings in which children learn and grow. Purpose of the Present Study Previous researchers have shown support for eliciting parent report about child development (Bricker & Squires, 1989b; Bricker et al., 1988; Snyder et al., 1992; Squires & Bricker, 1991 ; Squires et al., 1997). Few empirical studies have investigated parent and

PAGE 28

28 professional congruence as measured by a multi dimensional developmental screening instrument completed in the same way by each informant. As will be discussed further in Chapter 2 research that employs defensible methods for gathering parent and professional report, measuring child development, and analyzing correspondence between parent and professional judgments is needed. To address this gap in the literature, the following res earch questions will be addressed in the present study: 1. When using the same multi dimensional developmenta l screening instrument, what is overall congruence between two informants who interact wi th the child regularly ? 2. Does congruence between parents and early childhood educators vary based on child age cohort ? 3. Does congruence between parents and early childhood educators vary depending on the child developmental domain assessed ( i.e., communication, gross motor, fine motor, problem solving, personal soc ial)? 4. Are there differences in the pattern of response selection across informant types (i.e., parents and ECEs)? 5. Does child age and length of relationship with the early childhood educator predict item level congruence between parents and ECE? In summa ry, the increasing number of women entering the workforce has resulted in children spending more time in contexts with other caregivers. Children may spend up to 40 hours a week being cared for in other settings, such as an early education and care program Given that young children spend significant time in at least two microsystems, the reciprocal exchange of information about child development and learning could facilitate connections between caregivers (i.e., parents and ECEs) and ensure each is informe development, learning, and behavior within and across contexts. Open communication enhances practice skills across contexts and people.

PAGE 29

29 Dev elopmental screening instruments are increasingly being used in early education and care settings as a result of recommendations by prominent organizations in the medical, educational, and early childhood fields (AAP, 2001; Council on Children with Disabil ities et al., 2006; NAEYC, 2012; NASP, 2009; National Research Council, 2008). While the primary purpose of developmental screening instruments is to inform decisions about whether children need additional supports or services, they potentially have an add ed value of serving as a formal method of gathering and comparing perspectives of primary caregivers from different settings. When microsystems are connected through frequent and high quality connections between adults, a multiply linked mesosystem is form enhanced (Bronfenbrenner, 1979).

PAGE 30

30 CHAPTER 2 LITERATURE REVIEW To provide a rationale for the need for the present study and situate it within the larger body of literature focused on developmental screening, a systematic descriptive review of the literature was conducted. This review explored three main topics: issues and trends in developmental screening, examination of studies that involved parents and professionals in developmental screening, and empirical st udies focused on the caregiver c ompleted developmental screening instrument used in the present study. Identification of Articles for Review General Procedure To address each of the topic areas outlined above, three separate electronic searches were conduc ted. The same procedure was applied to each search. EBSCOhost was used to search simultaneously Academic Search Premier, PsycInfo, and Professional Development Collection databases. Acceptable sources were limited to periodicals and books. The age of parti cipants was limited to infancy (2 23 months), preschool (3 5 years), and childhood (birth 12 years). The results of each search were exported to RefWorks to identify duplicate sources. After the removal of duplicate sources, the titles and abstracts of ea ch source were reviewed for specific criteria related to each topic. If an article met all specified criteria, the reference list was used to conduct an ancestral search. Topic One: Issues and Trends in the Developmental Screening Literature The following search terms were entered into the electronic databases: (a) developmental screen*, screen* and young children; (b) screen* and infants; screen* and preschool*; and (c) screen* and es were reviewed for the presence of five criteria: (a) written in the United States or sampled participants

PAGE 31

31 exclusively from this country; (b) categorized as either a research study, position statement, book chapter, or book; (c) primarily focused on scre ening; (d) described a multi dimensional developmental screening instrument that measured at least four domains of development; and (e) was not an investigation of the psychometrics of one or more screening instruments. Of the 3,204 abstracts reviewed, 93 articles/books/book chapters met criteria for inclusion in this review. Topic Two: Multi Dimensional Developmental Screening Conducted by Parents and Professionals The following search terms were entered into EBSCOhost: (a) developmental screening tests a nd parents; and (b) congruence* and development*. To be included in the present review, the articles from peer reviewed journals had to meet all of the following criteria: (a) a professional (e.g., examiner, pediatrician, nurse, teacher, psychologist) comp leted a developmental screening instrument; (b) a parent/caregiver completed a developmental screening instrument by report or interview; (c) instruments used by professionals and parents were multi dimensional developmental screening instruments, meaning that items sample at least four domains of development (e.g., motor, cognitive, communication, social, adaptive); and (d) the selected multi dimensional developmental screening instrument(s) was used in its English version and had psychometric data reporte d in either the published report or the test manual. The titles and abstracts of these articles were reviewed for pre specified inclusion criteria. A total of 10 articles were identified for this topic area. Topic Three: Empirical Studies on Ages and Stag es Questionnaires The Ages and Stages Questionnaires Third Edition (ASQ 3) was created for use in screening or monitoring programs in settings such as health clinics, pediatric practices, early education and care settings, home visiting programs, and com munity based programs (Squires, Twombly, Bricker, & Potter, 2009). The purpose of the ASQ 3 is to gather information about

PAGE 32

32 child development from individuals who interact regularly with the child. It is used in many early education and care programs as a u niversal screener, including the program in which the present study was conducted. To identify articles related to the ASQ, the following search terms were entered in or or or Bricker, Diane. Articles were screened for four criteria: (a) the source must be available in English, (b) the source reports on empirical data collected by the listed authors, (c) the data were publishe d in a journal or test manual, and (d) the primary research question(s) were focused on the application of the ASQ (or Infant Monitoring Questionnaire). After applying these four criteria, 45 of 186 sources were identified for inclusion in this review. As sources were reviewed, studies were classified as either an investigation of the technical properties of the ASQ or of implementation of the ASQ in applied settings. Within each of these categories, sources were separated if they were conducted with childr en from countries other than the United States. The present review only briefly summarizes the studies conducted outside of the United States, as these studies were primarily translations of the ASQ and subsequent investigations of applicability with sampl es outside the United States. Scope and Sequence of Literature Review Using the search procedure described above, articles that met inclusion criteria were reviewed in their entirety. The remainder of Chapter 2 is organized to review the history of develop mental screening, issues and trends in the contemporary literature on developmental screening, empirical research on the completion of multi dimensional developmental screeners by parents and professionals, and support for the use of the Ages and Stages Qu estionnaire as the primary study instrument of interest.

PAGE 33

33 Historical Context for Developmental Screening of Children The 1960s were an era of progressive reform because of growing concern for the poor educational and health outcomes of families who were de scribed as being disadvantaged 1 The achievement gap between students considered to be advantaged and disadvantaged was a prominent educational issue, prompting leg islation for federal funding of Head Start to improve educational opportunities for young children from disadvantaged backgrounds (Frey & Walker, 2006). Studies conducted during this time highlighted the prevalence of medical conditions or diseases and sec ondary conditions that could be prevented among this population. In 1967, a provision to the Medicaid Act established the Early and Periodic Screening, Diagnosis, and Treatment (EPSDT) program with the primary purpose of early identification of and interve ntion In addition to the growing concern for children whose families were characterized as disadvantaged, the 196 0s and 1970s represented a time of heightened attention to the conditions of state institutions for individuals with disabilities (United States Department of Education, 2010). As part of the Elementary and Secondary Education Act (ESEA), the Education of Handicapped Amendments of 1974 (PL 93 380) established Child Find efforts for identifying children with disabilities from birth through 21 years, but there was no requirement for the provision of services (Paget & Nagle, 1986). Similar to EPSDT, Child Find arose from a response to growing awareness of the benefits of early identification and intervention for children with or at risk for disabilities. Unlike EPSDT, Child Find was intended to identify any 1 the 1960s to refer to persons experiencing environmental risk due to limited finances. This terminology is used in this section to preserve historical accounts.

PAGE 34

34 child who might have a disability. Although developmen tal screening arose from two different pieces of legislation often governing different disciplines (i.e., medicine and education), professionals in both fields faced similar implementation issues. Early and Periodic Screening and Diagnosis Program EPSDT w on Medicaid that went beyond paying bills to assuring that eligible children received periodic es early and as characterized as disadvantaged were entitled under EPSDT to periodic physical examinations and screenings. EPSDT was the first major federal commitm ent to reach large numbers of children in need of basic health services as well as developmental and medical screenings. Several years after the initial implementation of EPSDT, program evaluations conducted separately by the American Orthopsychiatric Ass concluded EPSDT was not adequately meeting the needs of the children who it was created to serve. Many barriers were identified at administrative and professional levels (AOA, 1978; Moore, 1978). EPSDT model demonstration programs heavily emphasized screening for cognitive development, primarily in relation to preparation for school and learning. Although this developmental area was essential, it did not take into account the importan ce of examining more than one domain of child development given the widely acknowledged recognition of the interrelationships among developmental domains (Shonkoff & Phillips, 2000) and the relationships of multiple developmental domains to later school su ccess EPSDT program mirrored the goals of Head Start at that time and not the broader goal of illness prevention that EPSDT was created to address.

PAGE 35

35 Another challe nge in the implementation of EPSDT was the little value medical professionals placed on developmental screening as part of comprehensive health services (AOA, 1978). Even if developmental screening occurred under EPDST, professionals were often faced with a lack of available resources to conduct additional in depth diagnostic assessments with a developmental focus or to offer services and supports to families if a child was identified as being at risk or having a developmental delay or disability (AOA, 1978 Fund, 1978). As discussed in the EPSDT guidelines, screening alone did not equate to offering help (Frankenburg & North, 1974). If families were not being connected to further evaluations or resources, then screening to identify risk, the need for additional diagnostic assessment, or the procurement of resources becomes questionable from an ethical perspective. After conducting evaluations of EPSDT programs across the nation, the AOA (1978) and that the most significant area for improvement was informing and involving parents in the screening processes and information gathering related to making informed decisions. Overall, parents were reported to be minimally included in the screening process. In some EPSDT sites, for example, parents were not present for the screening provided at times deemed inappropriate by the program evaluators. Within the primary heal th care setting, little or no time was allocated to inform parents of the screening results, provide recommendations for what to do next, or advise on how to address screening data that indicated risk or the need for additional assessment. Thus, the lack o f continuity among screening, assessment, and intervention was due to insufficient resource allocation at the broad administrative level and within individual clinical or primary care practices.

PAGE 36

36 Individuals with Disabilities Education Act Several legisla tive acts were passed to assist states in educating children with disabilities as an alternative to home schooling, institutions, or non public segregated schools. As part of 142) was passe d in 1975 to guarantee a free, appropriate public education to children with disabilities in every state, ages 6 to 18 (U.S. Department of Education, 2010). Subsequent amendments made to EHA included changing its name to the Individuals with Disabilities E ducation Act (IDEA) in 1990. The passage of EHA brought incentives for the identification of preschoolers with disabilities through the Preschool Incentive Grant and guaranteed access to special education and related services following the determination of eligibility for services (Kenny & Culbertson, 1993; Paget & Nagle, 1986). Child Find was the term designated for the process for identifying children, ages birth through 21 years, with delayed development and who subsequently, may benefit from special edu cation services (Katz, 1989; Paget & Nagle, 1986). Although screening was not specifically described in this legislation, a national survey of implementation of early school screening requirements with lead educational agencies reported 33 states mandate d some form of screening for preschoolers for physical, emotional, behavioral or learning problems (Gracey, Azzara, & Reinherz, 1984). The majority of these states reported conducting screenings primarily one time, upon entry into preschool or kindergarten while only eight states conducted periodic (ongoing) screening. More than half of the states collected screening information from growth and development (Dia mond, 1993; Diamond & Le Furgy, 1988; Diamond & Squires, 1993). In 1986, the reauthorization of EHA through the passage of Public Law 99 457 contained provisions to expand the availability of early intervention services to eligible infants and toddlers

PAGE 37

37 and established the preschool program for children with disabilities (IDEA, 20 U.S.C. 1419 § 300, 1986 ; U.S. Department of Education, 2010 ). With the new amendments to IDEA, free, appropriate public education was now available for children 3 through 21 yea rs, and the Part H (now Part C) program was available for eligible children from birth through age 2 and their families. participating members of the assessment and interven tion team (Katz, 1989). Thus, legislation was increasingly acknowledging the importance of including parents in the decision making process of determining whether a child needed early intervention or special education services. Kochanek (1993) conducted a review of state policies present in 1991 pertaining to early identification and Child Find. He reviewed written policies for the presence of 10 exemplary screening indicators: multiple data sources, multiple data reporters, multivariate decision making, c oordination with other screening efforts, family as informant, cultural sensitivity, screening as services, periodicity, score reliability and validity, and training. The last four indicators were not specified in federal regulations, but were considered r ecommended practices in assessment (DEC, 1993). Kochanek concluded that there was considerable variability in state policies on the inclusion of knowledge based, exemplary components, and notably, state policies were less likely to make reference to screen ing indicators if they were not explicit in federal regulations. The majority of states did not reference periodicity, psychometric properties of screening instruments, or screening as an initial step in comprehensive assessment that was linked to interven tion. More importantly, approximately 30% of states did not make reference to families being major informants regarding child and family status, needs, and goals. The majority of states made reference to this indicator, but the policies were often vague an d did not provide specific procedures and implementation strategies related to family involvement.

PAGE 38

38 procedures, and outlined quality assurance standards. Althou gh it is possible that written policies did not reflect actual implementation in the field, findings from the Kochanek study indicated the need for guidance at the state level to support the use of recommended practices in screening. Parallel Themes in Me dical and Educational Legislation The passing of EPSDT and EHA/IDEA reflected the enactment of federal policies that explicitly acknowledged the importance of early identification of children at risk for or with disabilities. Identification of these young children was facilitated by the implementation of developmental screening by professionals properly trained in administration and interpretation. While there was widespread acceptance of the concepts of early identification and the use of developmental scr eening, there was not universal implementation by all EPSDT programs or local education agencies in states. There was evidence of the early acknowledgement of improved accuracy in decision making when professionals include parent input (AOA, 1978; Paget & Nagle, 1986). However, professionals in the health care and education fields had struggled to implement this practice effectively. Policies typically offered few guidelines on how to include parents as informants, leading to minimal or non existent parent involvement in developmental screening conducted in health care clinics and schools. To improve practices that engage parents in the developmental screening process, a need existed for clarity and specificity in policies and recommendations. Recommended Practices in Developmental Screening Several organizations have offered recommended assessment practices for use with young children: American Academy of Pediatrics (2001; Council on Children with Disabilities et al., 2006), Division for Early Childhood (2 007; Neisworth & Bagnato, 2005), National Association of School Psychologists (2009), and the National Association for the Education of Young

PAGE 39

39 Children (2003). Of these, AAP and NASP make specific reference to screening while the remaining statements refer broadly to various assessment purposes and processes. Notably, although NAEYC (2003) does not specifically include screening in its recommended practices, it is a required practice for early education and care programs that are applying for accreditation s tatus (NAEYC, 2012). Collectively, these organizations unanimously support the gathering of multiple sources of evidence to inform screening or assessment decisions, and specifically the inclusion of parent involvement in the decision making process. Speci fic recommended practices and guidelines will be reviewed in conjunction with other relevant literature. Criteria for S electing Developmental Screening Instruments The Standards for Educational and Psychological Testing (1999), a joint publication by the American Educational Research Association (AERA), American Psychological Association (APA), and the National Council on Measurement in Education (NCME) is commonly considered to be the prominent authority of standardized test construction. Some researcher s have applied these standards to specifically address developmental screening instruments (Brassard & Boehm, 2007; Glascoe, 1991; Glascoe, 2007a; Salvia & Yssledyke, 1985). Jointly, the recommendations of these associations and researchers outline definit ions and recommended criteria for developmental screening instruments for standardization, score reliability, score validity, sensitivity, and specificity (Table 2 1). Three reviews have been conducted on the psychometric properties of commercially availa ble developmental screening instruments. The first review, conducted by Katoff and Reuter (1980), included 21 developmental screening instruments. A decade later, Glascoe, Martin, and Humphrey (1990) selected a review panel of pediatricians and a special e ducator to examine the psychometric properties of 19 developmental screening instruments available at that time. Most recently, Macy (2012) reviewed published reports from peer reviewed journals on 14

PAGE 40

40 developmental screening instruments. Comparisons of the results of these three reviews show an increase in the number of developmental screening instruments that meet appropriate test standards (Glascoe et al., 1990; Katoff & Reuter, 1980; Macy, 2012). The selection of an appropriate developmental screening in strument requires the knowledge to critique the characteristics of an instrument using a set of recommended criteria. Training A quality developmental screening program begins with professionals who have been properly trained in administration and interpre tation of developmental screening instruments (Kenny & Culbertson, 1993). In order to be effective, professionals who are using developmental screening instruments must receive training about developmental differences of infants, toddlers, and preschoolers that could affect the interpretation and use of screening results in conjunction with other information (Brassard & Boehm, 2007; Nagle, 2007; Paget & Nagle, 1986). An understanding of early childhood assessment practices is beneficial regardless of whethe r the professional is relying on direct elicitation methods or caregiver report. If direct assessment is conducted by the professional, additional training on developmentally appropriate assessment techniques would become necessary to ensure score validity and reliability (Bradley Johnson & Johnson, 2007; Brassard & Boehm, 2007; Paget & Nagle, 1986) Using caregiver report to complete a developmental screening instrument would eliminate the necessity of specialized training, which could potentially translate to savings in cost and personnel time. Time for A dministration Multi dimensional developmental screening instruments should include items that reflect a broad sampling of skills across multiple domains of development (Boan et al., 2007; Council on Childr en with Disabilities et al., 2006; Glascoe, 2010; Meisels & Wasik, 1990; Neisworth & Bagnato, 2005). Taking this premise into consideration, instruments should also be relatively

PAGE 41

41 brief in order to fit within the constraints of a busy clinical practice, lar ge caseload, or early education and care program (Brooks Gunn & Lewis, 1983; First & Palfrey, 1994; Meisels & Wasik, 1990). This includes considerations related to ease of scoring and interpretation of results (Brooks Gunn & Lewis, 1983; Scott & Carran, 19 89). While many developmental screening instruments take, on average, about 15 minute to complete (Council on Children with Disabilities et al., 2006), there is often limited time for a professional to build adequate rapport with the child to gather inform ation in a direct elicitation format. This limitation might preclude valid inferences from screening data gathered using direct assessments. Pediatricians frequently report that although they consider early identification of developmental problems importan t, administration time deters them from routinely using developmental screening instruments (Glascoe & Dworkin, 1993; Smith, 1978). As a result, pediatricians will frequently rely on clinical judgment when deciding whether to refer a child for additional e valuation (Glascoe & Dworkin, 1993; Sand et al., 2005; Scott Lingaraju, Kilgo, Kregel, & Lazzari, 1993; Smith, 1978). This approach is heavily prone to error and might lead to the under identification of children, particularly those with more subtle delays The use of standardized screenings is recommended over informal assessments (Glascoe, 1991; Marks, Hix Small, Clark, & Newman, 2009; Meisels & Wasik, 1990; Squires, Nickel, & Eisert, 1996; tandardized screening instruments can take the form of parent or professional achievement of developmental milestones, or parent concerns, or by direct elicitation and ty to complete specified tasks. Parent completed instruments can be completed prior to arriving at a scheduled appointment, leaving more time during the appointment for discussion or other evaluations. Although parent completed instruments are

PAGE 42

42 typically wr itten at a functional reading level (Squires et al., 2009), practitioners should be sensitive to the needs of some parents who may require verbal administration or clarification of questions. Serial U se Developmental screening should be repeated over tim e to take into account age related manifestations of delay (Council on Children with Disabilities, et al., 2006; Glascoe, 1991; Glascoe, 2005; Glascoe, 2007a; Jackson & Needelman, 2007; Kenny & Culbertson, 1993; Kochanek, 1993; NASP, 2009). When developmen tal screening is used as a one time activity, such as screening upon entry into a program, children whose difficulties are not evident to informants until a later point might be overlooked. Repeated screenings provide other opportunities to identify childr en who have received false negative results on prior screens, should have been. Multi S o urce and Multi M ethod The screening process should use an ecological f ramework that accounts for the multiple contexts in which children grow: the home, community, and school (Bagnato, 2007; Boan et al., 2007; Colligan 1981; DEC, 2007; Glascoe, 1991; Hall, Bramlett, Barnett, & Cox, 1994; Kochanek, 1993; NASP, 2009; Neisworth & Bagnato, 2005; Paget & Nagle, 1986; Taylor 1993). Information should be gathered by using multiple modes of assessment (i.e., observation, interview, direct elicitation), multiple sources (i.e., early childhood educators, family members, professionals), and sample behaviors from multiple contexts (i.e., community, school, home). those contexts improves the ecological validity of screening decisions.

PAGE 43

43 Parent I nv olvement ecommended that throughout the screening process, communication with parents is frequent, clear, and culturally sensitive (Eddey, Robey, Zumoff, & Malik, 199 5 ; Moore, 1978; Neisworth & Bagnato, 2005; Scott & Carran, 1989). Parents would benefit from having a clear understanding of the process, having their concerns attended to, and receiving immediate feedback about results. If the results of a screening indicate the child may potentially have issues that require further evaluation, parents should be linke d to assistance, either by a referral to the appropriate agency or by a list of recommendations for follow up (Dworkin, 2006; Moore, 1978). As screening serves as the entry point into comprehensive assessment processes (Bagnato, 2007), interactions with pa rents set the tone for expectations of future encounters with other health and can produce anxiety for parents and it is important to be sensitive to their needs during this process. Parents should be invited to choose their preferred role in the assessment of their child, allowing for active engagement as team members in screening and subsequent procedures when desired (Bagnato, 2007; Brassard & Boehm, 2007; Koch anek, 1993; Ireton, 1990; Moore, 1978; Neisworth & Bagnato, 2005; Thurlow & Gilman, 1999). Parents can be included by asking them Diamond & Squires, 1993; Marks et a l., 2009). Given that screening is a relatively brief procedure, there often is limited time for a professional to build adequate rapport with the child to gather information in a direct elicitation format. This limitation might preclude valid

PAGE 44

44 inferences f across time, contexts, and people than the professional is able to observe and elicit during a short period of time. Accurate information is likely to be gathered when parents are asked about observable behaviors, particularly behaviors that occur frequently. When asking parents about behaviors, it is recommended that parents select response choices rather than respond using open ended formats (Diamond & Squires, 1993). Histo rical and Contemporary Issues in Developmental Screening The value of developmental screening in identifying children who may benefit from further evaluation and, if needed, early prevention or intervention, has been recognized in legislation and in positi on statements disseminated by several national organizations. There has been an increase in the specification of rigorous standards and guidelines that should be used to evaluate developmental screening instruments and their implementation within a develop mental screening model. Prominent organizations, particularly AAP and DEC, have published policy statements and recommended practices to guide practitioners in selecting and using developmental screening instruments. Improvements have been made in the numb er of available instruments with adequate psychometric qualities. Since the early implementation of EPSDT and IDEA legislation, many challenges related to implementing parent involvement practices in developmental screening remain. From the 1970s to the p resent day, experts have recognized that parent involvement throughout the assessment process is important in promoting ecological validity of results and forming positive relationships with families (Bagnato 2007; DEC, 2007; Glascoe, 1991). Over reliance on professionally administered developmental screening instruments has resulted in inconsistent implementation of parent involvement in the assessment process mainly due to time constraints in busy clinical practices or early education and care programs. H aving parents

PAGE 45

45 complete developmental screening instruments addresses many obstacles to consistent implementation. Parents are able to observe a wider range of behaviors, across time, and across multiple contexts and people and can provide this information to professionals in a shorter time. Including parents at this early stage of the assessment process is an opportunity to provide awareness and knowledge about important developmental milestones, regardless of the results of the screening. Many children ar e spending large amounts of time in early education and care environments with non parental caregivers. Developmental screening instruments can serve as a informan professional unfamiliar to the child or family is unable to observe in a short period of time. As a result of the growing need to integrate information from multiple contexts, it becomes child development when measured with the same developmental screening instruments. Multi Dimensional Developmental Screening Conducted by Parents and Professionals described as congruence between parent and profes sional judgments. Congruence has been Rule, 1994; Snyder et al., 1992). The term congruence is preferred over accuracy as the latter implies an evaluative statem percentage of a greement or disagreement in identification, and item analysis ( Dinnebeil & Rule,

PAGE 46

46 1994 ). As discussed in Chapter 1, the term congruence is used to capture similarities in ratings without the implication that agreement is the ultimate goal. Reviews conducte d by Snyder et al. (1992) and Dinnebeil and Rule (1994) have supported the validity of eliciting parent report, however, the studies reviewed by these authors typically examined the relationship between parent report as elicited through a questionnaire or interview and a professionally completed, comprehensive, developmental assessment that was often directly administered to the child. To date, there has not been a published review of studies examining congruence when both informants complete the same multi dimensional developmental screening instrument. Using the search procedures and criteria described earlier, 10 studies were identified that examined the relationship between parent and professional (e.g., examiner, pediatrician, nurse, day care provider, psychologist) perspectives of child development across multiple domains using a developmental screening instrument. Findings from a descriptive analysis of these 10 studies are reported below These 10 studies were categorized as either studies that prima rily investigat ed interrater score reliability or studies that primarily investigat ed congruence. To provide clarification when discussing these studies, the following terminology will be applied: will be used to refer broadly to the group of 10 studies, regardless of the stated purpose and (b) primary purpose Partici pants Seven of 10 articles compared parent report to an examiner who did not regularly interact with the child (i.e., Bricker et al., 1988; Bricker & Squires, 1989b; Burgess, Asher, Doucet, Reardon, & Daste, 1984; Chiu & DiMarco, 2010; Glascoe, 2002; Squir es & Bricker, 1991; Squires, Bricker, & Potter, 1997). Examiners represented the disciplines of medicine,

PAGE 47

47 psychology, and education or were reported to be professionals trained in assessment. In two studies, the disciplines or training of examiners were no t described. One of the studies sent screening forms to day care providers or kindergarten teachers and made comparisons to parent report (Coghlan et al., 2003). Two of the studies used both of these approaches by making parent examiner comparisons and par ent teacher comparisons (Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976). None of the studies provided information The characteristics of each study sampl e are shown in Table 2 2. Across all studies, distribution of ages within the sample or explored if results varied by age or age group. Half of the studies sampled children with identified risk factors (Bricker & Squires, 1989b; Bricker et al., 1988; Chiu & DiMarco, 2010; Squires et al., 1997; Squires & Bricker, 1991). Risk factors were either medical (e.g., remaining in the neonatal intensive care unit for a minimu m of three days) or environmental (e.g., extreme poverty). In four studies, only a subsample of participants had screening forms completed by parents and a professional (Bricker & Squires, 1989b; Bricker et al., 1988; Squires et al., 1997). In two of the 1 0 studies, authors compared parent examiner and parent teacher pairs, but only a subsample of children had received teacher completed ratings (Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, and Dick 1976). For six studies, it was not reporte d whether the characteristics of the subsample used to address study questions related to interrater score reliability differed from the larger study sample. Methods to Examine Parent Professional Relationship Although all 10 studies had moderate to high levels of agreement or score congruence, a variety of approaches were used to explore the relationship between parent and professional judgments or ratings. In the sections that follow, the multi dimensional developmental screening

PAGE 48

48 instruments implemented in each of the 10 studies and the analyses used to interpret results are reviewed. Measures of parent professional relationship The following multi dimensional developmental screening instruments were represented in this review: Ages and Stages Questionna ire, 2 nd Edition (ASQ; Bricker & Squires, 1999), previously known as the Infant/Child Monitoring Questionnaires (ICMQ; Bricker & Squires, 1989a), Brigance Infant and Toddler Screen (Brigance & Glascoe, 2002), Denver Developmental Screening Test (DDST; Fran kenburg & Dodd, 1975), Denver II (Frankenburg et al., 1992), Denver Prescreening Developmental Questionnaire (PDQ; Frankenburg, van Doornick, & Liddell, 1976), the Revised Denver Prescreening Developmental Questionnaire (R PDQ; Frankenburg, 1987) and Paren Table 2 3 shows the instruments used in each study. Eight of the studies administered the same instrument to both parents and professionals (Bricker & Squires, 1989b; Bricker et al., 1988; Co ghlan et al., 2003; Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976; Glascoe, 2002; Squires & Bricker, 1991; Squires et al., 1997). Frankenburg et al. (1987) and Frankenburg, van Doornick, Liddell, and Dick (1976) compared two pa irs of scores (i.e., parent teacher and parent examiner). Parents and teachers completed the PDQ or R PDQ, while the examiner administered the DDST. These eight studies were all standardization or validation investigations of new or revised screening instr uments. Although the primary purpose of these studies was to establish interrater score reliability, analyses were conducted that permitted an examination of the relationship between parent and professional scores. The remaining two studies primarily focus ed on the congruence between parents and professionals when using multi dimensional developmental screening

PAGE 49

49 instruments (Burgess et al., 1984; Chiu & DiMarco, 2010), but neither article used the term congruence to describe these comparisons. Analyses of p arent professional relationship The analyses used in each study are shown in Table 2 3. The majority of studies evaluated congruence by calculating percent agreement on the classification of children (Bricker & Squires, 1989b; Bricker et al., 1988; Chiu & DiMarco, 2010; Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976; Squires & Bricker, 1991; Squires et al., 1997). Screening instruments classified children as either in need of further evaluation, at some risk of needing further ev aluation, or no need for further evaluation at the present time The remaining studies used other analysis procedures. To calculate parent examiner agreement, Frankenburg, van Doornick, Liddell, and Dick (1976) calculated item level agreement between the 1 0 PDQ questions with the corresponding DDST questions. Compared to the studies described above, Chiu and DiMarco (2010) reported agreement on classification for each developmental domain represented on the screening instrument Similarly, Coghlan et al. (2 003) reported item level percent agreement, with each item representing a different developmental domain. Chiu and DiMarco (2010) and Coghlan et al. (2003) were the only two studies in which domain specific results were reported. Few studies used tests of statistical significance to evaluate the relationship between parent and professional scores. Burgess et al. (1984) used a formula for calculating a developmental quotient (DQ) to translate PDQ scores (Frankenburg, van Doornick, & Liddell, 1976) and DDST (Frankenburg & Dodd, 1975) in order to calculate correlations for three pairs of informants. Frankenburg, van Doornick, Liddell, and Dick (1976) tested for differences between the positive predictive values of parent and teacher completed screening forms when compared to the professionally administered DDST. In the standardization of the Brigance Infant

PAGE 50

50 and Toddler Screen, Glascoe (2002) averaged scores across all domains for parent completed and examiner elicited screens. F test statistics were used to ev aluate whether there were statistically significant differences between means. In addition to reporting item/domain level percent agreement, Coghlan et al. (2003) calculated kappa values to correct for chance agreement. Findings of Studies Primarily I nv est igating C on gruence Only two of the 10 studies reviewed had a primary research question concerning the comparison of parent and professional scores (Burgess et al., 1984; Chiu & DiMarco, 2010). In neither of these studies was the same screening instrument c ompleted by parents and professionals. Burgess et al. (1984) and Chiu and DiMarco (2010) used the professionally administered DDST (Frankenburg et al., 1975) or Denver II (Frankenburg et al., 1992). Additionally, both of these studies compared parent repo rt to examiners who did not typically interact with the child. Despite these limitations, these studies provide preliminary information about score congruence between parents and professionals when multi dimensional developmental screening instruments are used. Burgess et al. (1984) were interested in comparing the relationship between three pre screening methods with a professionally administered DDST (Frankenburg et al., 1975). Two of the methods involved administering a pre screen either completed indep endently by mothers (PDQ) or in an interview format (PDQ M). The third group was prescreened using the Alpern Boll Development Profile (Alpern, Boll, & Shearer, 1980). PDQ and DDST results were translated into developmental quotient (DQ) scores to calcul ate correlations between screening instruments. The three prescreening tests correlated significantly with the DDST DQ (Burgess et al., 1984). The strongest correlation was between Alpern Boll and DDST scores ( r = .68), followed by the PDQ M ( r = .60) and the PDQ ( r = .48).

PAGE 51

51 These results are shown in Table 2 3. Using the DDST as the criterion, the number of correct referrals (positive results on the pre screen and screen) and the number of under referrals (negative prescreen and positive screen) were the hi ghest for PDQ M. The authors concluded that the PDQ M was the most acceptable means for pre screening children. Administering the completing the PDQ. The second stud y of congruence compared two developmental screening instruments, one completed by parents and the other completed by a nurse, using a sample of children who were homeless (Chiu & DiMarco, 2010). Children received screenings at a homeless shelter. Mothers independently completed an ASQ 2 (Bricker & Squires, 1999) for their child. A nurse administered the Denver II (Frankenburg et al., 1992). The results were compared using a calculation of percent agreement on the classification of children for further eval uation by each developmental domain. Agreement was highest across parents and nurses in personal social and gross motor domains (95% each), followed by fine motor (71%). The lowest percent agreement was reported for the language domain (67%). The ASQ 2 alo ne classified children as needing further evaluation in the areas of personal social, gross motor, and fine motor/adaptive. The Denver II identified more children for referral in the area of language than the ASQ 2. The authors hypothesized that this resul 2) even when others cannot (i.e., nurses not crediting items on the Denver II). Findings of Studies with a Focus on I nterrate r Score Agreement or R eliability When examining overall agreement between those completing developmental screening instruments on the classification of children, results show moderate to high agreement. This methodology was used in six studies (Bricker & S quires, 1989b; Bricker et al., 1988;

PAGE 52

52 Frankenburg et al., 1987; Squires & Bricker, 1991; Squires et al., 1997) and was included as one of the analyses by Frankenburg, van Doornick, Liddell, and Dick (1976). When agreement was examined at the item level, two studies showed moderate to high agreement between parents and professionals (Coghlan et al., 2003; Frankenburg, van Doornick, Liddell, & Dick, 1976). Statistical tests of significance were utilized in three studies and reported mixed findings (Coghlan et al., 2003; Frankenburg, van Doornick, Liddell, & Dick, 1976; Glascoe, 2002). Three of these studies compared parent reports to reports of teachers or day care providers (Coghlan et al., 2003; Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976), while the remaining studies conducted comparisons with professional examiners that did not regularly interact with the child. The results of these studies are summarized in Table 2 3 and described in detail below. Frankenburg, van Doornick, L iddell, and Dick (1976) validated a newly created instrument, the PDQ (Frankenburg, van Doornick, & Liddell, 1976), which was developed through the selection and adaptation of 10 items from the DDST (Frankenburg et al., 1975). Parents completed the PDQ whi le a subsample ( n = 202) was selected to receive a teacher completed PDQ. Item level agreement between parents and examiners was 93%, with a range of 68% 100%. To determine parent teacher agreement, researchers calculated classification agreement on the children, which was 59%. Based on the PDQ results, teachers referred more children than parents, while PDQs completed by parents had a slightly higher positive predictive value. When both of the PDQs were compared to the examiner administered DDST, positiv e predictive value was 15.8% for parents and 10.8% for day care providers, but the difference between these values was not statistically significant. Comparisons between parent examiner and teacher examiner scores for all children, regardless of classifica tion, were not available.

PAGE 53

53 The PDQ was revised almost a decade later as the R PDQ (Frankenburg, 1987). Instead of answering 10 questions as on the PDQ (Frankenburg, van Doornick, & Liddell, 1976), parents continued answering R PDQ questions until a ceiling yes or no no opportunity validation of the original instrume nt, parents completed the R PDQ while a subsample ( n = 71) was selected to complete a teacher completed R PDQ. Classification agreement between parents and teachers was 83%. There were almost equal numbers of parent ratings identifying children as having d elays as there were with teacher ratings identifying delays. In total, 193 children received R PDQ screenings completed by parents and DDST screenings completed by examiners. Data available in tables published with the study indicated that classification a greement between parents and examiners was 74%. Comparing the classification agreement across all of the children shows that parents have higher agreement with another caregiver who regularly interacts with the child. Using the IMQ/ASQ, agreement ranged f rom 87% 97% between parents and professional examiners who had not previously met the child (Bricker & Squires, 1989b; Bricker et al., 1988; Squires & Bricker, 1991; Squires et al., 1997). In three of these studies, a portion of the original sample had t o be eliminated from the analysis because one or more test domains had two or more uncompleted answers by the examiner (Bricker & Squires, 1989b; Squires & Bricker, 1991; Squires et al., 1997). Examiners reported that they had little opportunity to observe the child engaged in certain activities reflected in ASQ items. The PEDS is comprised of 10 questions to elicit concerns in nine developmental domains (Coghlan et al., 2003; Glascoe, 2007b). Item/domain level agreement between parents and

PAGE 54

54 teachers and kap pa values were calculated for each of these concerns. In all categories, parents reported more concerns than teachers. Percentages of agreement were considered to be high (>75%) for all of the categories except for behavior (64.9%). Kappa values calculated for each category were low to modest ( 0.02 0.40); suggesting that the high percent agreement reported may have been due to chance. It is also possible that lower kappa values may be due to a low base rate of response in some of the categories (Streiner 2003). These values may have differed if there were a higher number of at risk children included in the sample or if overall agreement for the PEDS had been examined. During the initial investigation of the psychometric properties of the Brigance Infan t and Toddler Screen (Brigance & Glascoe, 2002), versions were administered to parents to complete by self report or interview and to examiners for completion by observation or direct elicitation (Glascoe, 2002). Scores for each screen were averaged across six domains and compared. There were no statistically significant differences in results using either form of screening administration after controlling for age and psychosocial risk ( F[ 5,170] = 0.543). Discussion and Limitations of Current Research Dev elopmental screening is increasingly becoming a part of common practice in health care and early education and care programs (National Research Council, 2008). While developmental screening has been traditionally used as a method of identifying children wh o are in need of additional supports, services, and evaluations, it can also serve as an avenue to share and exchange information about child development. When completed by caregivers representing the multiple contexts in which children spend time, develop mental screening can serve as a information is collected from parents and teachers, discussions are informed by screening results with enhanced ecological validity. A review of literature related to developmental screening

PAGE 55

55 score congruence was conducted to understand what is known about the relationship between parent and professional scores when these two informant groups use multi dimensional developmental screening instruments Through a systematic search of the literature, 10 studies were identified in which parent and professionals completed a multi dimensional developmental screening instrument. Overall, studies showed moderate to high levels of professional and pa rent agreement or congruence, however, there is a clear need for additional studies that administer the same screening instruments to two individuals who are in regular contact with the child. Only three studies paired parent responses with teachers from d ay care centers or other settings (Coghlan et al., 2003; Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976). Using the PEDS (Glascoe, 2002), seven of 10 items had percentages of agreement above 80%, although kappa values were gener ally low (Coghlan et al., 2003). The study conducted by Frankenburg, van Doornick, Liddell, and Dick (1976) provided data for parent examiner (PDQ DDST) and parent teacher (PDQ PDQ) agreement, however, it is difficult to compare these numbers because of th e different methods employed to calculate each of these percentages. When validating the Revised PDQ, similar methods were applied to measure parent teacher (R PDQ R PDQ) and parent examiner (R PDQ DDST) classification agreement (Frankenburg et al., 19 87). Although results showed higher agreement between caregivers who regularly interact with their child, parents and examiners were completing two instruments with differing item content and administration methods (i.e., parent report versus direct elicit ation). Additional research is needed to examine congruence by having parents and professionals complete the same screening instrument using the same administration methods, as differences in agreement may be a

PAGE 56

56 function of the items and structures of the i nstruments rather than reflecting differences in perspectives about child development. Improved methods for examining congruence are needed. Studies need to be conducted that use analyses that provide information beyond classification agreement. Likewise, calculating agreement by using an overall average masks potential differences by domain. The results of Chiu and DiMarco (2010) and Coghlan et al. (2003) demonstrate that congruence can differ depending on the skill being addressed. Both of these studies h ad high agreement for gross motor skills and low agreement in the rating of language skills. It is possible that congruence on time the early care and education p rovider has known the child, but these differences are unknown because results in previous studies were not differentiated by age of child or length of contact. The majority of the studies reviewed primarily focused on demonstrating the score reliability of a new or revised instrument. Interrater score reliability was calculated by comparing ratings of parents and professionals. These studies provided some information as to how errater score reliability operates under the assumption that informants need to be interchangeable and are randomly selected (Suen et al., 1995). Interrater score reliability is a reflection of the stability of a score across informants and high interrater score reliability suggests similar scores would be obtained independent of the informant used. This concept does not apply when informants are not assumed to be interchangeable, as in the case of parents and professionals. Whether or not the professional is someone who interacts with the child regularly, parents and ECEs likely will consider child development with different sets of experiences, backgrounds, and perspectives.

PAGE 57

57 Alternatively, it is possible that parents and ECEs will have similar perspectives about child development due to the significant amount of time each person spends with the child, sharing many of the caretaking responsibilities. Ultimately, similarities and differences in perspectives about child development should be used to help info rm how perspectives should be gathered, shared, and used to create a microsystems communicate information that promotes the compatibility of expectations and activiti es (Bronfenbrenner, 1979). The relationships that occur within the mesosystem are critical scores obtained from the two informants can serve as a starting point fo r conversations about why children might exhibit skills differently based on context or how caregivers can support their generalization across contexts. In addition, understanding variables that affect congruence of perspectives about child development mig ht help inform activities related to establishing and maintaining parent professional partnerships. Multi Dimensional Developmental Screening with the Ages and Stages Questionnaires The present study focused on the congruence between parent and profession al perspectives about child development and the ASQ 3 (Squires & Bricker, 2009) was used as the multi dimensional developmental screening instrument. The ASQ 3 was already being implemented as part of a universal screening program at the early education an d care program that served as the 3 at designated intervals as part of ongoing activities at the center. Prior to implementing the developmental screening instrument with parents from the center as part of the present study, a review of the ASQ empirical literature was conducted to support its continued use.

PAGE 58

58 Overview of ASQ Studies Twenty seven articles investigated the use of the ASQ in other countries; 17 sources sampled participants residi ng within the United States. Seven of the 17 sources included research questions concerning the technical properties of the ASQ. Seven sources implemented the ASQ in applied settings such as pediatric clinics or home visiting programs. Three sources were i nterested in the ability of ASQ score to describe the development of a special population of children. Studies E xamining the Technical Properties of the ASQ Seven sources were identified that investigated the technical properties of the Infant Monitoring Questionnaire (IMQ, a precursor to the ASQ) and the ASQ (Bricker & Squires, 1989b; Bricker et al., 1988; Gollenberg, Lynch, Jackson, McGuinness, & Msall, 2010; Squires & Bricker, 1991; Squires et al., 1997; Squires, Potter, Bricker, & Lamorey, 1998; Squir es et al., 2009). The majority included articles published in journals ( n = 6). One source was the technical manual for the current edition of the ASQ (Squires et al., 2009). The authors of the ASQ 3 conducted seven of the eight psychometric investigations reviewed here. Each study was reviewed and data were gathered on demographics of participants and findings of score reliability (i.e., test retest, interrater, and internal consistency) and score validity (i.e., construct validity, sensitivity, specific ity, false positive ratio, false negative ratio) of the IMQ or ASQ. Definitions of each of these psychometric terms and recommended criteria were adapted from a review of seminal literature in standards for assessment (AERA et al., 1999; Salvia & Ysseldyke 1985) and literature on developmental screening instruments (Brassard & Boehm, 2007; Glascoe, 1991; Glascoe, 2007a). Detailed definitions and criteria are listed in Table 2 1.

PAGE 59

59 Standardization sample Revisions for the current edition of the ASQ (i.e., AS Q 3, Squires et al., 2009) were based on 18,572 questionnaires completed by parents of children between 1 and 66 months of age. Within this sample, 3,434 children had more than one completed questionnaire. Therefore, the total demographic sample consisted of 15,138 children. Data were collected across four years in either paper or electronic format via a project website. Respondents represented a wide geographic distribution, from all 50 states, and several U.S. territories. When broken down by age, sample sizes for questionnaires ranged from 352 (2 months) to 2,088 (12 months) children (Squires et al., 2009). There were no questionnaire intervals that fell below the recommended criteria of 100 sample participants per age group (Brassard & Boehm, 2007; Glas coe, 1991; Salvia & Ysseldyke, 1985). It is recommended that the normative sample be representative of the population for which a screening instrument is intended (AERA et al., 1999; Brassard & Boehm, 2007; Glascoe, 1991). The ethnicity of the ASQ 3 norma tive sample of children was comparable to the U.S. Census estimates for 2007, suggesting representativeness of the general population (Squires et income, which were proportionately similar to the most recent U.S. Census. Children included in the sample were identified as having one or more medical or environmental risk factors. Although these terms were not defined in the current test manual, studies of previous editi ons defined medical risk factors of infants to be at least three days spent in the neonatal intensive care unit or having been referred for an evaluation due to medical or developmental concerns (Bricker & Squires; 1989b; Bricker et al., 1988; Squires et a l., 1997). Environmental risk factors include: (a) household income below the federal poverty line, (b) maternal age younger than 19

PAGE 60

60 the 12 th grade, and (d) a history of parental involvement with Child Protective Services. The description of the normative sample suggests that the authors made adequate effort to support the use of ASQ 3 with children from diverse ethnic, geographic, and socioeconomic distribution s. Descriptions of participants from studies of previous editions of ASQ. Compared to past studies of the ASQ, revisions for the third edition constituted the largest sample, spanned across more age intervals, and provided the most detail on participant c haracteristics (Table 2 4). These differences were partly due to the addition of age intervals with each subsequent revision; the IMQ had eight questionnaires from 4 to 36 months and the current edition, the ASQ 3, has 21 questionnaires from 2 60 months (Bricker & Squires, 1989b; Squires et al., 2009). Investigations of special populations. Three studies have examined whether ASQ scores correctly identify children within special populations. The ASQ has been used to explore the frequencies of detection fo r children conceived by intracytoplasmic sperm injection and in vitro fertilization (Squires, Carter, & Kaplan, 2003). The ASQ identified children within this population who were in need of referral at the same frequency of detection as the normative sampl e. Researchers investigated the implementation of the ASQ to identify developmental problems in children aged birth 3 years, who have experienced child maltreatment and compared screening results when conducted by child welfare (CW) or early intervention (EI) providers (McCrae, Cahalane, & Fusco, 2011). The ASQ identified children in need of further evaluation at similar frequencies as a national sample. However, frequencies of referral by EI providers were more accurate, as compared to national data, wit h EI providers identifying children at over three times the frequency of CW providers. Authors suggested that CW providers might lack the knowledge or training in child development issues needed to conduct

PAGE 61

61 developmental screenings effectively. Additionally the authors suggested the differences might focus of services. Children of teen mothers are also considered a special population because of the increased probability of environmental risks (Ryan Krause, Mead ows Oliver, Sadler, & Swartz, 2009). standardized instrument, there was little to no agreement. A high frequency of delay was identified when using the standardi zed instrument, but not the parent completed ASQs. The and inaccurate information about child development. In a different study with at risk mothers, it was dem sessment. Mothers at risk who completed at least one ASQ had significantly higher levels of agreement than mothers at risk who did not complete an ASQ. The low agreement reported in Ryan Krause et al. (2009) may indicate the need for improved education abo ut child development for mothers at risk before or as part of completing multi dimensional developmental screening instruments. Validity All seven studies examining the technical properties of the ASQ reported some validity data. In each of these studies, validity was addressed by comparing the classification of children based on their performance on a standardized developmental assessment with their classification based on their parent completed ASQ (Bricker & Squires, 1989b; Bricker et al., 1988; Gollenb erg, et al., 2010; Squires & Bricker, 1991; Squires et al., 1997; Squires et al., 1998; Squires et al., 2009). Studies typically mailed or gave the age appropriate ASQ to parents to complete at home. Evaluations were scheduled within 2 weeks of receiving t he completed ASQ.

PAGE 62

62 Criterion instruments included the Battelle Developmental Inventory (BDI; Newborg, Stock, Wnek, Guidubaldi, & Svinicki, 1984, 2004), Bayley Infant Scales of Development (BSID; Bayley, 1969, 1993), Gesell and Armatruda Developmental and N eurological Examination 1972), and the Stanford performance was typically classified as fail on the standardized test when scores were 1.5 SD below the mean (standard score <75), with the exception of one study where fail was designated for scores 2 SD below the mean (Gollenberg, Lynch, Jackson, McGuinness, & Msall, 2010). Criterion validity For al l studies, criterion validity was reported as a percent agreement on classification of children as either pass or fail Results are summarized in Table 2 5. Classification agreement between the ASQ 3 and the BDI was 86% (Squires et al., 2009). For prior ed itions of the ASQ, agreement ranged from 79 93% (Bricker & Squires, 1989b; Bricker et al., 1988; Gollenberg, et al., 2010; Squires & Bricker, 1991; Squires et al., 1997; Squires et al., 1998). In a study that investigated the effectiveness of the ASQ wit h mothers of low income status versus middle income status, there were no statistically significant differences in percent agreement with the criterion instrument between these two groups (Squires et al., 1998). One study reported correlations between the ASQ and criterion instrument (Gollenberg et al., 2010). Statistically significant correlations were shown between the BSID II Mental Scale and ASQ communication ( r = .52, p < .001) and personal social ( r = .45, p < .01) domains. The BSID II Motor Scale sig nificantly correlated with the gross motor domain ( r = 0.46, p < .01) of the ASQ. Sensitivity The sensitivity of a screening instrument is defined as its ability to accurately refer children who are identified on a standardized instrument as having a dela y (Glascoe, 1991; Glascoe, 2007a). Sensitivity is calculated by the formula: true positives/(true positives + false

PAGE 63

63 negatives). In the reference to screening results, positives designate children who fail screening or standardized evaluation (i.e., childre n who are positive for a condition) and negatives designate children who pass screening or standardized evaluation (i.e., children who are negative for a condition). Sensitivity for the current edition of ASQ is .86 (Squires et al., 2009), exceeding the re commended criteria of .70 .80 (Brassard & Boehm, 2007; Glascoe, 1991; Glascoe, 2007a). Studies investigating prior editions of the ASQ showed sensitivities ranging from .54 1.00 (Bricker & Squires, 1989 b ; Bricker et al., 1988; Gollenberg et al., 2010; Squires et al., 1997). Results are displayed in Table 2 5 Specificity The specificity of a screening instrument is an indicator of its ability to identify children (using the designation of pass ) who are not identified as having a delay in a standardize d evaluation (Glascoe, 1991; Glascoe, 2007a). Specificity is calculated by using the formula: true negatives/(false positives + true negatives). The ASQ 3 has demonstrated specificity of .86 (Squires et al., 2009). As shown in Table 2 5, studies of prior e ditions showed specificities of .80 .95 (Bricker & Squires, 1989b; Bricker et al., 1988; Gollenberg et al., 2010; Squires et al., 1997; Squires et al., 1998). The recommended criteria for specificity of screening instruments is .70 .80 ( Glascoe, 1991; Glascoe, 2007 a). All of the studies of the ASQ have demonstrated results that have met this criterion. Other measures of accuracy Another method to evaluate the accuracy of a screening instrument is to examine indicators that show the proportion of child ren who are misclassified: false positive ratio, false negative ratio, underreferral rates and overreferral rates. False positive ratios are the percentage of children who receive failing scores on a screening instrument, but receive passing scores on a st andardized developmental assessment (Glascoe, 1991; Glascoe, 2007a). False negative ratio is the percentage of children who receive a passing score on a

PAGE 64

64 screening instrument, but have failing scores on a standardized developmental assessment. Both of these numbers are used to calculate the proportion of children who are under or overreferred. During the standardization of the ASQ 3, false positive and negative ratios were both reported to be 14% (Squires et al., 2009). This resulted in an underreferral rat e of 6% and overreferral rate of 8%. Rates for prior editions of the ASQ are displayed in Table 2 5. Reliability For developmental screening instruments, the recommended criterion for interrater and test retest score reliability is 80% (Glascoe, 1991; Gla scoe, 2007a; Salvia & Ysseldyke, 1985). Five of the seven ASQ studies reported interrater or test retest score reliability data (Bricker & Squires, 1989b; Bricker et al., 1988; Squires & Bricker, 1991; Squires et al., 1997; Squires et al., 2009). The resul ts of these studies are summarized in Table 2 6. Interrater score reliability Parents were given the ASQ to complete at home and return by mail (Bricker & Squires, 1989b; Bricker et al., 1988; Squires & Bricker, 1991; Squires et al., 1997; Squires et al. 2009). Within 2 weeks, parents brought the children to an evaluation in which a standardized instrument was administered by a trained examiner who was blind to the results of the parent completed ASQ. After administration was completed, examiners complet ed an ASQ. Interrater score reliability was measured as percent agreement on the classification of children. Based on 107 children, interrater score agreement on the ASQ 3 was 93% (Squires et al., 2009). The first and second editions showed levels of agree ment ranging from 87 100% (Bricker & Squires, 1989b; Bricker et al., 1988; Squires & Bricker, 1991; Squires et al., 1997). In three of these studies, a portion of the sample could not be included in the analyses because examiners were unable to complete some of the items (Bricker & Squires, 1989b; Squires & Bricker, 1991; Squires et al., 1997). Agreement was not reported separately by domain.

PAGE 65

65 Test retest reliability As mentioned above, parents were given the ASQ to complete at home. For a subsample of c hildren, parents completed a second ASQ within 2 weeks during the 1997; Squires et al., 2009). Based on 142 children, the ASQ 3 has test retest score reliabilit y of 93% (Squires et al., 2009), while prior editions have reported between 91 99% (Bricker & Squires, 1989b; Bricker et al., 1988; Squires et al., 1997). Validated for use in other countries The ASQ 3 is available in Spanish for use in the United Stat es (Guiberson & Rodriguez, 2010; Squires et al., 2009). The English version of the ASQ has been investigated for use in Canada and Australia (Elbers, Macnab, McLeod, & Gagnon, 2008; Frisk et al., 2009; Skelle rn, the ASQ has been translated for investigation of utility in Denmark (Klamer, Lando, Pinborg, & Greisen, 2005), France (Dionne, Squires, Leclerc, Peloquin, & McKinnon, 2006; Flamant et al., 2011), Korea (Heo, Squires, & Yovanoff, 2008), Netherlands (Kerstje ns et al., 2009), Norway (Janson & Squires, 2004; Richter & Janson, 2007), Spain (Campos, Squires, & Ponte, 2011), Taiwan (Tsai, McClelland, Pratt, & Squires, 2006) and Turkey (Kapci, Kucuker, & Uslu, 2010). Generally, the technical properties and the dist ribution of scores of children from other countries have been similar to findings with samples from the United States. Studies that Examined the Utility of the ASQ within a Screening Process Seven of the 45 studies implemented the ASQ in applied settings as part of a screening process. The primary research questions concerned its feasibility and acceptability with different contexts and children. The next several sections describe aspects of these studies in relation to settings, detection of children in n eed of referral, time and cost for implementation, potential

PAGE 66

66 benefits for children and families, and implementation of the ASQ in screening programs located in other countries. Settings The ASQ has been primarily investigated as a screening instrument wit hin a developmental surveillance model in primary care clinics (Earls, Andrews, & Hawy, 2009; Hix Small, Marks, Squires, & Nickel, 2007; Jee et al. 2010; Marks et al., 2009). One study investigated the effect of systematic education about the use of ASQ o knowledge and preferences (Thompson, Tuli, Saliba, DiPietro, & Nackashi, 2010). The ASQ has been implemented in a Healthy Start home visiting program that supports children and families from birth through 5 years old (Squires, Katzev & Jenkins, 2002). Home visitors assisted families with reading and answering questions. Detection of children in need of referral After implementing a parent completed screening instrument as part of developmental surveillance, several studies reported increased rates of detection of children needing further evaluation (Earls et al., 2009; Hix Small et al., 2007; Jee et al., 2010; Marks et al., 2009). These primary care centers typically relied on clinical judgment to determine if a referral was necessar y. Prior to the use of the ASQ, physicians were most likely to refer only if the child had delays in multiple domains, only gross motor delays, or only communication delays (Earls et al., 2009; Hix Small et al., 2007). Jee et al. (2010) reported increased referral rates in all domain areas, but the greatest increase was observed in fine motor and personal social referrals. Systematically collecting information from parents about all domains of development improved kill areas in which they may have been less familiar. Another study examined identification rates in children born pre term compared to a group of children born full term and found increases in detection rates were greater for pre term children

PAGE 67

67 (Marks et a l., 2009). Early identification within this at risk population is important and it appeared that incorporating information from parents was beneficial for the physician. Time and cost for implementation Besides considering the technical properties of a scr eening instrument, it is recommended that screening instruments are relatively brief in order to ease adoption into the constraints of a busy practice or other early childhood setting (Brooks Gunn & Lewis, 1983; First & Palfrey, 1994; Meisels & Wasik, 1990 ; Scott & Carran, 1989). This includes the time it takes to score and interpret results. In two studies, it took very little time to implement the ASQ when totaling the amount of time it took to give the ASQ to parents, explain its purpose, provide instruc tions, score, interpret results, and enter results into the electronic medical chart (Hix Small et al., 2007; Jee et al., 2010). Training office staff on the entire screening process took approximately 30 minutes. Additionally, the itemized cost per patien t ranged from $1.61 to $2.43 (Hix Small et al., 2007). Costs varied depending on whether the ASQ was completed in office or mailed back and the amount of time required for physician follow up. Benefits for children and families In order for children and f amilies to benefit from the implementation of the ASQ, health care clinics and early education and care centers need to be using it in practice. Despite the numerous criticisms on its technical adequacy, the DDST or Denver II remains the developmental scre ening instrument of choice, particularly by physicians (Glascoe et al., 1992; Thompson et al., 2010). To address this gap between research and practice, the ASQ was incorporated into a pediatric resident education program (Thompson et al., 2010). Residents received a didactic lecture about three screening tools: Denver II, ASQ, and PEDS (Glascoe, 2007b). Following the lecture, residents completed a 30 minute structured clinical observation on the administration of each instrument and completed four evaluati ons in which all three tools

PAGE 68

68 were used with each child. Prior to the lecture, residents reported having little knowledge of the ASQ, whereas the majority reported having experience with the Denver II. Following the four evaluations, residents reported the highest preference for the ASQ, particularly for its quality incorporating the ASQ into developmental surveillance improved the likelihood that they will use this practice. The ASQ was incorporated into a home visiting program for first time parents (Squires et al., 2002). Beginning when the child was approximately 4 months of age, home visitors assisted parents in reading and completing the items during sessions. After scoring the items together, a sheet of activities was given to the parents to try before the next screening interval. Staff reported incorporating the ASQ into their sessions assisted in providing structure for the visits and when suggesting games a nd activities for the parents. They shared that parents from low income families were able to complete the items with little trouble. Staff noted the ASQ engaged parents in viewing their children using a strength based perspective. Implementation in othe r countries International researchers have investigated the feasibility and acceptability of incorporating the ASQ in developmental screening programs or in the identification of special populations of children. Studies have been conducted in Australia (G asson & Piek, 2003; 1996; Rydz et al., 2006), Denmark (Plomgaard, Hansen, & Greisen, 2006), Ecuador (Handal, Lozoff, Breilh, & Harlow, 2007a, 2007b), and France (Troude La Rochebrochard, 2011). The ASQ was used in a multinational study that included 19 countries (Yu et al., 2007). The findings of these studies suggest that the ASQ was equally effective in these countries when incorporated into developmental screening programs.

PAGE 69

69 Support for the Use of the Ages and Stages as a Multi Dimensional Developmental Screening Instrument Upon review of the empirical studies on the ASQ, the ASQ generally meets recommended standards for technical adeq uacy as a screening instrument. The normative sample of the ASQ 3 was a large, relatively representative sample of children and the technical manual reported adequate levels of concurrent score validity, specificity, sensitivity, and other measures of accu racy. Reports of interrater and test retest score reliability showed the ASQ can yield similar information from different informants who have observed the child. Several studies examined the utility of the ASQ as a tool to systematically incorporate parent developmental screening programs in primary care and home visitation settings. Researchers reported improved rates of detection of children needing further evaluation. The ASQ has been shown to be a time and cost efficient procedure fo r screening children. Most importantly, incorporation of the ASQ represented improvements in services for children and families. Parents were actively engaged in child assessment and, in Squires et al. (2002), use of the ASQ resulted in reports of improved development and learning. Summary Growing numbers of children are being cared for in contexts outside of their home, particularly in early education and care programs. As developmental scre ening is increasingly incorporated into these programs in response to recommendations by prominent organizations and changes in federal and state policies screening can serve dual purposes of early identification and a method to share formally perspective provided across multiple settings. A review of the historical background of developmental

PAGE 70

70 screening reveals that while pare nt involvement has been long regarded as a recommended practice in screening, it has not been implemented consistently. Difficulties with incorporating parent input have been partially due to vague policies and guidelines for practitioners. In the past d ecade, numerous organizations have provided policy recommendations and position statements on either developmental screening or the assessment of young children, including screening. Collectively, these organizations unanimously support the gathering of mu ltiple sources of evidence to inform assessment, specifically, the inclusion of parent input in the decision making process. Using parent completed developmental screening instruments fits within the constraints of busy medical practices, early education a nd care programs, and school districts and is likely to improve the ecological validity of results. An investigation of the literature on the relationship between parent and professional ratings as measured with multi dimensional developmental screening i nstruments identified 10 studies demonstrating moderate to high agreement or score congruence However, several limitations were identified in these studies that might preclude valid inferences about parent ECE congruence. Typically, professionals in the s tudies were individuals who had limited contact with the child, which might affect score validity. There were weaknesses in the methodology used to gather and interpret score congruence. The majority of studies used percent agreement between parents and pr ofessionals on overall classifications of children, without investigating differences by age, developmental domain, or length of contact with the child. Understanding variables that affect similarities and differences in ratings can inform how information should be gathered, shared, and used between adults who spend significant time with young children in different contexts (e.g., home and school). A review of the empirical literature on the Ages and Stages Questionnaire has provided evidence of its utilit y as a multi dimensional

PAGE 71

71 developmental screening that can be used in different contexts. The literature review also highlighted the need to conduct additional investigations of congruence and to explore the potential of the ASQ as a tool for facilitating t development and learning across home and early education and care contexts. The current literature comparing developmental screening instruments completed by parents and professionals cannot adequately address the se issues, as the professionals included in research studies typically were unfamiliar with the child outside of assessment context.

PAGE 72

72 Table 2 1. Select technical properties and recommended criteria for screening i nstruments Technical p roperty Definitio n Recommended c riteria Standardization Clear description of groups included in standardization sample ; standardization sample representative of the population for which the test is intended; children with disabilities are included in sample or special stu dies are reported in the manual. 100 subjects per age group. Score r eliability The degree to which scores are free of measurement error for a given group. Test retest and interrater reliability of .80 or greater. Score v alidity The degree to which accu mulated evidence and theory support the proposed interpretation of test scores. Correlation of .70 .80 with a test battery that samples the same range of developmental tasks. Sensitivity The proportion of cases in which an instrument refers a child who i s later determined to have a delay. Sensitivity is calculated by true positives/(true positives + false negatives). .70 to .80 Specificity The proportion of cases in which an instrument does not refer a child who does not have a delay. Specificity is c alculated by true negatives/(false positives + true negatives). .70 to .80

PAGE 73

73 Table 2 2. Sample c haracteristics of studies examining the relationship between parent and professional informants using developmental screening instruments. Cita tion Sample characteristics Ages n Bricker & Squires, 1989b Infants with medical risk factors 4 24 months 39 Bricker, Squires, Kaminski, & Mounts, 1988 Infants with medical risk factors 4 24 months 73 Burgess, Asher, Doucet, Reardon, & Daste, 198 4 Sampled from pediatric clinic 6 60 months 60 Chiu & DiMarco, 2010 Homeless children, none chronically ill or with developmental delay 4 60 months 21 Coghlan, Kiing, & Wake, 2003 Sampled from child care centers and kindergartens in Australia 18 72 months 262 Frankenburg, Fandal, & Thornton, 1987 a (parent examiner) Sampled from pediatric clinics 2 weeks 72 months 19 3 Frankenburg et al., 1987 a (parent teacher) Sampled from day care centers 2 weeks 72 months 71 Frankenburg, van Doorninck, Liddell, & Dick, 1976 a (parent examiner) Sampled from community 3 72 months 1,115 Frankenburg, von Doorninck, Liddell, & Dick, 1976 a (parent teacher) Sampled from day care centers 3 72 months 202 Glascoe, 2002 Sampled across United States 2 weeks 24 months 408 Squires & Bricker, 1991 Children with environmental risk factors 4 30 months 14 Squires, Bricker, & Potter, 1997 Children with medical risk factors, environmental risk factors, and normative sample 4 36 months 112 Note. Medical risk factors include more than 3 days spent in the NICU or referred for evaluation of medical or developmental concerns. Environmental risk factors included household income below the federal poverty line, maternal age <19 years at time of birth, maternal educ ation <12 th grade, or history of parental involvement with Child Protective Services. a Studies selected a subsample of participants to receive additional screenings from day care providers. Subsamples are reported for parent examiner agreement and parent t eacher agreement.

PAGE 74

Note. IMQ = Infant Monitoring Questionnaires. DDST = Denver Developmental Screening Test. R PDQ = Revised Denver Prescreening Developmental Questionnaire. DDST = Denver Developmental Screening Test. PDQ = Denver Prescreenin g Developmental Questionnaire. PDQ M = Denver Prescreening Developmental Questi onnaire, by maternal interview. pmental Status. *p v alue < 0.05 **p value < 0.01 Table 2 3 Summary of s tudies examining parent professional relationship on developmental screening instruments. Citation Professional r ole Professional s creener Parent s creener Analysis Results Bricker & Squires, 1989b Examiner IMQ I MQ Classification agreement Overall agreement 87% Bricker, Squires, Kaminski, & Mounts, 1988 Examiner IMQ IMQ Classification agreement Overall agreement 97% Burgess, Asher, Doucet, Reardon, & Daste, 1984 Examiner DDST PDQ or PDQ M Correlation of DQ r = 0.478** (PDQ) r = 0.597** (PDQ M) Chiu & DiMarco, 2010 Examiner Denver II ASQ Classification agreement Personal social 95% Gross Motor 95% Fine Motor 71% Language 67% Coghlan, Kiing, & Wake, 2003 Day care or Kindergarten provider PEDS PEDS It em agreement ; Kappa values Other concerns 95.8%, = 0.02 Gross motor 94.7%, = 0.40** Global/cognitive 92.4%, = 0.04 Fine motor 91.6%, = 0.11 Receptive language 90.8%, = 0.18** School 90.8%, = 0.28** Self help 87%, = 0.23** Soc ial emotional 79.8% = 0.37** Expressive language 77.5%, = 0.26** Behavior 64.9%, = 0.13* Frankenburg, Fandal, & Thornton, 1987 Day care provider R PDQ R PDQ Classification agreement Parent teacher 83% Examiner DDST R PDQ Classification agreement Parent examiner 74% Frankenburg van Doornick, Liddell, & Dick, 1976 Day care teacher PDQ PDQ Classification agreement Parent teacher 59% Examiner DDST PDQ Item Agreement Parent examiner 93% Glascoe, 2002 Examiner Brigance Infant an d Toddler Screen Brigance Infant and Toddler Screen Comparison of mean scores F[5,170] = 0.543 Squires & Bricker, 1991 Examiner IMQ IMQ Classification agreement Overall agreement 100 % Squires, Bricker, & Potter, 1997 Examiner ASQ ASQ Classification ag reement Overall agreement 94%

PAGE 75

Note. N rep resents the total number of ASQ s included in the anal yses when children were sampled over time. Normative samples were defined as having no previous history of developmental or serious health problems, full term (>37 weeks) at birth, and never assigned to a neonatal intensive care unit (NICU). Medical risk f actors include more than three days spent in the NICU or referred for evaluation of medical or developmental concerns. Environmental risk factors included h ousehold income below the federal poverty line, maternal age <19 years at time of birth, maternal ed ucation <12 th grade, or history of parental involvement with Child Protective Services a Included ASQ sample from Bricker et al., 1988 and Bricker & Squires, 1989b. Table 2 4 Demographic information of participants descri bed in psychometric studies of A ges and S tages Q uestionnaires Child characteristic Risk factor represented Additional information Citation N n 2 24 months 25 36 months 37 60 months Ethnicity Medical Environmental Normative sample Maternal education Family income Infant Monitoring Questionnaires Bricker, Squires, Kaminski, & Mounts (1988) 653 264 Bricker & Squires (1989b) 743 265 Squires & Bricker (1991) 18 18 Ages and Stages Questionnaires Second Edition Squires, Bricker, & Potter (1997) a 4,145 2,008 Squires, Potter, Bricker, & Lamorey (1998) 247 96 Gollenber g Lynch, Jackson, McGuinness, & Msall (2010) 40 40 Ages and Stages Questionnaires Third Edition Squires, Twombly, Bricker, & Potter (2009) 18,572 15,138

PAGE 76

Note. instrument when scores were 1.5 SD below the mean. Gesell = Revised Gesell and Armatruda Developmental and Neu rologic Examination; BSID = Bayley Scales of Infant Development; Stanford Binet = Stanford Binet Intelligence Test; MSCA = a Included sample from Bricker et al., 1988 in analysis b Included samples from Bricker et al., 1988 and Bricker & Squires, 1989b in analyses. c Fail was designated by scores 2 SD below the mean. Table 2 5 Results of validity studies of Ages and S tages Questionnaires (in percentages) Citation N Criterion instrument Agreement on classification Sensitivity Specificity False positive False negative Under referral Over referral Infant Monitoring Questionnaires Bricker & Squires (1989b) 743 Gesell, BSID, Stanford Binet 79 92 63 a 91 a 7 6 Bricker, Squires, Kaminski, & Mounts (1988) 653 Gesell 79 94 54 95 3 16 1 9 Squires & Bricker (1991) 18 BSID 89 Ages and Stages Questionnaire Second Edition Squires, Bricker, & Potter (1997) b 4,145 Gesell, BSID, S tanford Binet, MSCA 84 75 86 14 2 3 4.0 11.9 Squires, Potter, Bricker, & Lamorey (1998) 247 BSID 80 93 8 5 89 0 5 6 20 Gollenberg Lynch, Jackson, McGuinness, & Msall (2010) 40 BSID II 88 c 100 93 0 8 Ages and Stages Questionnaire Th ird Edition Squires, Twombly, Bricker, & Potter (2009) 18,572 BDI 86 86 86 14 14 6 8

PAGE 77

77 Table 2 6 Results of score reliability s tudies of Ages and Stages Questionnaires Citation Interrater Reli ability ( n ) Test Retest ( n ) C GM FM PS P Infant Monitoring Questionnaires Bricker & Squires (1989b) 87% (39) 91% (107) Bricker, Squires, Kaminski, & Mounts (1988) 97% (73) 99% (68) Squires & Bricker (1991) 10 0% (14) Ages and Stages Questionnaires Second Edition Squires, Bricker, & Potter (1997) a 94% (112) 94% (175) .63 .75 .53 .87 .49 .79 .52 .75 .52 .68 Ages and Stages Questionnaires Third Edition Squires, Twombly, Bricker, & Potter (2009) 93% (107) 92% (145) .57 .83 .57 .87 .51 .83 .53 .78 .51 .71 Note. C = communication domain; GM = gross motor domain; FM = fine motor domain; PS = problem solving domain; P = personal social domain. a Included sample from Bricker et al 1988 and Bricker & Squires, 1989b.

PAGE 78

78 CHAPTER 3 METHOD S Few empirical studies have been conducted to investigate parent and professional congruence as measured by a multi dimensional developmental screening instrument when the same instrument is used and completed in the same way by each informant. The purpose of the present study was to contribute to the research on congruence between parent and professional judgment by recruiting a group of professionals (i.e., ECEs) who have consistent and f requent contact with the children. The following research questions were addressed in the present study: 1. When using the same multi dimensional developmenta l screening instrument, what is overall congruence between two informants who interact wi th the chil d regularly ? 2. Does congruence between parents and early childhood educators vary based on child age cohort ? 3. Does congruence between parents and early childhood educators vary depending on the child developmental domain assessed ( i.e ., communication, gross motor, fine motor, problem solving, personal social)? 4. Are there differences in the pattern of response selection across informant types (i.e., parents and ECEs)? 5. Does child age and length of relationship with the early childhood educator predict item le vel congruence between parents and ECEs? This chapter include s information about the design of study, setting participants, measures, procedures and data analyses The description of procedures includes sampling and recruitment of subjects, instructions provided to participants, and data collection activities Design of Study A cross sectional, exploratory correlational research design was used. Parents and ECEs from two university based early education and care centers were recruited at a single point in time to compare their responses about child developmental status on a multi dimensional developmental screening instrument. In the current study, congruence was defined as the

PAGE 79

79 correspondence between the judgments of two informants. This terminology was selected to avoid using the value estimation or under development. Setting This study w as conducted at two university based early education and care centers. These centers provide early education and care to children age s 6 weeks through 5 years. The centers serve families who attend or are employed by a large university located in the southeastern United States. O ne center serv es the entire university community while the second center serves families affi liated with a n academic health sciences center. At the time of recruitment enrollment across the t wo centers was 268 infants, toddlers, and preschoolers. Both centers have accreditation from the National Association for the Education of Young Children (NA EYC). The centers utilize a team teaching approach; two lead teachers collaboratively plan and teach all lessons and activities within each classroom. As a mini mum requirement, ECEs must have a Child Development Associate (CDA) credential. To fulfill the NAEYC accreditation requirement specifying that centers must conduct forms of assessment are used in the program ECEs complete the Ages and Stages Questionnaire ( ASQ 3; Squires & Bricker, 2009) for all children serially and at regular intervals to screen for making related to the need to monitor more o refer for more comprehensive evaluation. All full time staff are required to attend yearly training, provided by the student investigator of this study, on using and interpreting the ASQ 3. In the present study, ECEs and parents consented to complete AS Qs for children enrolled in the early education and care

PAGE 80

80 centers. Both informants completed ASQs for the same child, even if the child was not due for his or her regular ASQ screening interval at the center. Participants Early Childhood Educators ECEs w e re eligible to participate in the study if they worked full time (i.e., approximately 40 hours per week) in a classroom and h ad completed an ASQ 3 training provided by the center within the past 2 months. Fifty one full time early childhood educators were employed across the two centers at the time the present study was conducted, however, six of the ECEs were not regularly assigned to one classroom and were therefore not eligible for the present study. Forty four of the 45 eligible ECEs consented to partic ipate in the present study. During data collection, four teachers were removed from the pool of eligible ECEs because of maternity leave ( n = 1) and reassignment to a new classroom within the center ( n = 3). These ECEs were each able to complete 1 to 2 ASQ s and therefore are included in the description of participants below. Of the 44 participating ECEs, the majority were female ( 97.7%) ECE participants identified their race/ethnicity as European or Caucasian (68.2%), Latino or Latino American (15.9%), Afr ican or African American (9.1%), or mixed ethnicities (6.8%). The majority of special education (31.8%) or child development (20.5%). Forty three percent of th e ECEs reported having a child development associate (CDA) credential. ECEs had been employed with the centers for an average of 26.2 months (range = 4 150, SD = 32.4). Thirty one ECEs had taught young children prior to working at the centers. Mean years of experience working with young children were reported to be 8.9 years (range = 1 40, SD = 8.1).

PAGE 81

81 Parents and Children Eligible parents for the present study were those who had already consented for their child to participate in regular screenings at the center and those not employed by the centers. In addition, the child of each eligible parent needed to be enrolled in an classroom for a t least 1 month before the parent ECE pair completed the ASQ 3. Some parents completed for ASQs for siblings, and subsequently, there were 110 parent participants in the study (detailed further below) The parents who completed the ASQ 3 forms represented 91 mothers/step mothers (82.7%), 17 fathers/step fathers (15.5%), 1 legal guardian (.9%), and 1 fos ter parent (.9%). Of these respondents, 83.6% were female ( n = 92) and 16.4% were male ( n = 18). Parents self reported their affiliation with the university where the early education and care centers were based: 28.2% were faculty, 26.4% were staff, 26.4% were graduate or professional students, 8.2% were fellows, 7.3% were categorized as other (e.g., not affiliated due to merger with an off campus center or other type of hospital affiliation), and 2.7% were undergraduate students. The majority of parents w ere born in the United States (68.2%). The race/ethnicity of the parents was reported as follows: 65.5% European or Caucasian, 17.3% Asian/Pacific Islander or Asian American, 7.3% Latino or Latino American, 4.5% mixed ethnicities, and 4.5% African or Afri can American. Over one quarter ( n = 28) of the parents reported speaking a language other than English in the home. There were 126 children in study sample. Table 3 1 shows the distribution of children across the four age cohorts. The largest age cohort r epresented was 42 60 months ( n = 44), followed by 27 36 months ( n = 38), 14 24 months ( n = 27), and 2 12 months ( n = 17). The mean age of the study sample was 32.9 months (range = 3.6 64.5, SD = 17.3). In the sample, 58.7% of the children were ma le and 41.3% were female. The race/ethnicity of the children were reported as follows: 57.9% European or Caucasian, 18.3% mixed ethnicity, 14.3% Asian/Pacific

PAGE 82

82 Islander or Asian American, 6.3% Latino or Latino American, and 3.2% African or African American. Thirty six of the children were considered to be dual language learners and of these children, 12 spoke English as their second language. Data were collected on two risk variables identified as relevant for interpreting ASQ scores in the ASQ 3 manual (Sq uires et al., 2009). These risk variables were prematurity and having a history of concerns. There were 11 children (8.7%) who were born three or more weeks prematurely. Of these children, only one child received an ASQ that was adjusted for prematurity, a s recommended in the ASQ 3 manual (Squires et al., 2009); the remaining children regularly received ASQ screenings based on their chronological age. Seventeen children (13.5%) had been evaluated at least once for medical or developmental concerns. Measure s Demographic Forms After giving written informed consent, ECEs and parents completed demographic forms. Parents demographic forms were included in the packet that included the ASQ 3 forms. For each child who received an ASQ 3, the ECE completed an addit ional demographic form to collect information on the length of his or her relationship with the specified child. Ages and Stages Questionnaire The Ages and Stages Questionnaire Third Edition ( ASQ 3; Squires & Bricker 2009) is a multi dimensional devel opmental screening instrument for use with children 1 to 66 months of age The ASQ 3 was designed with the intent that parents, or other caregivers who know the child well, complete the questionnaires Parents or other caregivers complete the ASQ 3 by read ing and responding to a series of questions about developmental skills appropriate for children at different age intervals. They are instructed to indicate if the child is doing the skill

PAGE 83

83 regularly (as yes ), sometimes, or not yet T here are 21 questionnaire intervals that are used for children from 1 66 months (Squires et al., 2009). Each questionnaire indicates the age range applicable for use F or example, the 12 month questionnaire is used for children between the ages of 11 months, 0 days through 12 months, 30 days. Each questionnaire has 30 items divided evenly between five developmental domains: communication, gross motor, fine motor, problem solving, and personal social. The distribution of the number of items represented in each developmental domain remains consistent across the age intervals, but the pool of items varies for each age interval. Some specific items may be repeated across only two age intervals, while other items repeat more frequently. The ASQ is scored b y converting r esponses to points : each yes response is 10 points, each sometimes response is 5 points, and each not yet response is 0 points. Points are summed to obtain a score for each domain and transferred to a summary page. Age based cut offs are prov ided on the summary page and indicate whether scores for each developmental domain fall in one of the following categories: above the cutoff close to the cutoff (referred to as monitoring zone or at risk ) and below the cutoff Scores that are above the c utoff are interpreted as the is on schedule and no further action is needed. Scores close to the cutoff fall in the monitoring zone and it is suggested that learning activities and additional monitoring be conducted If scores are below the cutoff further assessment might be needed. Validity of ASQ scores A detailed review of score validity studies conducted for the three editions of the ASQ was provided in Chapter 2. This section focus es on the development of the ASQ questionnaires comparison of performance on questionnaires by non risk and risk groups, and the determination of cutoff scores. This section end s with a brief review of classification agreement between ASQ

PAGE 84

84 3 and the first and second editions of the Battelle Developmental Inventory (Newborg, et al., 1984, 2004). Develo pment of the ASQ The content of the ASQ was informed by review ing (a) standardized developmental tests published when the first version of the instrument (i.e., Infant Monitoring Questionnaire, IMQ) was dev eloped, (b) nonstandardized tests focus ed on early development (c) textbooks, and (d) literature about early developmental milestones (Squires et al., 2009). Items developed for the ASQ had to meet the following criteria: (a) skills could be easily observ ed or elicited by parents, (b) skills were highly likely to occur within homes and child care settings, and (c) skills reflected important developmental milestones. Developmental s kills that met these criteria were written into items using concrete terms a nd vocabulary that did not exceed a sixth grade reading level. To narrow the large pool of potential items, a developmental quotient (DQ) was calculated for each item using Equation 3 1. (age equivalent)/(age interval of ASQ item) X 100 = DQ (3 1 ) In this formula, the age equivalent was derived from the original source of the item (Squires et al., 2009). For example, if the item was derived from the Battelle Developmental Inventory (Newborg et al., 1984, 2004), the age equivalent specified in the BDI for a skill was the one used in this formula. The final pool of items for each age ASQ interval questionnaire was selected if the DQ range d from 75 to 100. This criterion was selected to restrict items for a particular questionnaire interval to those t hat targeted a skill at the middle to low end of the developmental range for that particular chronological age interval. The rationale was that a child who was not observed to use a skill calculated at a DQ of 75 to 100 might benefit from further assessmen t.

PAGE 85

85 Comparison of non risk and risk groups. A sample of 18,572 ASQ questionnaires for children aged 1 to 66 months was used to determine the cutoff scores for the ASQ 3 (Squires et al., 2009). At the time of ASQ 3 completion, families were asked to indica te whether the child had any environmental or medical risk factors. Environmental risk was defined as family income level below the federal poverty guideline, maternal age of 19 years or younger at the time of the e 12th grade, or a history of involvement with child protective services. Medical risk was defined as prematurity (< 39 weeks gestation) or infant birth weight less than 3 pounds, 5 ounces. Using these guidelines, 19% ( n = 2,186) of the ASQ standardization sample ( n = 18,572) had two or more risk factors. When mean scores were compared between the non risk and risk groups, means for the non risk group were generally higher than the means for the risk group. The ASQ (Squires et al., 2009) prov ides graphic displays of the means of non risk and risk groups across all ages, for each developmental domain. The authors noted that there were some exceptions in which the risk group had consistently higher means than the non risk group, although they at tributed these anomalies to low representations of children at risk at certain age intervals. Determination of cutoff scores. Non risk and risk groups were combined for analyses to determine cutoff scores (Squires et al., 2009). The authors u sed two stra tegies to determine cutoff points: (1) receiver operating characteristic (ROC) curves and (2) comparing the percentage of children identified as needing further evaluation using three different cutoff scores. In both of these strategies the authors compar ed cutoff scores for each questionnaire interval that were at 2, 1.5, and 1 standard deviations below the mean score ROC analyses were applied to compare conditional probabilities for sensitivity, specificity, false positive proportion, false negative p roportion, under identification, and over

PAGE 86

86 identification, for each of the cutoff points (Squires et al., 2009). Analyses were conducted for each questionnaire interval. The authors were interested in selecting a cutoff score that represented a reasonable b alance between over identification and under identification rates. Over identification was defined as the proportion of children identified as needing further assessment by the ASQ, but who received scores in the typical range (scaled score above 75) on th e BDI or BDI 2. Under identification was defined as the proportion of children identified as not needing further assessment by the ASQ, but receiving scores below the typical range (scaled score of 75 and below) on the BDI or BDI 2. When cutoff scores wer e set at 1 and 1.5 SD below the mean, the over identification rate increased as the under identification rate decreased. Cutoff scores were set at 2 SD below the mean because they were the most balanced in terms of the true positive and false positive prop ortions for all 21 questionnaires. The second strategy entailed comparing the percentage of children that fell below each of the cutoff scores when set at 2, 1.5, and 1 SD below the mean (Squires et al., 2009). The authors referenced U.S. Census Bureau a nd Centers for Disease Control and Prevention prevalence data for young children with developmental disabilities to set target identification percentages. It was projected that questionnaires should identify 12% 16% of children for referral in one develo pmental area, while 2% 7% of children should be identified for referral in two or more areas. Using these percentages, 2 SD below the mean was selected to be the cutoff point for all 21 questionnaires. C lassification agreement. For the ASQ 3, classific ation agreement was measured by comparing the classification of children based on their performance on the B attelle Developmental Inventory first and second editions ( BDI, Newborg et al., 1984 ; BDI 2, Newborg et al., 2004) with their classifications on the ASQ 3 (Squires et al., 2009). Agreement meant that

PAGE 87

87 the child either received typical scores on both the BDI and ASQ 3 or received eligible (for referral) scores on the BDI and ASQ eligible if the c performance on the ASQ 3 was designated as typical above the cutoff or monitoring zone and as eligible below the cutoff A total of 18,572 questionnaires were used for analyses. Overall classification agreement between BDI and ASQ 3 was 85.8%. Classification agreement also was examined for four age cohorts: 2 12 months, 14 24 months, 27 36 months, and 42 60 months. Across these intervals classification agreement ranged from 82.6% (14 24 months) to 88.9% (2 12 months). Reliability of ASQ scores This section d escribe s the internal consistency test retest, and interrater reliability of scores for the A SQ 3 as reported in th e ASQ (Squires et al., 2009) Internal consistency score reliability. The internal consistency score reliability of all questionnaires was examined through correla coefficient alpha (Squi res et al., 2009). For 20 ASQ intervals (9 and 10 month questionnaires were combined), Pearson product moment correlation coefficients were calculated for developmental domain scores with overall ASQ scores. For each developmental domain score correlatio ns with ASQ overall scores were as follows when all questionnaire intervals were combined: communication ( r = .76), gross motor ( r = .65), fine motor ( r = .73), problem solving ( r = .78), and personal social ( r = .79). All correlations were statistically s ignificant at p < .01. These findings suggest moderate to strong internal consistency between developmental domain scores and total test score. For each of the 20 questionnaire intervals, Cronbach coefficient alphas were calculated for developmental do main scores (Squires et al., 2009). Alphas ranged from .51 to .87 across all

PAGE 88

88 developmental areas and age intervals. Table 2 6 shows the alpha ranges for each developmental domain. A lphas specific to each questionnaire are available in the ASQ de (Squires et al., 2009) These alphas indicate that the ASQ items have good to acceptable internal consistency score reliability. Test retest score reliability. To examine test retest score reliability, 145 parents were asked to complete a second quest ionnaire within 2 weeks of completing an initial questionnaire. Agreement was defined as having the same classification (i.e., typical or eligible ) on both questionnaires. The percent agreement for the 145 parents was 92% Intraclass correlation coefficie nts (ICCs) calculated for classification agreement ranged from .75 to .82. Similar results were found in other studies examining test retest reliability of all editions of the ASQ and are displayed in Table 2 6. Interrater score reliability. To examine i nterrater score reliability, 107 children were selected to have questionnaires completed by both parents and trained test examiners (Squires et al., 2009). Parents completed the ASQ at home and mailed in the forms. Within 2 weeks of completion, children we re brought in for a standardized assessment with a trained test examiner who administered the BDI 2 to the child (Newborg et al., 2004) Immediately following the assessment, the examiner completed the age appropriate ASQ for the same child Interrater sco re agreement was defined as having the same classification (i.e., typical or identified) on both parent and examiner completed questionnaires. Across all questionnaires, the percent agreement between parents and examiners was 93 % and the ICCs of developme ntal domain classification agreement ranged from .43 (communication) to .69 (personal social). Results of the ICCs for other developmental domains were not provided.

PAGE 89

89 Procedures Sampling and Recruitment For the present study, a statistical power analysis s uggested a minimum of 30 Parent ECE pairs, up to a maximum of 35 parent ECE pairs would be optimal to recruit within each of the following age cohorts : 2 12 months, 14 24 months, 27 36 months, 42 60 months. These age cohorts were selected to repli cate the age intervals used in validity studies with the ASQ 3 (Squires et al., 2009). The age cohort affiliation for each participant was based on the applicable ASQ 3 age interval for the child For example, according to ASQ 3 f orms, a child who is 25 mo nths, 5 days old would receive the 24 month questionnaire (Squires et al., 2009), and therefore, the parent ECE pair would b e categorized within the 14 24 month age cohort ECEs were recruited prior to sending informational letters about the study home with children enrolled in their classrooms. The student investigator for the present study attended each participate. Of the 45 eligible ECEs, 44 consented to study parti cipation and signed informed consent forms. Recruitment procedures for parents included sending home informational letters and informed consent forms. Distribution of these forms was accomplished by placing them in all n = 262 ), having blank consent forms available in the classrooms and at the front desk of each center, and greeting parents outside of classrooms to inform them about the study. Six families were not eligible to participate in the present study because they were employed by the centers. Signed consent forms were received for 145 children, representing 54.1% of eligible children across both centers. Refer to Table 3 2 for percentages of recruited eligible children by age cohort.

PAGE 90

90 To be in cluded in analyse s, the E CE and parent must have completed the ASQ 3 within 2 weeks of each other This time frame was selected based on interrater score reliability studies of previous and current editions of the ASQ (Bricker & Squires, 1989b; Bricker, et al., 1988; Squires & Bri cker, 1991; Squires et al., 1997; Squires et al., 2009). Five parent ECE pairs returned ASQ forms outside of the 2 week window and were not included in analyses. Fourteen parent ECE pairs had incomplete forms, meaning that consent had been received from th e ECE and parent, an ASQ 3 packet had been given, but either one or both of the packets were not returned. Data for 126 parent ECE pairs across the two centers w ere included in analyses (Table 3 2). In structions for ECE Practitioners and P arents ASQ 3 packets for ECEs and parents had identical set s of instructions describing the completion of the form and the scoring rubric to be used (Appendix) Parents w ere informed they were permitted to discuss the items with other caregivers, but c ould not consult with any staff from the early education and care center and vice versa for ECEs The same respondent was asked to complete the demographic form and ASQ 3 form. The instructional form include d clarification on the three scoring responses ( yes, sometimes, an d not yet ) along with example s that illustrate d how to select the appropriate response on the ASQ 3 questionnaire Data Collection After parents and ECEs consented to participate in this study, each parent ECE pair was given packets containing instructio nal forms (Appendix) demographic forms, and ASQ 3 forms to complete for the same child. The ASQ 3 summary sheets (used to determine cutoff scores and assign classification categories) were not included in either packet, so parents and ECEs were unable to determine scoring classification categories. To ensure that forms were returned within

PAGE 91

91 the 2 week time period, parents and ECEs received a verbal and written reminder within 1 week of receiving the forms. Of the 110 parents who completed an ASQ 3 95 pare nts (86 %) completed an ASQ 3 for 1 child, 14 parents (13%) completed an ASQ 3 for 2 children and 1 parent completed an ASQ 3 for 3 children. ECEs completed ASQs for an average of 2.9 children (range = 1 6, modes = 1, 3). These centers utilize a team tea ching model, subsequently, 44 E CEs were distributed across 24 classrooms. Within these classrooms, ASQs were completed for an average of 5.3 children (range = 1 10, mode = 4). The distribution of questionnaires across the available age intervals is displa yed in Figure 3 1. The questionnaires were not administered at the 2 month or 9 month intervals, as there were no consented participants within the applicable age range for these two age intervals. For the remaining 19 age intervals, the number of question naires distributed ranged from 1 (at the 20 month interval) to 15 participants (at the 33 month interval). Data Analyses Using the scoring procedures described above, parent and ECE completed ASQ 3 forms were scored by the author and checked independent ly by a second person Discrepancies in scor ing were checked against hard copies of the forms and corrections w ere made on the form. Demographic form data were entered into Microsoft Excel and checked using a double data entry procedure. ASQ 3 forms w ere entered into a REDCap database (Harris et al., 2009) by the author and checked by a second person. A function is available within REDCap by which forms initial data entry, a form was m arked as unverified. After a second person looked over the electronic entry and verified all responses ha d been correctly transferred from the hard copy form into the database, the form was marked as ve rified. After this procedure was completed for all p arent and ECE

PAGE 92

92 forms, the RedCap database was exported into an Excel spreadsheet Demographic and ASQ Excel sheets w ere exported into SPSS Version 20 for analyses The following questions were used to guide analyses conducted in the present study: 1. When us ing the same multi dimensional developmental screening instrument, what is overall congruence between two informants who interact with the child regularly? a. What is the item level agreement between parents and ECEs across all ASQ items? b. What is the classifi cation agreement between parents and ECEs across all 5 developmental domains? 2. Does congruence between parents and early childhood educators vary based on child age cohort? a. What is the item level agreement between parents and ECEs for each age cohort? b. What is the classification agreement between parents and ECEs for each age cohort? 3. Does congruence between parents and early childhood educators vary depending on the child developmental domain assessed (i.e., communication, gross motor, fine motor, problem so lving, personal social)? a. What is the item level agreement between parents and ECEs for each developmental domain? b. What is the classification agreement between parents and ECEs for each developmental domain? c. What is the item level agreement between parents and ECEs within each developmental domain for each age cohort? d. What is the classification agreement between parents and ECEs within each developmental domain for each age cohort? e. What is the relationship between parent and ECEs developmental domain raw sco res? 4. Are there differences in the pattern of response selection across informant types (i.e., parents and ECEs)? a. As a group of informants, do parents select responses (i.e., yes, sometimes, not yet, omitted) at similar frequencies to ECEs across all ASQ i tems?

PAGE 93

93 b. As a group of informants, do parents select responses at similar frequencies to ECEs across all ASQ items for each age cohort? c. As a group of informants, do parents select responses at similar frequencies to ECEs within each developmental domain? d. As a group of informants, do parents select responses at similar frequencies to ECEs within each developmental domain for each age cohort? 5. Does child age and length of relationship with the early childhood educator predict item congruence between parents and ECEs? Congruence Examined using Percent Agreement Congruence was examined by calculating percent agreement between parents and ECEs using two approaches. The first approach examined parent and ECE congruence on ASQ items. The second approach examined pare nt and ECE congruence on classification decisions. Congruence on ASQ items Item level agreement between parents and ECEs was calculated across ASQ items to generate a percentage. When parents and ECEs selected the same response, agreement was coded as 1; when different responses were selected, agreement was coded as 0. If one informant omitted the item, this was also coded as 0. Item level agreement was calculated in two ways: (1) by using the three original response choices ( yes, sometimes, not yet ) prov ided in the ASQ 3 (Squires & Bricker, 2009) and (2) by using two response choices generated by considering yes and sometimes as agreement. When calculating agreement for two response choices, agreement was coded as 1 when both informants selected either ye s or sometimes Research questions for which item level agreement was calculated are shown in Table 3 2. To examine overall congruence on ASQ items, the total number of agreements was summed and divided by the total number of items ( v = 30) to generate a percentage (Equation 3 2). Overall item agreement = (Total number of agreements)/30 X 100 (3 2)

PAGE 94

94 Overall item agreement was calculated for each parent ECE pair using Equation 3 2 and then averaged across all 126 parent ECE pairs. To examine differen ces by child age cohort, children were divided into one of four age cohorts based on ASQ interval: birth 12 months, 14 24 months, 27 36 months, and 42 60 months. Agreement was calculated by averaging Equation 3 2 over the parent ECE pairs in each ag e cohort. To examine item level agreement by developmental domain, Equation 3 2 was adjusted to account for the number of items in each domain. The 30 ASQ items are divided evenly among the five domains so that each domain is comprised of six items (Squire s et al., 2009). The formula is shown below (Equation 3 3). Developmental domain agreement = (Number of agreements)/6 X 100 (3 3) To examine developmental domain differences within each age cohort, Equation 3 3 was applied for each of the five devel opmental domains, and then percentages were averaged across the number of parent ECE pairs in each age cohort. Congruence on classification decisions Classification decisions were determined by age based cutoffs for developmental domain scores specified o n the summary sheets (Squires et al., 2009). Classification agreement was calculated in two ways: (1) using either the three original descriptive categories ( above the cutoff, monitoring zone/at risk, and below the cutoff) provided in the ASQ 3 (Squires & Bricker, 2009) or (2) using two categories generated by considering above the cutoff and monitoring zone as agreement. For the latter method, the two classification categories were renamed to typical (i.e., either above the cutoff or monitoring zone) and e ligible (i.e., below the cutoff ) to parallel terminology used in validity studies of the ASQ 3 (Squires et al., 2009). When parent and ECE domain scores translated to the same classification category, agreement was coded as 1; when domain scores translated to different classification categories, agreement was coded as 0. When

PAGE 95

95 calculating classifications using two categories, agreement was coded as 1 if either above the cutoff or monitoring zone was selected. Research questions for which classification deci sion congruence was calculated are shown in Table 3 2. To calculate overall classification agreement, the total number of agreements on each classification decision was added for all 5 domains and divided by the total number to generate a percentage (Equa tion 3 4). Percentages were averaged across all parent ECE pairs in the sample. Overall classification agreement = (Total number of agreements on classification decisions)/5 X 100 (3 4) To examine classification agreement for each age cohort, o verall classification agreement was calculated and averaged across the number of pairs in each age cohort using Equation 3 4. This equation was simplified when calculating classification agreement by developmental domain. Classification decisions are based on the total score within each developmental domain. Therefore the only possible values for each parent decision for each developmental domain is either 0% or 100%. The number of parent ECE pairs who agreed for the c lassification decision were summed and divided by the total number of children ( N = 126) to represent the proportion of pairs with agreement on domain classification (Equation 3 5). This formula was applied for each developmental domain. Developmental doma in classification agreement = ( n pairs with agreement )/ N X100 (3 5) When examining developmental domain differences within each age cohort, N represented the total number of pairs within a specific age cohort. Congruence on Devel opmental Domain Raw Scores Many researchers recommend the use of more than one index to examine score reliability (Portney & Watkins, 2000). In addition to calculating percent agreement on items and

PAGE 96

96 classification decisions by developmental domain and age cohort, an intraclass correlation domains. ICC is calculated using variance estimates obtained through analysis of variance, thus it reflects both degree of cor respondence and agreement among ratings. The ICC ranges from 0.00 to 1.00, with higher numbers indicating that two groups of informants rate similarly. For the purposes of this study, parents were considered one group of informants and ECEs are the compari son group of informants. Several models are available to calculate ICCs and the selection of model is based on the nature of the study. Equation 3 6 was used to calculate ICC because each child is assessed by a set of different informants (i.e., unique set s of parent ECE pairs) and informants were considered as if they were chosen randomly (Portney & Watkins, 2000). This model is based on a one way analysis of variance in which children are treated as the independent variable. Total variance is composed of differences between children and error variance, which represents the variation within a child across informants. Some of the error component could be due to true scores changing across informants, some could be due to informant error, and some is unexplai ned. The ANOVA does not discriminate between these sources of error. In Equation 3 6, BMS is the between subjects mean square, EMS is the error mean square, RMS is the between informants mean square, and k is the numbers of informants. ICC = BMS EMS (3 6) BMS + ( k 1)EMS Item Response Pattern of Parents and ECEs For each ASQ item, informants could select one of the three provided response choices or omit the item. Across the ASQ items, the total number of times each possible respo nse choice (e.g., yes, sometimes, not yet, or omitted ) was selected was totaled for each informant type (i.e., parent and ECE). Research questions for which item response patterns were calculated are shown

PAGE 97

97 in Table 3 2. To calculate the overall item respon se pattern for each informant type, Equation 3 7 was applied using n = 126. Percentage of items with response choice = (3 7) (Total number of items response choice was selected )/( n 30) X 100 The divisor in the formula represented the total num ber of possible items across all participants, totaling 3780 items. To examine overall item response patterns by age cohort, Equation 3 7 was adjusted so that n represented the number of participants in each informant type for each age cohort. To calculat e item response patterns within each developmental domain, the total number of times each possible response choice (yes, sometimes, not yet, or omitted) was selected across the 6 items was totaled for each informant type (i.e., parent and ECE) using the fo rmula shown in Equation 3 8. Percentage of developmental domain items with response choice = (Total number of items response choice was selected )/ ( n 6) X 100 (3 8) The divisor in the formula represented the total number of possible items in each developmental domain across all participants, totaling 756 items. Calculations were repeated for each developmental domain. This equation was also applied to examine developmental domain differences for each age cohort. Equation 3 8 was adjusted so that n represented the number of participants within each age cohort. Predictors of Congruence To address research question five, beta regression was used. Regression analysis is used when testing whether an independent variable (IV) predicts a dependen t variable (DV; Younger, 1979). Ordinary least squares (OLS) regression is commonly applied to test the linear relationship between two variables. One of the primary assumptions in OLS regression is that the

PAGE 98

98 conditional distribution of the DV is normal. Th is assumption is violated when the DV is a proportion (Smithson & Verkuilen, 2006), as is the case with percentages. When the DV is a proportion, the function relating the DV to the IV is not likely to be linear. In cases such as this, it is more appropria te to use beta regression in which the predicted proportions are between 0 and 1 (Simas, Barreto Souza, & Roucha, 2010; Smithson & Verkuilen, 2006). Beta regression assumes that the residuals follow a beta distribution that is bounded by 0 and 1, which can be symmetric or skewed. Beta regression consists of two sub models. The first sub model tests the conditional mean of the DV while the second sub model tests the precision parameter (Smithson & Verkuilen, 2006). Testing the conditional mean of the DV perm its examination of whether the IV scores (e.g., child age and length of relationship) predict the proportions of the DV (e.g., item level percent agreement). The test for the precision parameter asks whether the IVs predict the variance of the conditional beta distributions. Including the second sub model allows for the fact that if the DV is related to the IV, the variance will also be related. As opposed to other analyses used in research questions two and three, child age was defined by number of months to provide a comparative unit of analysis to length of relationship with the ECE, also defined in months.

PAGE 99

99 Table 3 1. Comparison of ASQ 3 standardization sample distribution with the present sample distribution by age cohort. Age cohort ASQ 3 standardiza tion s ample n (%) S ample n (%) 2 12 months 7,158 (39%) 17 (14%) 14 24 months 5,743 (31%) 27 (21%) 27 36 months 3,064 (17%) 38 (30%) 42 60 months 2,607 (14%) 44 (35%) Total N 18,572 126 Note. T here are 21 ASQ 3 questionnaire intervals tha t are used for children from 1 66 months (Squires Bricker, Twombly, & Potter, 2009). Each questionnaire indicates the age range applicable for use Age cohorts were formed by referencing the grouping of age intervals used in validity studies of the ASQ 3.

PAGE 100

100 Table 3 2. Number of children who were eligible for inclusion in the present study, whose parents provided consent, and who were included in analyses. Eligible Children N Consent Received n (% of eligible) Included in Analyses a n (% of elig ible) 0 12 months 32 22 (69%) 17 (53%) 14 24 months 57 31 (54%) 27 (47%) 27 36 months 66 40 (61%) 38 (58%) 42 60 months 107 52 (49%) 44 (41%) Total across ages 262 145 (55%) 126 (48%) Note. a To be in cluded in analyse s, the parent and ECE must have completed the ASQ 3 within 2 weeks of each other Five parent ECE ASQ 3 pairs returned ASQ forms outside of the 2 week window and were not included in analyses. Fourteen parent ECE pairs had incomplete forms, meaning that consent had been received fr om the ECE and parent, an ASQ 3 packet had been given, but either one or both of the packets were never returned.

PAGE 101

101 Table 3 3. Summary of research aims by ASQ unit of analysis, variable of interest, and analysis procedure. ASQ Unit Variable of Interest Analysis Research Aim Items CD Raw Score Child Age DD LR % agree ICC BR % of RS 1a 1b 2a 2b 3a 3b 3c 3d 3e 4a 4b 4c 4d 5 Note. CD = classification decisions; DD = developmental domain; LR = length of relationship between child and ECE; % agree = percent agreement between parent and ECE; ICC = intraclass correlation coefficient; BR = beta regression; % of RS = percent of response selection.

PAGE 102

102 Figure 3 1. The number of completed questionnaires for each ASQ 3 age interval. There are 21 ASQ 3 questionnaire i ntervals that are used for children from 1 66 months (Squires, Bricker, Twombly, & Potter, 2009). Each questionnaire indicates the age range applicable for use.

PAGE 103

103 CHAPTER 4 RESULTS The purpose s of the present study were to (a) examine overall congruence in the judgment based ratings of two informants (i.e., parent and early childhood educator) when they each completed the same multi dimensional developmental screening instrument for a child (Ages and Stages Questionnaire 3 [ASQ 3]; Squires & Bricker, 2 009), (b) identify whether congruence between the two informants varied based on child age cohort, (c) examine if congruence between the two informants var ied by developmental domain included on the screening instrument (e.g., communication, fine motor, gr oss motor, problem solving, and personal social), ( d ) compare patterns of responses between the two types of informant s, and (e) explore whether child age and length of relationship with the early childhood educator predict ed congruence To conduct the stu dy, parents and early childhood educators ( ECEs ) from two university based early education and care center s were recruited to complete separately a developmental screening instrument for the same child. The multi dimensional developmental screening instrum ent used was the A SQ 3 Of the 262 eligible children enrolled in the centers, signed informed consent forms were received for 145 children (54.1%). Questionnaires were included for analyses if parent ECE pairs completed the forms within 2 weeks of each oth er. Five parent ECE pairs returned ASQ 3 forms outside of the 2 week window and were not included in analyses. Fourteen parent ECE pairs had incomplete forms, meaning consent had been received from the ECE and parent, an ASQ 3 packet had been given, but ei ther one or both of the packets were not returned. ASQ 3 data for 126 parent ECE pairs across the two centers w ere used for data analyses.

PAGE 104

104 Overall Congruence B etween Parent and ECE Ratings Descriptive analyses were conducted to address the first researc h questio n related to congruence of parent and ECE judgment based ratings on the ASQ 3. One analysis involved examining overall item level agreement and the other overall classificatio n agreement. Overall Item Level Agreement Overall item level agreement b etween parents and ECEs was calculated across all 30 ASQ items to generate a percentage agreement score. Item level agreement was calculated in two ways: (1) by using the three original response choices ( yes, sometimes, not yet ) provided in the ASQ 3 (Squi res & Bricker, 2009) and (2) by using two response choices generated by considering yes and sometimes as agreement. When parents ECE pairs selected the same response for each ASQ item, a score of 1 was applied and when they selected responses that differed from each other a score of 0 was applied. If one informant omitted the item, this was also scored as 0. Scores were summed and then divided by 30 and multiplied by 100 to obtain a percent agreement score for each informant pair. When examining congruence using the three response choices, overall percent agreement averaged across 126 parent ECE pairs was 78% ( SD = 14 %) When calculating congruence using two response choices (i.e., yes/sometimes, not yet ), overall percent agreement averaged across 126 paren t ECE pairs was 92% ( SD = 9 %) Overall Classification Agreement On the ASQ 3, a total domain score can be derived based on informant responses to each item (Squires et al., 2009). If an ASQ 3 item is marked yes a score of 10 is given, if the item is marke d sometimes a score of 5 is given, and if the item is marked not yet a score of 0 is given. Scores for the six items in each domain are summed and can range from 0 to 60. Different total domain scores correspond to one of three classification categories ( a bove the cutoff, monitoring

PAGE 105

105 zone, below the cutoff ) using age based cutoffs provided on the summary forms. The classification categories are used to provide guidance about whether or not the child should receive further evaluation or monitoring. In the pr esent study, congruence about o verall classification decision s re ferred to the average agreement between parents and ECEs for the classification categories across the five developmental domains included on the ASQ 3 Classification decision congruence was calculated for each developmental domain by applying a score of 1 when parents and ECEs agreed on the classification category and applying a score of 0 when they did not agree on the classification category. These scores were summed and then divided by fiv e and multiplied by 100 to obtain an averag e classification percent agreement score across all ASQ domains When calculating classification agreement using three categories (i.e., above the cutoff, monitoring zone, and below the cutoff) average percent ag reement for the 126 parent ECE pairs was 86%, SD = 21 % When calculating agreement using two categories (i. e., typical, eligible ) created by considering above the cutoff and monitoring zone as agreement average percent agreement for the 126 parent ECE pai rs was 93%, SD = 1 3% Parent and ECE Congruence by Child Age Cohort Two types of descriptive analyses were conducted to address the second research question related to whether ASQ 3 score congruence between parents and ECEs var ied based on child age cohor t : overall item level agreement by age cohort and overall classification agreement by age cohort. Similar to the analyses described for research question one, overall item level agreement and overall classification agreement w ere examined separately. Varia tion in item level and classification agreement based on child age was examined using four age cohorts, which were consistent with those used in validity and reliability studies reported in the ASQ 3

PAGE 106

106 manual (Squires et al., 2009) : (a) birth 12 months, (b ) 14 24 months, (c) 27 36 months, and (d) 42 60 months. Overall Item Level Agreement by Age Cohort The distribution of the child sample in the present study across the four age cohorts is shown in Table 4 1. When examining congruence of parent ECE ASQ 3 r atings across age cohorts using three item level response choices (i.e., yes, sometimes, not yet ), overall item level agreement for each age cohort was (a) M = 83%, SD = 11% for children 42 60 months of age ( n = 44); (b) M = 82%, SD = 13% for children 27 3 6 months of age ( n = 38); (c) M = 71%, SD = 12% for children 14 24 months of age ( n = 27); and (d) M = 69%, SD = 14% for children 0 12 months of age ( n = 17) When examining congruence using two item level response choices (i.e., yes/sometimes; not yet ) o verall item level percent agreement for each age cohort was (a) M = 96%, SD = 6% for children 42 60 months of age; (b) M = 93%, SD = 9% for children 27 36 months of age; (c) M = 88%, SD = 8 % for children 14 24 months of age; and (d) M = 86%, SD = 10% for c hildren 0 to 12 months of age (Table 4 1). Overall Classification Agreement by Age Cohort To examine congruence related to c lassification agreement for each of the four age cohorts, the analytic procedures for overall classification agreement described ab ove were applied to each of the four age cohorts When classification decision congruence was examined using the three classification categories (i.e., above the cutoff, monitoring zone, and below the cutoff ) the classification agreement percentages by ag e cohort were (a) M = 91%, SD = 15% for children 42 60 months of age; (b) M = 91%, SD = 16% for children 27 36 months of age; (c) M = 79%, SD = 23% for children 14 24 months of age; (d) M = 69%, SD = 27% for children 0 12 months of age When congruence wa s calculated based on two classification categories (i.e., typical, eligible ) classification agreement percentages by age cohort were (a) M = 96%, SD = 10% for

PAGE 107

107 children 42 60 months of age; (b) M = 94%, SD = 12% for children 27 36 months of age; (c) M = 90%, SD = 17 % for children 14 24 months of age; (d) M = 88%, SD = 16% for children 0 12 months of age. Table 4 2 shows data for classification agreement by age cohort. Parent and ECE Congruence for Five Developmental Domains Five descriptive analyses we re conducted to address research question three related to whether congruence across parent ECE judgment based ratings on the ASQ 3 varied across developmental domains: (a) item level agreement for each developmental domain, (b) classification agreement fo r each developmental domain, (c) item level agreement by developmental domain within age cohorts, (d) classification agreement by developmental domain within age cohorts, and (e) intraclass correlation coefficients (ICCs) by developmental domain. Item Lev el Agreement by Developmental Domain For each of the five developmental domains on the ASQ 3, congruence between parent ECE ratings was evaluated by calculating a percent agreement score for the six items in each domain. The analytic methods used for overa ll item level agreement were applied but the divisor was 6 rather than 30 (Equation 3 3). Agreement was calculated for three response choices (i.e., yes, sometimes, not yet ) and two response choices (i.e., yes/sometimes, not yet ). For the three response ch oices, average percent agreement for the 126 parent ECE pairs for each domain were : (a) gross motor M = 86% ( SD = 16%), (b) problem solving M = 79% ( SD = 22%), (c) communication M = 78% ( SD = 22%), (d) fine motor M = 75% ( SD = 23%), and (e) personal so cial M = 73% ( SD = 24%). For two response choices, average percent agreement for the 126 parent ECE pairs for each domain were : (a) gross motor M = 96% ( SD = 9%), (b) problem solving M = 92% ( SD = 13%), (c) fine motor M = 91% ( SD = 15%), (d) personal s ocial M = 91% ( SD = 14%), and (e) communication M = 90 % ( SD = 16%). Considering the response

PAGE 108

108 choices of yes and sometimes as agreement resulted in the smallest increase in percent agreement across informant pairs for the gross motor domain when comparing with increases made in other developmental domains (Table 4 1). Classification Agreement by Developmental Domain For e ach developmental domain score on the ASQ 3, age based cutoffs provided on the summary forms (Squires & Bricker, 2009) are used to deter mine the equivalent classification category (i.e., above the cutoff, monitoring zone, below the cutoff) At the domain level, percent agreement for parent ECEs classification decision was either 0% or 100% for each parent ECE pair To examine classificatio n congruence by domain agreement was averaged across 126 parent ECE pairs (Equation 3 5). When three classification categories were used percent agreement for classification decision by domain was (a) problem solving M = 88% ( SD = 33%) ; (b) fine motor M = 88% ( SD = 33%) ; (c) gross motor M = 87% ( SD = 33%) ; (d) personal social M = 83% ( SD = 3 8 %) ; and (e) communication, M = 83% ( SD = 38%). When two classification categories were used percent agreement for each domain was (a) problem solving M = 97% ( S D = 18%) ; (b) fine motor M = 95% ( SD = 21%) ; (c) personal social M = 94% ( SD = 23%) ; (d) communication M = 90% ( SD = 30%) ; and (e) gross motor M = 89% ( SD = 32%) Data related to classification congruence by domain are shown in Table 4 2. Item Level Ag reement by Developmental Domain and Child Age Cohort For the youngest age cohort ( 0 12 months ) there were 17 parent ECE pairs. For this age cohort item level percent agreement (three response choices) by developmental domain was (a) gross motor M = 8 1% ( SD = 13%) ; (b) fine motor M = 73% ( SD = 25%) ; (c) problem solving M = 71% ( SD = 21%) ; (d) communication M = 62% ( SD = 20%) ; and (e) personal social M = 58% ( SD = 28%). When calculating item level percent agreement for two response choices percent agreement increased across all developmental domains, but the order of the last two

PAGE 109

109 developmental domains changed, with communication having the lowest percent agreement. Complete results for item level agreement by developmental domain and child age coho rt are shown in Table 4 1. The age cohort of 14 24 months had 27 parent ECE pairs. For this age cohort item level agreement (three response choices) by developmental domain was (a) gross motor M = 87% ( SD = 18%) ; (b) fine motor, M = 75% ( SD = 22%) ; (c ) communication M = 70% ( SD = 20%), (d) problem solving M = 61% ( SD = 25%), and (e) personal social M = 61% ( SD = 25%). When calculating item level agreement for two response choices, percent agreement increased across all developmental domains, but the third (communication) and fourth (problem solving) developmental domains changed order from a e above (Table 4 1). T he age cohort of 27 36 months had 38 parent ECE pairs. For this age cohort item level agreement (three response choices) by development al domain was (a) communication M = 90% ( SD = 16%) ; (b) gross motor M = 85%, ( SD = 18%) ; (c) problem solving M = 80% ( SD = 18%) ; (d) personal social M = 80% ( SD = 18%) ; and (e) fine motor M = 74% ( SD = 25%). When calculating item level agreement for tw o response choices, percent agreement increased across all developmental domains and the order of developmental domains remained the same as a e above (Table 4 1). The age cohort of 42 60 months had 44 parent ECE pairs. For this age cohort, item level agreement (three response choices) by developmental domain was (a) problem solvin g, M = 92 % ( SD = 1 2%) ; (b) gross motor M = 86% ( SD = 1 5 %) ; (c) personal social M = 81 % ( SD = 2 1%); (d) communication M = 7 9 % ( SD = 2 3 %) and (e) fine motor M = 7 8 % ( SD = 2 3%) When calculating item level agreement for two response choices, percent agreement increased across

PAGE 110

110 all developmental domains and the order of developmental domains remained the same as a e above (Table 4 1). Classification Agreement by Developmental D omain and Child Age Cohort For the 0 12 month age cohort, there were 17 pairs. For this age cohort, classification agreement (three categories) by developmental domain was (a) fine motor M = 82% ( SD = 39%) ; (b) communication M = 76% ( SD = 44%) ; (c) pr oblem solving M = 76% ( SD = 44%) ; (d) gross motor M = 65% ( SD = 49%) ; (e) and personal social M = 47% ( SD = 51%). When calculating classification agreement for two categories (i.e., typical, eligible ), percent agreement by developmental domain was (a) p roblem solving, M = 100% ( SD = 0%); (b) personal social, M = 94% ( SD = 24%); (c) fine motor, M = 88% ( SD = 33%); (d) communication, M = 82% ( SD = 39%); and (e) gross motor, M = 76% ( SD = 44%). Classification agreement for two categories was higher across a ll developmental domains, but the order of the first (problem solving) and third (fine motor) developmental domains changed from a e above. For the age cohort of 14 24 months, classification agreement (three categories) by developmental domain was (a) g ross motor M = 89% ( SD = 32%) ; (b) fine motor M = 89% ( SD = 32%) ; (c) problem solving M = 74% ( SD = 45%) ; (d) personal social M = 74% ( SD = 45%) ; and (e) communication M = 70% ( SD = 47%). When calculating classification agreement for two categories, p ercent agreement by development domain was (a) fine motor, M = 93% ( SD = 27%); (b) problem solving, M = 89% ( SD = 32%); (c) communication, M = 89% ( SD = 32%) ; (d) personal social, M = 96% ( SD = 19%); and (e) gross motor, M = 81% ( SD = 40%) Classification agreement for two categories was higher across all developmental domains, but the order of the developmental domains changed when considering highest to lowest classification agreement percentages.

PAGE 111

111 For t he 27 36 month age cohort, classification agreement (three categories) by developmental domain s was (a) communication M = 95% ( SD = 23%) ; (b) problem solving M = 95% ( SD = 23%) ; (c) personal social M = 92% ( SD = 27%) ; (d) gross motor M = 87%, ( SD = 34% ); and (e) fine motor M = 87% ( SD = 34%) When ca lculating classification agreement for two categories, percent agreement by developmental domain was (a) fine motor, M = 100% ( SD = 0%); (b) problem solving, M = 97% ( SD = 16%); (c) communication, M = 95% ( SD = 23%); (d) personal social, M = 92% ( SD = 27%) ; and (e) gross motor, M = 87% ( SD = 34%). Classification agreement for two categories was higher across all developmental domains, but the order of the developmental domains changed when considering highest to lowest classification agreement percentages. For the 42 60 month age cohort classification agreement (three categories) by developmental domain was (a) gross motor M = 95% ( SD = 21%) ; (b) problem solving M = 95% ( SD = 21%) ; (c) personal social M = 93% ( SD = 26%) ; (d) fine motor M = 91% ( SD = 29%) ; and (e) communication M = 82% ( SD = 39%). When calculating classification agreement for two categories, percent agreement by developmental domain was (a) gross motor M = 100% ( SD = 0%) ; (b) problem solving M = 100% ( SD = 0%) ; (c) personal social M = 95% ( SD = 21%), (d) fine motor M = 95% ( SD = 21%) ; and (e) communication M = 91% ( SD = 29%). Classification agreement for two categories was higher across all developmental domains and the order remained the same as a e above. Results for classifica tion agreement by developmental domain and age cohort using three and two classification categories are shown in Table 4 2. Relationship s between Parent and ECE Raw Scores on Developmental Domains As noted previously, raw scores can be calculated for each ASQ 3 developmental domain by summing the item level scores. A yes response to an item receives a score of 10, a sometimes response receives a score of 5, and a not yet response receives a score of 0. Scores on each

PAGE 112

112 domain, therefore, range from 0 to 60 g iven there are six items associated with each of the five ASQ 3 domains. An intraclass correlation coefficient (ICC) was calculated for each developmental domain using the raw scores for parents and ECEs. A one way ICC model for single measurement was sele cted for this analysis (Equation 3 6) Several models are available to calculate ICCs and the selection of model is based on the nature of the study. A one way ICC model for single measurement was selected because each child is assessed by a set of differe nt informants (i.e., unique sets of parent ECE pairs) and informants were considered as if they were chosen randomly (Portney & Watkins, 2000). ICCs range from 0 to 1.00. ICCs were as follows: (a) gross motor domain = 73, F (1, 125) = 6.06, p <.001 ; (b) c ommunication domain =. 64, F (1, 125) = 5.32, p <.001 ; (c) problem solving domain ICC = .56, F (1, 125) = 3.73, p <.001 ; (d) personal social domai n ICC = .54, F (1, 125) = 3,34, p <.001 ; and (e) fine motor domain ICC = .46, F (1, 125) = 2.71, p <.001. These res ults differ from those reported for overall item level percent agreement by domain because the ICC accounts for chance agreement between parents and ECEs (Portney & Watkins, 2000) and also is based on correspondence between domain scores rather than on ite m level responses. Comparison of Frequencie s and Percentages of Response Selection between Types of Informant For each of the 30 item s on the ASQ 3, there are three available response choices: yes, sometimes, and not yet. T he informant may only omit 1 to 2 item s in each domain. In these cases, the ASQ 3 provides adjusted scoring guidelines for each domain If more than 2 items are omitted, the developmental domain cannot be scored ( Squires et al., 2009 ). To address research question four related to frequen cy of response selections for each type of informant (i.e., parent or ECE), the number of times one of the four possible responses (i.e., yes, sometimes, not yet, and omitted ) was selected was totaled. Frequencie s were calculated by totaling the number of

PAGE 113

113 times that a response choice was selected. Using these frequencies, percent of response selection w as calculated across all of the 30 items, by age cohort, by developmental domain, and by developmental domain with in age cohorts using Equations 3 7 or 3 8 a s described in Chapter 3 In the following sections, percentages of response selection are reported to facilitate comparisons across age cohorts and the reader may refer to Table 4 3 for frequencies of response selection. Frequency and Percentage of Respo nse Selection for All ASQ Items The frequency and percentage of i tem response selection was calculated for all 30 ASQ 3 items (Table 4 3 and Table 4 4) When comparing the frequency and percentages of response s selected across parents and ECEs, parents se lected sometimes more often than ECEs (+2%) and omitted 20 more items (+.5%). ECEs selected yes more often than parents (+2%). Both parents and ECEs selected not yet 7% of the time ( Table 4 4 ). Frequency and Percentage of Response Selection by Age Cohort The frequency and percentage of responses selected by each informant type within each age cohort was calculated using an adjusted version of Equation 3 7 by totaling the number of ASQ 3 items and multiplying by the number of participants in the age cohort (30 x n children in each age cohort ). As shown in Tables 4 3 and 4 4, ECEs selected not yet more often than parents for children 0 12 months (+2%) 14 24 months (+1%) and 42 60 months (+1%) and parents selected not yet more often than ECEs for child ren 27 36 months (+2%). ECEs selected sometimes more often than parents for children 4 2 60 months (+1%) and parents selected sometimes more often than ECEs for children 0 12 months (+3%), 14 24 months (+4%), and 27 36 months (+3%). ECEs selected yes more often than parents for three age cohorts : children 0 12 months (+1%), 14 24 months (+3%), and 27 36 months (+5%). Both parents and ECEs selected yes for 88% of the 30 response opportunities (i.e., 30 ASQ 3 items) for children 42 60 months.

PAGE 114

114 Parents were more likely than ECEs to omit items across all four age cohorts For children ages 0 12, parents omitted 2 items and ECEs did not omit any. In the 14 24 month age cohort parents omitted 4 items and ECEs did not omit any. For children ag es 27 36 months, parents omitted 8 items and ECEs omitted 1 item, while for the 42 60 month age cohort parents omitted 10 items and ECEs omitted 3 items. The patterns of parent and ECE omissions are shown in Table 4 3 Frequency and Percentage of Res ponse Selection by Development al Domain Frequencies and p ercentages associated with response selection patterns were calculated based on the number of items in each domain multiplied by the number of parent ECE pairs (6 x 126 = 756) using Equation 3 8 As shown in Tables 4 3 and 4 4, ECEs selected not yet more often than parents in the communication (+4%) and gross motor (+1%) domains. Parents selected not yet more often than ECEs in the fine motor domain ( + 3%). Both parents and ECEs selected not yet for 8% of response opportunities in the problem solving domain. Parents selected sometimes more often than ECEs in four domains: communication (+2%), gross motor (+2%), problem solving (+3%), and personal social (+3%). In the fine motor domain, parents and ECEs both selected sometimes for 12% of the response opportunities In terms of selecting yes, parents chose this response more often than ECEs in the communication (+3%) domain. ECEs selected yes more often than parents in four domains: gross motor (+1%), fine motor (+5%), problem solving (+4%), and personal social (+3%). Overall, across developmental domains, parents omitted 24 items while ECEs omitted 4 items (Table 4 3) Parents and ECEs each omitted one item in the communication domain. ECEs omitted one it em in fine motor and two items in personal social. There were no similarities in the patterns of omissions by ECE informants. Comparatively, parents omitted more items in gross motor (+2 items), fine motor (+12 items), problem solving (+6 items), and perso nal social (+3

PAGE 115

115 items) domains The most commonly omitted item among parents w as from the fine motor domain : whether the child was stringing small items (4 times) and able to us e child safe scissors (2 times). Frequency and Percentage of Response Selection by Developmental Domain across Age C ohorts When examining the frequency of responses and percentages of response selection categories for the 0 12 month age cohort across all d omains, ECE s selected not yet and yes more frequently than parents and parent s selected sometimes more frequently than ECEs (Tables 4 3 and 4 4) When considering each domain, ECEs selected not yet more often than parents in the communication (+12%) problem solving (+2%) and personal social domain (+5%) ECEs selected yes more o ften than parents in the gross motor (+5%) and fine motor (+9%) domain. The overall pattern of responses changed in the following ways when examined by developmental domain: parents selected not yet more often than ECEs in gross motor (+3%) and fine motor (+4%) and selected yes more often than ECEs in communication (+2%) and personal social (4%) Overall percentage r esults and percentage results by developmental domain within this age cohort are shown in Table 4 4 When examining the pattern of responses f or the 14 24 month age cohort across all domains parents selected sometimes more frequently than ECEs and ECEs selected not yet and yes more frequently than parents (Table 4 3) When examining response pattern percentages by developmental domain, t he over all pattern changed in the following ways: (a) ECEs selected sometimes more often than parents in fine motor (+1%); (b) parents selected not yet more often than ECEs in fine motor (+1%) and problem solving (+4%); (c) parents selected yes more often than EC Es in gross motor (+1%) and communication (+5%); and (d) parents and ECEs selected not yet at equal percentages i n personal social (Table 4 4 ).

PAGE 116

116 When examining frequencies and percentages of responses for the 27 36 month cohort across all domains, p arents selected not yet and sometimes more often than ECEs and ECEs selected yes more often than parents (Tables 4 3 and 4 4) When examining response patterns by developmental domain, this pattern was consistent for fine motor and personal social domains. The ov erall pattern changed in the following ways for the other developmental domains : ECEs selected not yet more often in gross motor (+1%) and problem solving (+4%) and parents and ECEs selected sometimes at equal percentages in the communic ation domain (Table 4 4 ). When examining frequencies and percentages of response patterns across all developmental domains for the 42 60 month age cohort, ECEs selected not yet more often than parents and parents selected sometimes more often than ECEs. P arents and ECEs se lected yes equally. As shown in Tables 4 3 and 4 4, ECEs selected not yet more often than parents in communication (+3%) and gross motor (+1%) ECEs selected sometimes more often in communication (+1%) fi ne motor (+1%), and problem solving (+2%), and pare nts and ECEs selected yes at equal percentages in the gross motor domain P arents selected not yet more often than ECEs in fine motor (+1%) and personal social (+1%) parents selected sometimes more often than ECEs in personal social (+3%) and parents sele cted yes more often than ECEs in communication (+4%) and problem solving (+2%) P arents and ECEs selected sometimes at equal percentages in the gross motor domain Child Age and Length of ECE Child Relationship as Predictors of Congruence The fifth resear ch question focused on whether child age and length of time the early childhood educator had known the child when the ASQ 3 was completed were noteworthy predictors of parent ECE item agreement. To conduct exploratory analyses related to this question, bet a regression was use d and consist ed of two sub models. Model 1 tested whether the conditional mean of the dependent variable (i.e., overall item agreement between informants

PAGE 117

117 expressed as a proportion) increase d as the categories of the independent variable (i.e., age and length of time ECE knew child when ASQ 3 completed) increase d Model 2 tested the precision parameter: whether the dependent variable predicted the variance of conditional beta distributions. Child age and length of ECE child relationship (independent variables, IVs) were measured in months and congruence (dependent variable, DV) was a proportion between 0 and 1.0. When length of ECE child relationship was not included in the model child age significantly predicted the conditional mean of parent ECE congruence b = .016, t (124) = 3.94, p < .01, and was a non significant predictor of the conditional variance of the beta distribution, b = .007, t (124) = 0.96, p = .337. When child age was not included in the model length of ECE child relation ship significantly predicted the conditional mean of parent ECE congruence b = .028, t (124) = 3.50, p < .01, and significantly predicted the conditional variance of the beta distribution, b = .060, t (124) = 3.57, p < .01. When child age and length of EC E child relationship were tested as predictors of parent ECE congruence child age ( b = .013, t (124) = 3.00, p < .01, ) and length of child ECE relationship ( b = .018, t (124) = 3.03, p = .040) significantly predicted the conditional mean of parent ECE cong ruence Estimating the precision parameter model with both variables showed that length of ECE child relationship was a significant predictor ( b = .049, t (124) = 3.02, p < .01) and child age was a non significant predictor of the conditional variance of th e beta distribution (Table 4 5 ). Summary The purpose s of the present study were to examine overall congruence in the judgment based ratings of two informants (i.e., parent and ECE) when they each completed the same multi dimensional developmental screening instrument for a child (ASQ 3; Squires & Bricker, 2009) and whether there were variations in congruence due to developmental domain, child age cohort, or child age and length of relationship with the ECE. Overall item level agreement between

PAGE 118

118 parents and E CEs was generally high as was classification agreement. Variability was noted in item and classification agreement across child age cohorts. Overall item level percent agreement was higher for parents and ECEs of preschool aged children relative to infant and toddler age cohorts. Findings showed differences in item agreement across developmental domains and agreement was consistently lower for fine motor skills. Examination of response patterns showed some differences across parent and ECE informants, with parents selecting sometimes more often than ECEs and ECEs selecting yes and not yet more often than parents. Findings showed that the overall response pattern was generally consistent across developmental domains and age cohorts, but some changes in respon se pattern by developmental domain were noted. Child age and length of child ECE relationship significantly predicted the conditional mean of parent ECE congruence, expressed as a proportion of item agreement, but only length of child ECE relationship was a significant predictor of the conditional variance of the beta distribution.

PAGE 119

119 Table 4 1. M ( SD ) of item level percent agreement by age cohort and developmental domain. Developmental Domain 0 12 month age cohort ( n = 17) 14 24 month age cohort ( n = 2 7) 27 36 month age cohort ( n = 38) 42 60 month age cohort ( n = 44) Total across age cohorts ( n = 126) 3 Response Choices Communication 62 (20) 70 (20) 90 (16) 79 (23) 78 (22) Gross Motor 81 (13) 87 (18) 85 (18) 86 (15) 86 (16) Fine Mo tor 73 (25) 75 (22) 74 (25) 78 (23) 75 (23) Problem Solving 71 (21) 61 (25) 80 (18) 92 (12) 79 (22) Personal Social 58 (28) 61 (25) 80 (18) 81 (21) 73 (24) Total across Domains 69 (14) 71 (12) 82 (13) 83 (11) 78 (14) 2 Respon se Choices Communication 78 (18) 83 (17) 96 (11) 93 (15) 90 (16) Gross Motor 91 (10) 96 (10) 95 (11) 97 (6) 96 (9) Fine Motor 89 (16) 90 (14) 89 (17) 93 (13) 91 (15) Problem Solving 87 (13) 85 (17) 93 (12) 98 (6) 92 (13) Personal Socia l 83 (20) 86 (14) 90 (14) 97 (7) 91 (14) Total across Domains 86 (10) 88 (8) 93 (9) 96 (6) 92 (9)

PAGE 120

120 Table 4 2. M ( SD ) classification agreement by age cohort and developmental domain. Developmental Domain 0 12 month age cohort ( n = 17) 14 24 mo nth age cohort ( n = 27) 27 36 month age cohort ( n = 38) 42 60 month age cohort ( n = 44) Total across age cohorts ( n = 126) 3 Classification Categories Communication 76 (44) 70 (47) 95 (23) 82 (39) 83 (38) Gross Motor 65 (49) 89 (32) 87 ( 34) 95 (21) 87 (33) Fine Motor 82 (39) 89 (32) 87 (34) 91 (29) 88 (33) Problem Solving 76 (44) 74 (45) 95 (23) 95 (21) 88 (33) Personal Social 47 (51) 74 (45) 92 (27) 93 (26) 83 (38) Total across Domains 69 (27) 79 (23) 91 (16) 91 (15) 86 (21) 2 Classification Categories Communication 82 (39) 89 (32) 95 (23) 91 (29) 90 (30) Gross Motor 76 (44) 81 (40) 87 (34) 100 (0) 89 (32) Fine Motor 88 (33) 93 (27) 100 (00) 95 (21) 95 (21) Problem Solving 100 (0) 89 (32) 97 (16) 100 (0) 97 (18) Personal Social 94 (24) 96 (19) 92 (27) 95 (21) 94 (23) Total across Domains 88 (16) 90 (17) 94 (12) 96 (10) 93 (13)

PAGE 121

121 Table 4 3. Frequency of response selection by parents (P) and early childhood educators (ECE) on the Ages and Stages Questionnaire 3. 0 12 months ( n = 17) 14 24 months ( n = 27) 27 36 months ( n = 38) 42 60 months ( n = 44) Total across Ages ( n = 126) Response Selected P ECE P ECE P ECE P ECE P ECE Communication Domain Not Yet 6 18 27 41 16 14 5 13 54 86 Sometimes 18 9 20 14 11 11 26 28 75 62 Yes 78 75 115 107 200 203 233 222 626 607 Omit ted 1 1 1 1 Gross Motor Domain Not Yet 14 11 7 12 14 15 3 6 38 44 Sometimes 10 8 14 10 18 8 18 19 60 45 Yes 78 83 141 140 196 205 241 239 656 667 Omitted 2 2 0 Fine Motor Domain Not Yet 7 3 10 8 27 14 10 7 54 32 Sometimes 14 10 19 21 28 22 33 38 94 91 Yes 80 89 130 133 170 191 216 219 596 632 Omitted 1 3 3 1 5 12 1 Problem Solving Domain Not Yet 6 8 18 11 19 20 4 6 47 45 Sometimes 14 13 34 16 25 15 11 15 84 59 Yes 81 81 109 135 181 193 248 243 619 652 Omitted 1 1 3 1 6 0 Personal Social Domain Not Yet 17 22 17 18 21 18 4 1 59 59 Sometimes 21 20 31 29 32 24 33 27 117 100 Yes 64 60 114 115 174 186 225 234 577 595 Omitted 1 2 2 3 2 Total across domains Not Yet 50 62 79 90 97 81 26 33 252 266 Sometimes 77 60 118 90 114 80 121 127 430 357 Yes 381 388 609 630 921 978 1163 1157 3074 3153 Omitted 2 4 8 1 10 3 24 4

PAGE 122

122 Table 4 4. Percentage of response selection by parents (P) and early childhood educators (ECE) on the Ages and Stages Questionnaire 3. Note. Percentages were calculated based on the number of items in each domain multiplied by the number of participants in each age group. 0 12 months ( n = 17) 14 24 months ( n = 27) 27 36 months ( n = 38) 42 60 months ( n = 44) Total across Ages ( n = 126) Response Selected P ECE P ECE P ECE P ECE P ECE Communication Domain Not Yet 6 18 17 25 7 6 2 5 7 11 Sometimes 18 9 12 9 5 5 10 11 10 8 Yes 7 6 74 71 66 88 89 88 84 83 80 Omitted <1 <1 <1 <1 Gross M otor Domain Not Yet 14 11 4 7 6 7 1 2 5 6 Sometimes 10 8 9 6 8 4 7 7 8 6 Yes 7 6 81 87 86 86 90 91 91 87 88 Omitted 1 <1 Fine Motor Domain Not Yet 7 3 6 5 12 6 4 3 7 4 Sometimes 14 10 12 13 12 10 13 14 12 12 Yes 78 87 80 82 75 84 82 83 79 84 Omitted 1 2 1 <1 2 2 <1 Problem Solving Domain Not Yet 6 8 11 7 8 9 2 2 6 6 Sometimes 14 13 21 10 11 7 4 6 11 8 Yes 79 79 67 83 79 85 94 92 82 86 Omitted 1 1 1 <1 1 Personal Social Domain Not Yet 17 22 11 11 9 8 2 <1 8 8 Sometimes 21 20 19 18 14 11 13 10 15 13 Yes 63 59 70 71 76 82 85 89 76 79 Omitted <1 <1 <1 Total across domains Not Yet 10 12 1 0 11 9 7 2 3 7 7 Sometimes 15 12 15 11 10 7 9 10 11 9 Yes 75 76 75 78 81 86 88 88 81 83 Omitted <1 <1 1 <1 1 <1 1 < 1

PAGE 123

123 Table 4 5. Predictors of parent ECE item agreement. Model 1 Conditional Mean Model 2 Precisio n Parameter Variable b 95% CI b 95% CI Child Age .016** [.008, .024] .007 [ .007, .021] Length of ECE Child Relationship .028** [.012, .044] .060** [.027, .093] Combined Model Child Age .013** [.005, .022] Length of ECE Child Relation ship .018* [.001, .035] .049* [.017, .082] Note. p < .05 ** p < .01

PAGE 124

124 CHAPTER 5 DISCUSSION The primary aim of the present study was to evaluate congruence between ratings of two informants who spend significant amounts of time with young children (i.e ., parent and ECE s ) when they each complete d the same multi dimensional deve lopmental screening instrument, the Ages and Stage Questionnaire 3 (ASQ 3; Squires & Bricker, 2009). A secondary aim was to examine whether c ongruence varied based on the developme ntal domain assessed, child age (by cohort or in months), or length of relationship between children and ECEs. To explore these aims, descriptive and exploratory analyses were used. This chapter interprets findings of the present study and discusses them in the context of related literature. Limitations of the present study are described and implications for future research and practice are presented. Overall Congruence B etween Parent s and ECE s Using the ASQ 3 Overall Item Level Congruence Overall item lev el agreement was 78% for three response choices ( yes, sometimes, and not yet ) and 92% when calculated for two response choices ( yes/sometimes, not yet ). The latter agreement percentages are similar to those of Frankenburg, van Doornick, Liddell, and Dick ( 1976), which was the only research study located that examined congruence using item level agreement between parents and examiners and focused on children of similar age to those in the present study. Item agreement, based on two response choices ( yes, no ) between the PDQ (Frankenburg, van Doornick, & Liddell, 1976) and the DDST (Frankenburg et al., 1975) was reported to be 93%. Unlike in the present study where both informants completed the ASQ 3 in the same way, in the Frankenburg et al. study, parents a nd examiners completed two different developmental screening instruments that used different administration methods (i.e., parent

PAGE 125

125 questionnaire versus direct elicitation). The PDQ was created, however, through the selection of essential items from the DDS T (Frankenburg, van Doornick, Liddell, & Dick, 1976). Item agreement in the Frankenberg et al. study was calculated by comparing the responses on the PDQ only to the corresponding item on the DDST. In the present study, when using three versus two response choices to evaluate congruence, the majority of disagreements between parents and ECEs occurred between the response categories of yes and sometimes Comparing the increase in overall item level percent agreement when the response categories of yes and so metimes were considered an agreement revealed this response pattern The increase from 78% to 92% agreement indicates that 14% of the disagreements between parent and ECE informants were between yes and sometimes Only 8% of disagreements between these inf ormant pairs were between response choices of yes/sometimes versus not yet This finding shows almost twice as many disagreements occur between yes and sometimes as between yes/sometimes and not yet One possible explanation for the finding that agreemen t is higher when the response categories of yes and sometimes relation to a particular developmental skill reflected in an ASQ 3 item might influence whether a parent or ECE scores yes versus someti mes Four phases of learning have been described in the literature: acquisition, fluency, generalization, and adaptation (e.g., Haring, Lovitt, Eaton, & Hansen, 1978). In the acquisition phase, children are learning a new skill and are not yet able to use the skill consistently or with accuracy. During the fluency phase, children are accurate in their responses but need repeated practice to improve the speed of responding. When children are acquiring or becoming fluent with a skill, these skills might be ch aracterized as developing skills and, on the ASQ 3 might be scored as either not yet or sometimes occurring (Squires et al.,

PAGE 126

126 2009). Fluency sets the occasion for the generalization phase of learning (Haring et al., 1978). In this phase, children are accura te and fluent in using the skill and are beginning to use the skill consistently across contexts or people. Developmental skills in this phase might be indicated on the ASQ 3 as disagreements between yes and sometimes (Squires et al., 2009), because a skil l is emerging and beginning to generalize across contexts or people. Developmental skills in the last phase, adaptation, reflect a level of mastery where the child is learning how to adjust the skill for novel tasks or situations (Haring et al., 1978). Dev elopmental skills in this phase may be indicated on the ASQ 3 by responses of yes by both informants (Squires et al., 2009), because the child has mastered the skill across contexts or people and so both informants should share the same perspective. Given the possible relationship of the phases of learning to ASQ 3 (Squires & Bricker, 2009) response choices, the finding that the majority of disagreements between parents and ECEs occur between yes and sometimes suggest that differences in perspectives might be primarily occurring for emerging developmental skills. Although disagreements about child development may often instigate concern, it is important to obtain multiple perspectives to inform judgments or decisions about child development and learning ac ross contexts and people, rather than ensuring exact agreement about skill performance across informants (Suen et al., 1992). Pursuing exact agreement could negate the rich perspectives of each informant. Thus, the knowledge that parents and ECEs typically skills can inform the basis for important conversations between parents and ECEs about contexts.

PAGE 127

127 Using an ins trument such as the ASQ 3 completed by parents and ECEs to inform discussions about similarities or differences in skills demands or expectations across at least two contexts where many young children spend significant time (i.e., home and early care and e ducation program) might lead to the formation of a multiply linked mesosystem and support beneficial for these types of conversations between parents and ECEs to be shared discussions in information about similarities as well as differences in developmental skill demands and expectations across contexts, reaching consensus wh en appropriate. For example, a parent might score on the ASQ 3 that a child sometimes drinks from a cup unassisted, while an ECE might score that a child drinks from a cup unassisted (i.e., response choice of yes ). Discussions between the parent and ECE r eveal that at home the child is not always given an opportunity to drink from a cup unassisted given other parental and child demands or expectations at mealtimes. In the classroom, however, children are always provided a cup during mealtimes and they are expected to drink independently from the cup. In this particular example, the parent and ECE could discuss whether or not it is appropriate to align expectations for independent cup drinking across contexts but there is no need to be concerned about the ch independently. Overall Classification Congruence The ASQ 3 is primarily used to determine which children might need monitoring or further evaluation. This purpose focuses on classification decisions. In the present study, overall classification agreement between parents and ECEs was 86% for three classification categories ( above the cutoff, monitoring zone, below the cutoff ) and was 93% when two classification

PAGE 128

128 categories were used ( eligible, typical ). The terms eligible and typical were used to facilitate comparisons to validity studies reported in the ASQ 3 (Squires et al., 2009), and correspond to recommendations for whether or not to refer a child for further monitoring or evaluation. Eligible indicates that recommended that the child should be referred for further evaluation. Typical indicates that a item a greement congruence (described above) was lower when using three response choices (78%), the disagreements that occurred between parents and ECEs when rating discrete developmental skills did not result in clinically significant differences in whether or n ot to refer a child, as shown by 93% agreement between parents and ECEs in identifying children as eligible versus typical Using three versus two categories to evaluate classification congruence, comparisons showed that parent ECE disagreements occur at similar percentages for children broadly demonstrating lower developmental skills (i.e., disagreements between eligible and typical ) and higher levels of ability on developmental skills (i.e., disagreements between above the cutoff and monitoring zone ) Co mparing the increase in percent agreement when the classification categories of above the cutoff and monitoring zone were considered an agreement revealed this classification pattern The increase from 86% to 93% agreement indicates that 7% of the disagree ments were between above the cutoff and monitoring zone reflecting disagreements on children who broadly demonstrate higher developmental abilities or a greater number of emerging or mastered skills (Haring et al., 1978). The remaining 7% of disagreements were between eligible and typical which might reflect children who may broadly demonstrate fewer

PAGE 129

129 developmental skills in one context versus another or a greater number of developing or emerging skills that are not seen consistently across contexts. In t he validity studies reported in the ASQ (Squires et al., 2009), two classification categories ( typical, eligible ) were used to examine classification agreement between parents and examiners. The authors reported 93% classification agreement when administering the same instrument to both informants. The classification agreement percentage was the same in the present study when two classification categories were used. In the congruence literature reviewed in Chapter 2, classification agreement between two informants ranged from 59% (Frankenburg, van Doornick, Liddell, & Dick, 1976) to 97% (Bricker et al., 1988). When considering only studies that used informant procedures similar to those in the present study and that were conducted with the ASQ classification agreement ranged from 87% 97% (Bricker & Squires, 1989b; Bricker et al., 1988; Squires et al., 1997). Despite the fact that these ASQ studies were based on a comparison between parents and examiners who did not have extended contact with the child, classification agreement percentages in these studies were similar to those reported in the present study. Other than differences in the sample of professionals and the length of time they had known the child, there were not differences between the present study and prior ASQ studies with respect to the procedures used to gather data (i.e., an ASQ completed by two different informants within 2 weeks of each other). Similarities in overall classification decision agreement when two categories ( ty pical, eligible ) were used in previous studies and in the present study suggest that this type of agreement might not be significantly affected by the length of time an ASQ informant has known the child. Two studies included in the Chapter 2 literature re view examined overall classification agreement between parent and teacher: studies on the PDQ (Frankenburg, van Doornick, &

PAGE 130

130 Liddell, 1976) and the R PDQ (Frankenburg, 1987). Both studies used two classification categories ( fail and pass) to indicate whethe r or not a child should receive further evaluation or support. Parent teacher classification agreement on the PDQ was reported at 59% (Frankenburg, van Doornick, Liddell, & Dick, 1976) while parent teacher classification agreement on the R PDQ was reported as 83% (Frankenburg et al., 1987). These results reflect lower classification agreement than was found in the present study for three and two classification categories (86% and 93%, respectively). The teachers sampled in both of the Frankenburg studies we re from various day care centers, Head Start programs, and kindergartens in the community and data describing the characteristics of the teachers, including their experience, were not available (Frankenburg et al., 1987; Frankenburg, van Doornick, Liddell, & Dick, 1976). It was therefore not possible to compare characteristics of ECEs in the present study with characteristics of teachers in the Frankenburg studies in relation to differences in overall classification agreement percentages. Exploring Variati ons in Congruence of Parent ECE Ratings on the ASQ 3 A secondary aim of the present study was to explore if parent ECE congruence varied by developmental domain, child age (by cohort and in months), or length of child ECE relationship. For congruence by developmental domain and child age cohort, the descriptive analyses used were item level agreement, classification agreement, and examination of response patterns across informant types. Exploratory analyses included intraclass correlation coefficients to examine congruence by developmental domain raw scores and beta regression to examine predictors of congruence.

PAGE 131

131 Item Level Congruence by Developmental Domain Congruence for each ASQ 3 developmental domain was examined by determining agreement or disagreeme nt at the item level between parents and ECEs for the six items within each developmental domain, calculating percent agreement for each informant pair, and averaging percentages across the 126 parent ECE pairs. Findings from the present study showed congr uence varied across ASQ 3 developmental domains. Examining percentage agreement for the six items associated with each of the five ASQ 3 developmental domains, parents and ECEs had the highest agreement on the gross motor domain (86%), followed by the prob lem solving (79%), communication (78%), fine motor (75%), and personal social (73%) domains, respectively. Of the score congruence studies reviewed in Chapter 2, only one study examined item level agreement by developmental domain (Coghlan et al., 2003). I n the Coghlan et al. study, parents and day care providers completed the PEDS (Glascoe, 2007b), which had 10 items that each represented a different domain of development. There were three possible response choices ( no, a little, and yes ) for each PEDS ite m. Coghlan et al. (2003) reported gross motor (95%) and problem solving (cognitive, 92%) had the highest item level agreement compared to other developmental domains. Comparisons between PEDS items and other ASQ 3 domains are less straightforward because t he PEDS reports concerns for expressive and receptive language separately whereas these are represented on the ASQ 3 as one developmental domain (e.g., communication). The percentages for item agreement reported in Coghlan et al. (2003) ranged from 78% to 94% across the 10 items or domains, which is slightly higher than the range reported in the present study across the five developmental domains represented on the ASQ 3 (73% 86%). The differences might be attributed to having only 1 2 items comprise a de velopmental domain on the PEDS (Glascoe, 2007b) compared to six items per developmental

PAGE 132

132 domain on the ASQ 3 (Squires & Bricker, 2009). Nevertheless, despite differences in the number of items and types of items per domain across the PEDS and ASQ 3, in both the present study and the Coghlan et al. study, agreement between parents and care providers/ECEs was highest for gross motor and problem solving (cognition) relative to other developmental domains. In the present study, average overall item agreement wa s higher across all developmental domains when the response choices of yes and sometimes were considered an agreement. Comparing item agreement for three versus two response choices, the largest increases occurred for personal social (+18%) and fine motor domains (+16%), which indicate the percent of disagreements on emerging skills ( yes versus sometimes ) Parents and ECEs disagreed on developing skills ( yes/sometimes versus not yet ) 9% of the time for both personal social and fine motor domains. For these two developmental domains, parents and ECEs disagreed on emerging skills almost twice as often as on developing skills As noted previously, emerging skills are likely to result in disagreements on the ASQ 3 item responses between sometimes and yes since c hildren have acquired a skill but might not be consistently using the skill across contexts or people (Haring et al., 1978). Classification Congruence by Developmental Domain Congruence, as measured by classification agreement percentages, also varied acr oss ASQ 3 developmental domains, Using three classification categories, agreement was 88% for fine motor and problem solving, 87% for gross motor, 83% for communication, and 83% for personal social. When the classification categories of above the cutoff an d monitoring zone were considered as agreement, classification agreement increased by 7% 9% for all developmental domains except gross motor. For other developmental domains, calculating classification agreement using two categories increased percentages to 97% for problem solving, 95% for fine

PAGE 133

133 motor, 94% for personal social, and 90% for communication. Classification agreement for gross motor increased from 87% to 89% Although average item level agreement percentages were lower for fine motor and perso nal social domains compared to other ASQ 3 developmental domains, the lower item level agreement percentages were not associated with lower classification agreement percentages for these same two domains (i.e., 95% classification agreement for fine motor a nd 94% for personal social). The results reported above were compared to the only score congruence study on classification agreement by developmental domain. Chiu and DiMarco (2010) compared the responses of mothers on the ASQ 2 (Bricker & Squires, 1999) t Denver II (Frankenburg et al., 1992) and calculated classification agreement using the categories of pass and fail Cl assification agreement was reported for four of the five ASQ 2 developmental domains given the Denver II did not have a comparable domain for problem solving. Classification agreement results reported by Chiu and DiMarco were 95% for personal social and gross motor, 71% for fine motor and 67% for language (communication) The classification findings report ed by Chiu and DiMarco for personal social were the same as those reported for the personal social domain in the present study, while results for gross motor were only slightly higher than the present study (95% compared to 89%). Two findings reported by C hiu and DiMarco (2010) were different from those reported in the present study. First, a cross the four developmental domains compared in the Chiu and DiMarco study, language (communication) had the lowest classification agreement at 67%, which is substanti ally lower than reported in the present study (90%). There are several possible explanations for this difference. The authors hypothesized that low classification agreement was observed for

PAGE 134

134 terances even when adults who do not spend significant time with the child cannot. Higher classification agreement for communication might have been obtained in the present study given the study sample included ECEs who observe and interact with the childr en daily, sometimes up to 40 hours per week, compared to a nurse who is only interacting with the child at a single point in time. Second, Chiu and DiMarco reported classification agreement of 71% for fine motor, which is lower than reported in the present study ( evaluation of fine motor concerns more often than the nurse. D ivergent results from the present study may also be due to measurement differences between the ASQ 2 (Bricker & Squir es, 1999) and the Denver II (Frankenburg et al., 1992 ) given that the ASQ 2 is completed as a parent questionnaire while the Denver II is completed by direct elicitation. Moreover, although the developmental domains represented across the two instruments a re similar, the items that comprise each of the developmental domains are different, which likely precluded Chiu and DiMarco from examining item level agreement as was done in the present study. Item Level Congruence by Age Cohort Congruence, as defined b y overall item level agreement, may vary based on child age cohort. Interpretations offered in this section should be considered preliminary given the small numbers of infants and toddlers in the present study. Average percent agreement across ASQ 3 domain s for the two preschool age cohorts was 82% and 83%, respectively and for the infant and toddler age cohorts was 69% and 71%, respectively. For each of the age cohorts, comparisons were made for overall item level agreement percentages using three versus t wo response choices to evaluate congruence. When the response categories of yes and sometimes were considered as

PAGE 135

135 agreement item agreement increased by 17% for the infant and toddler age cohorts and increased by 11% and 13% for the preschool age cohorts. T hus, item agreement for two response choices was as follows: 86% for infants, 88% for toddlers, and 93% and 96% for the preschool age cohorts. Considering the large increase in percentages when yes and sometimes are considered as agreement, results show th at parents and ECEs are not in agreement when rating emerging skills that are children are beginning to generalize across contexts (Haring et al., 1978). The large increase in item agreement when using two response choices is consistent across all age coh orts, but is more predominant for the infant and toddler age cohorts. None of the studies reviewed in Chapter 2 focused on congruence investigated differences in parent professional item level congruence by age cohorts. Within each of the four age cohort s, congruence varied by developmental domain, with greater variability across developmental domains for infant and toddler age cohorts than preschool age cohorts. Average item agreement between parents and ECEs in the infant age cohort varied by 23% across ASQ 3 developmental domains (range = 58% 81%) and by 26% for the toddler age cohort (range = 61% 87%). For each of the preschool age cohorts, average item agreement varied by 16% for the 27 36 month age cohort (range = 74% 90%) and by 9% for the 4 2 60 month age cohort (range = 78% to 92%). The finding that there is greater variability across developmental domains for infants and toddlers is not unexpected when considering the progression of child development. growth consists of multiple complex processes and in order to effectively monitor the development of these processes, they are often classified into developmental domains (Bagnato, 2007). Although the use of categories implies discrete divisions between developmental domains, skill s can involve processes across multiple domains. For example, the ability to

PAGE 136

136 a problem solving skill related to exploration of objects (Squires & Bricker, 2009), but also samples the fine motor /pe rceptual motor skill s of grasping objects and hand eye coordination. During the infant years, skill development appears undifferentiated and globalized due to the immaturity of the nervous system, sensory abilities, and motor functions (Bagnato, 2007). As children grow, the progression of skill development reflects an increasing refinement and specialization of behaviors, and thus, developmental skills are more discrete and might be easier to identify. Compared to preschool age cohorts, infants and toddlers may have lower overall item congruen ce and greater variability across developmental domains because developmental domains are highly interrelated during infant and toddler years and therefore it may be more difficult to identify discrete skills. Additio nally, for some developmental skills, there are precise differences that mark the progression of skill development that may be difficult for informants to identify or agree on the frequency of its occurrence. For example, on the ASQ 3, items associated wit h several age intervals ask informants to observe the way a child picks up a small piece of food (i.e., Cheerio), however, earlier age intervals ask about whether a raking motion with all fingers is being used while later age intervals ask about whether th e tips of the thumb and fingers are being used (Squires & Bricker, 2009). A s skills become more discrete or differentiated as children age congruence between scores used by parents and ECEs might be higher than at earlier ages Within the infant and todd ler age cohorts, item level agreement was higher for the ASQ 3 gross motor domain (81% infant age cohort and 87% toddler age cohort), compared to other ASQ 3 developmental domains (see Table 4 1). This finding might be related to the types of skills reflec ted on ASQ 3 items in the gross motor domain for infants and toddlers. The skills are

PAGE 137

137 discrete (e.g., sits up straight for several minutes) and might be easier to rate when compared to skills in other developmental domains (e.g., copying activities that yo u do in the personal social domain). Additionally, differences across developmental domains and age cohorts may be due to the sequence and timing of development, with certain developmental domains assuming prominence during particular ages. Thus, parents a nd ECEs might be focused on the development of certain skills at particular ages leading to similar perspectives. For example, between approximately 6 15 months of age, children are beginning to crawl, sit, pull themselves up, stand, and walk (Bagnato, 20 07), which are developmentally relevant as well as observable skills reflected on the ASQ 3. For infant and toddler age cohorts, higher item level agreement for the gross motor domain relative to other domains might be explained by parents and ECEs height ened sensitivity to the developmental milestones included in the gross motor items on the ASQ 3. Item agreement between p arents and ECEs for infant and toddler age cohorts in the personal social skills domain was 58% and 61%, respectively One possible exp lanation for this relatively low level of agreement across informants might be the content of the items in the personal social domain in the ASQ 3, as well as the number of children cared for in an infant or toddler classroom versus a home environment. The personal social domain includes items related to behaviors that occur before and during feedings, diaper changes, and dressing (Squires et al., 2009). Within a classroom, as many as 8 infants between the ages of birth 15 months could be enrolled. NAEYC (2012) recommends a staff to child ratio for a group this size to be 1:4 and this ratio changes to 1:3 if there are 6 or fewer children in the classroom. Even when accounting for the required number of adults in the classroom, this is a large number of sam e aged children to care for, particularly given the caretaking responsibilities and level of supervision needed for

PAGE 138

138 infant and toddler age cohorts. Within the home, it is possible that there is only one child in the home, and even if there are siblings, a child may be the only one in his or her age range. Thus, it is possible that there are differing amounts of support and attention available to each child depending on the context. ECEs may be less focused on individual child behaviors related to personal s ocial skills as they are simultaneously conducting feedings, diaper changes, and dressing with other children. Within the younger preschool cohort (27 36 months), item agreement was higher for the communication domain (90%) relative to the other ASQ 3 d evelopmental domains (Table 4 1). Additionally, item agreement for communication was also higher than reported for the other three age cohorts of infants, toddlers, and older preschoolers (62%, 70%, and 79%, respectively). Relatively high congruence for th e communication domain may be due to parents and ECEs heightened sensitivity to these developmental skills between the ages of 27 36 months because children at this age are becoming increasingly verbal, including the rapid acquisition of new vocabulary a nd increasing length of phrases and sentences (Bagnato, 2007). For this age cohort, one possible explanation for higher item level agreement for the communication domain relative to other domains might be parents and ECEs heightened sensitivity to developm ental milestones reflected in the communication domain, leading to more similar perspectives. It may also be that the content of the communication domain for this age cohort reflects discrete skills that are easier to observe and to arrive at similar persp ectives. F or older preschoolers (42 60 months), item agreement between parents and ECEs for the ASQ 3 problem solving domain was 92% compared to agreement for the infant, toddler, and younger preschool cohort (71%, 61%, and 80%, respectively). This find ing might be explained by the heightened emphasis on pre academic skills for this age cohort with items on the ASQ 3

PAGE 139

139 focused on pre academic concepts such as numbers, letters, shapes, and sizes (Squires & Bricker, 2009). Across all age cohorts, item agree ment percentages were relatively consistent for the gross motor domain (range = 81% 87%) and the fine motor domain (range = 73% 78%) For the preschool age cohorts, fine motor was the developmental domain with the lowest item level agreement. This may be a ttributed to differing demands and expectations across contexts for fine motor skills The early education and care centers involved in the present study frequently planned activities in which to observe and support fine motor skill development (i.e., draw ing, writing, stringing beads), particularly at the preschool ages when these skills are considered important for school readiness. There were several parents who omitted the ASQ 3 item asking about stringing beads (Squires & Bricker, 2009), which is not a n activity that often occurs naturally in the home. Another parent omitted an item asking about how well a child cuts on a line (Squires & Bricker, 2009) and wrote in an explanation that they did not have child safe scissors in the home. When considering t he number of opportunities children have at early education and care centers to learn and refine fine motor skills reflected on the ASQ 3, the finding of slightly lower agreement percentages between parents and ECEs compared to other domains was not unexpe cted. Classification Congruence by Age Cohort Congruence, as measured by overall classification agreement, varied based on child age cohort, with higher percent agreement for preschool age cohorts than infant and toddler age cohorts. Preschool age cohorts each had 91% classification agreement between parents and ECEs, while agreement between these two sets of informants was 69% for infants and 79% for toddlers. Comparisons were made for classification agreement for three versus two classification

PAGE 140

140 categorie s. Across all four age cohorts, calculating classification agreement using two categories (e.g., eligible and typical) resulted in congruence ranging from 88% 96%. When comparing classification agreement of three versus two categories, the largest increa se in percent agreement occurred in the infant cohort (69% to 88%), indicating that 19% of the disagreements occurred between above the cutoff and monitoring zone with the remaining 11% of disagreements between eligible and typical Referencing the phases of learning, this may indicate that parents and ECEs are primarily disagreeing on children who broadly demonstrate higher developmental abilities or have more emerging skills (Haring et al., 1978). As reported in the previous section, parents and ECEs of the infant age cohort had the lowest overall item level agreement at 69%. While this result might warrant some concern, results of classification agreement show that low overall item level agreement generally did not result in clinically significant diffe rences in whether or not to refer a child, as shown by 88% classification agreement when using two classification categories. For the toddler and preschool age cohorts, using three versus two classification categories to evaluate congruence shows that pare nt ECE disagreements occur at similar percentages for children broadly demonstrating lower developmental skills (i.e., disagreements between eligible and typical) and higher levels of ability on developmental skills (i.e., disagreements between above the c utoff and monitoring zone ) As described in previous sections, this was determined by comparing the increase in percent agreement when above the cutoff and monitoring zone were considered agreement. None of the studies reviewed in Chapter 2 focused on cong ruence investigated differences in parent professional classification congruence by age cohorts. Within each age cohort, there was greater variability in classification agreement across developmental domains for infant and toddler age cohorts (35% and 19 %, respectively) than

PAGE 141

141 preschool age cohorts (8% and 13%, respectively). As described earlier, skill development appears undifferentiated and globalized during the infant years due to the immaturity of the nervous system, sensory abilities, and motor functi ons (Bagnato, 2007). As children grow, the progression of skill development reflects an increasing refinement and specialization of behaviors, and thus, developmental skills are more discrete and may be easier for informants to identify. When developmental skills are discrete and easily observable, it is likely that perspectives across informants will be similar. As noted previously, however, these preliminary infant and toddler cohort findings should be interpreted with caution given the small number of in fants and toddlers in the present study. For each age cohort, variability in classification agreement percentages across developmental domains decreased when three versus two classification categories were used. Additionally, using two classification categ ories resulted in s everal instances in which parents and ECEs had perfect agreement on classification : problem solving for the infant ( 0 12 mont h) and older preschool ( 42 60 month ) cohorts, fine motor for the young preschool ( 27 36 month ) cohort, and gross motor for the older preschool cohort. Classification agreement for two categories was generally high across all age cohorts and developmental domains with one exception. Parents and ECEs of children 0 12 months disagreed on classification decisio ns (e.g., eligible versus typical ) more often in the gross motor domain (76% classification agreement) compared to other developmental domains. Using three versus two classification categories to evaluate congruence, comparisons showed that parent ECE disa greements occur red more often for children broadly demonstrating lower developmental skills than for children demonstrating higher levels of ability on developmental skills Comparing increases in percent agreement when the classification categories of abo ve the cutoff and monitoring zone were

PAGE 142

142 considered an agreement revealed this pattern The increase from 65% to 76% indicates that 9% of the disagreements were between above the cutoff and monitoring zone reflecting disagreements on children who broadly de monstrate higher developmental abilities or a greater number of emerging or mastered skills (Haring et al., 1978). The remaining 24% of disagreements were between eligible and typical reflecting children who may broadly demonstrate lower developmental abi lity or a greater number of developing or emerging skills. Relationship s between Parent and ECE Raw Scores on Developmental Domains The summary scores of parents and ECEs were compared for each developmental d omain. On the ASQ 3, summary scores for each d evelopmental domain are derived by adding scores for ratings on each of the six items, so that total scores range from 0 to 60 (Squires et al., 2009). Intraclass correlation coefficients (ICCs) were used to compare summary score ratings of each parent ECE pair. ICCs ranged from .46 (fine motor) to .73 (gross motor) and F tests showed statistically significant all five developmental domains ( p < .001). As two groups of informants, the ICCs suggest fair to moderate r elationships between raw scores although the mean scores differed across type of informant. The ICC findings in the present study are comparable to the range of ICCs reported in the ASQ (Squires et al., 2009). The ASQ 3 authors compared par ent and examiner ratings on the ASQ 3 for 107 children, ages unspecified. Examiners completed the ASQ 3 after administering the BDI 2 (Newborg et al., 2004). Squires et al. reported ICCs ranging from .43 (communication) to .69 (personal social), but did no t specify the ICCs for gross motor, fine motor, and problem solving. Although the range of ICCs is similar to the range reported in the present study, the developmental domains associated with the upper and lower limits differ: fine motor had the lowest IC C in the present study compared to communication in Squires et al., while gross motor had the highest ICC in the present study compared to personal social in

PAGE 143

143 Squires et al. As described in previous sections, fine motor may have a lower congruence between p arents and ECEs due to differing demands across contexts, with more opportunities to observe and support fine motor skills occurring in early education and care centers. Additionally, the higher ICCs for communication domain compared to the other ASQ 3 dom ains might have been obtained in the present study given the study sample included ECEs who observe and interact with the children daily, compared to an examiner who is only interacting with the child at a single point in time. Examination of Informant Re sponse Patterns Overall average item level congruence between parents and ECEs across all age cohorts and ASQ 3 domains was 78%. Early research in score congruence typically interpreted disagreements between parents and professionals as being attributed t instruments (Johnson, Poteat, & Kushnick, 1986; Knobloch, Stevens, Malone, Ellison, & Risemberg, 1979; Miller, Manhal, & Mee, 1991) In the present study, comparisons of response patterns across informant types were conducted to provide additional information about the correspondence between parent and ECE perspectives. The frequency and percentage of the selection of a response choice ( y es, sometimes, not yet, and omitted ) was calculated for each informant type. Results of these analyses showed differences in the pattern of response selection across parents and ECEs Compared to parents, ECEs selected yes and not yet more often, while par ents selected sometimes more often when compared to ECEs. response patterns should correspond to parents frequently indicating response choices suggesting mastery of developmenta l skills (i.e., yes or sometimes ) with ECEs selecting response choices suggesting developing or emerging skills (i.e., sometimes or not yet ). Instead, the present study

PAGE 144

144 showed that ECEs selected response choices at the extremes while parents frequently sel ected the response choice ( sometimes) that reflects emerging skills. Parents might be more likely to select sometimes interactions across multiple contexts or people while ECEs selec t a response choice (either yes or not yet ) that is representative of only one context. Alternatively, parents might select sometimes more often because they are unsure if there child demonstrates the skill across contexts and people or they might not unde rstand completely the intent of the ASQ 3 item in relation to what differentiates sometimes from yes Regardless of which interpretation is considered, it is important to note that these findings represent global response patterns across the 126 informant pairs, not the differences in response patterns across informants for individual children. Item level agreement between parents and ECEs for the fine motor domain ranged from 73% to 78% across the four age cohorts. Comparisons of item level agreement for three versus two response choices showed that when agreement did not occur between parents and ECEs, it was primarily related to emerging fine motor skills. This finding is further supported by examination of response patterns for this developmental domain which differed from the overall response pattern reported above. Consistent with the item level agreement findings, response patterns showed ECEs selected yes more often than parents for fine motor items while parents selected sometimes more often than E CEs for the fine motor items. In addition, for fine motor domain items, response patterns showed that parents selected not yet more often than ECEs and omitted responses most often in the fine motor domain relative to the other four domains (i.e., 12 of th e 22 omissions occurring in the fine motor domain). Additionally, differences in response patterns across informant types for the fine motor domain were slightly larger than for other developmental domains, which had differences across informant types rang ing from 0% 4%

PAGE 145

145 (Table 4 4). In the fine motor domain, ECEs selected yes for 84% of responses compared to 79% of responses for parents, which is a difference of 5%. This fine motor response pattern was due to differences between parents and ECEs at the 0 12 month and 27 36 month age cohort (+9% each, respectively). Thus, parents were consistently rating fine motor skills using response choices that reflected developing skills (Haring et al., 1978), as shown by selecting not yet and sometimes more often than ECEs. Child Age and Length of ECE Child Relationship as Predictors of Congruence As described in Chapter 4, using beta regression, c hild age in months and length of relationship with the ECE were statistically significant predictors of the proportio nal means. O nly length of child ECE relationship was a statistically significant predictor of the proportional variance. The magnitude of the relationship between the predictors and the dependent variable s however, was small 2 (Younger 1979) provide d additional information about the magnitude of the relationship between the predictors (child age and length of ECE child relationship) and the dependent variable (proportion of overall parent ECE agreement) and was .09. The finding that th e magnitude of the relationship is relatively small may be due, in part, to the study inclusion criteria. ECEs had to know the child for at least 1 month to be enrolled in the present study. Had the sample included professionals who completed an ASQ 3 at o ne point in time and who did not know the child for at least 1 month, the magnitude of the relationship between this predictor variable and the dependent variables might have differed. None of the studies reviewed in Chapter 2 that focused on congruence in vestigated predictors of parent professional congruence.

PAGE 146

146 Limitations of the Present Study Several limitations of the present study should be considered in relation to reported results and interpretations. First, the results reported represent a single p oint in time completion of the ASQ 3 by two informants to examine congruence. It is possible that congruence between two informants might change over time. In a study where the ASQ was implemented with 36 at risk mothers, serial completions of the ASQ were associated with higher correlations between congruence might have been different had data collection occurred at a different time of year, which would have changed the length of the child ECE relationship, or if serial administration of the ASQ 3 had occurred throughout the year. Second, the informants did not have the same training and experiences related to the ASQ 3 before completing the instrument for ch ildren in the present study. For the majority of parents (67.5%) in the present study, it was the first time they completed the ASQ 3 for their child, while all of the ECEs in the study had training related to the ASQ 3 and experienced completing the ASQ 3 before completing it for a child in the present study. Congruence between two informants who have extended time with a child might differ when sampling informants who have similar experiences with a multi dimensional screening instrument. Third, the info rmants sampled in the present study are not necessarily representative of the general population of parents and ECEs involved in early education and care programs. The early education and care centers were based at a university, and therefore, were only av ailable to faculty, staff, and students affiliated with the university. All parents in the present study had some college experience, which is a contrast to parents represented in other score congruence studies reviewed in Chapter 2. When specified, approx imately 40% or less of the parent sample had at least some college experience ( Bricker & Squires, 1989; Bricker, et al. 1988; Burgess, et

PAGE 147

147 al. 1984; Chiu & DiMarco, 2010; Coghlan, et al. 2003; Frankenburg, et al. 1987; Frank en burg, von Doornick, Liddell & Dick, 1976; Glascoe, 2002; Squires & Bricker, 1991; Squires, et al. 1997) in the areas of early childhood/early childhood special education or child development The s ample of ECEs included in this study generally have more training than the general population of ECEs considering m any states have a minimum requirement of a high school diploma to work in a child care program ( Bureau of Labor Statistics, 2013; Chil It is estimated that approximately 20% of teachers in child care centers have a high school diploma or less (Child Care Aware of America, 2012). ECEs in the present study were employed by NAEYC accredited centers, while only 10 % of child care centers in the United States are nationally accredited (Child Care Aware of America, 2012). R esults from the present study might not generalize to the general population of informants or centers given the characteristics of the centers invo lved in the present study and the providers and parents affiliated w ith them. Fourth, while the total number of children and informant pairs in the present study ( n = 126) exceeded the minimum number specified in the power analysis ( n = 120), there were a disproportionate numbers of children in the preschool age cohorts relative to the other age cohorts. The power analysis suggested 30 children were needed in the 0 12 month age cohort and in the 14 24 month age cohort. Although thorough recruiting me thods were used, only 17 children in the 0 12 month age cohort and 27 children in the 14 24 month age cohort had complete data to be included in the analyses. As a result, the proportion of questionnaires for each age cohort was inversely disproportionate to the normative sample for the ASQ 3 (Squires & Bricker, 2009) as shown in Table 3 1. Research aims examining congruence by age cohorts or age cohorts and developmental domains should be interpreted with caution and results should be

PAGE 148

148 considered preliminar y. Questions concerning variability in congruence by child age cohort or age in months should be further explored in future studies with larger samples of infants (0 12 months) and toddlers (14 24 months). Implications of Findings The present study contr ibutes to the developmental screening score congruence literature by using procedures and analyses that permitted the exploration of similarities and differences in perspectives about child development al skills between parents and early childhood educ ators. Findings from the present study suggest that parents and ECEs generally share similar dimensional developmental screening instrument in the same way. Although congruence varied to some extent across developmental domains, age cohorts, and by age cohorts and developmental domains, when two response categories rather than three were examined for both items and classification, average percent agreement approximated or exceeded 80%. ICCs indicated 3 domains. Taken together, these data suggest that parents and ECEs shared relatively similar perspectives about this sample of young children. When differences in persp ectives existed, they often were associated with one informant scoring sometimes and the other informant scoring yes which suggests that differences are primarily occurring when children have emerging developmental skills, meaning that they are beginning to generalize an acquired skill across contexts and people (Haring et al., 1978). Thus, disagreements between parents and ECEs are not necessarily an instead, co mparing the similarities and differences across informants can provide information as

PAGE 149

149 Haring et al., 1978) or about the importance of sharing and verifying pe rspectives in relation to contextual expectations and skills observed within and across contexts. Although there has been a misconception that a goal of research focused on congruence is to provide evidence for parental estimates of child ability being sim ilar to professionals (see reviews by Dinnebeil & Rule, 1994; Snyder et al., 1992), proponents for multi informant approaches to assessment, including screening, recognize the need to incorporate information from both practitioners and parents. Striving f or exact agreement in judgments or ratings would negate the benefits of including multiple perspectives in assessment (Suen et al., 1995). Parents and professionals bring different sets of observations and expertise to the assessment process and each persp ective is essential to making informed decisions as a team. The interest in investigating congruence was to understand the relationship between the perspectives of two informants who spend extended time with the child across different contexts. Understandi ng the similarities and differences in the perspectives of parents and ECEs might help inform strategies that facilitate the formation of a multiply linked mesosystem in which there is a compatibility of expectations and activities for children (Bronfenbre nner, 1979). Continuity across settings may contribute to the development of mutual goals, including mutual goals related to development, learning, and behavior, thus increasing the frequency of opportunities for a child to experience developmentally enhan cing learning opportunities across different contexts and people. Given the typically brief interactions between parents and practitioners in early care and education settings, it often is difficult for parents and ECEs to regularly exchange information 1992). One possibility for enhancing information exchange in early care and education settings is to take advantage of existing or anticipated practices and incorporate strategies that actively

PAGE 150

150 Developmental screening is increasingly becoming a common practice in early education and care programs (National Resourc e Council, 2008). The inclusion of parents in this process fulfills recommendations of conducting multi informant assessment (DEC, 2007; NAEYC 2003, 2012) while also presenting a potentially effective and efficient strategy for establishing shared understa nding s to experiences provided across multiple settings (Bronfenbrenner, 1974, 1979) A review of the historical background of developmental screening show ed that although parent involvem ent has been long regarded as a recommended practice in screening, it has not been consistently implemented (AOA, 1978; Paget & Nagle, 1976) Difficulties with incorporating parent input have been partially due to vague policies and guidelines for practiti oners as well as, until recently, the limited availability of instruments, such as the ASQ 3, that meet standards for educational and psychological measurement yet permit judgment based scoring by informants lls (Glascoe et al., 1990; Katoff & Reuter, 1980; Macy, 2012) In the past decade, prominent organizations in the medical, psychological, and educational fields have unanimously supported the gathering of multiple sources of evidence to inform assessment specifically, the inclusion of parent input in the decision making process (DEC, 2007; Glascoe, 1991; NASP, 2009; Neisworth & Bagnato, 2005) Using parent completed developmental screening instruments could serve dual benefits of improving the ecological validity of screening results and enhancing information exchange between parents and ECEs (Bagnato 2007)

PAGE 151

151 The results of the present study can also inform developmental screening practices in early education and care centers, especially given recent fede ral level recognition of the need for universal developmental screening of young children ( AAP, 2001; Council on Children with Disabilities et al., 2006; NASP, 2009; Obama for America, 2012). To highlight the potential utility of sharing perspectives acros s contexts using developmental screening instruments, a case At the time of data collection, Child A was 34 months of age and had known the two ECEs in her classroom for 10 months. Child A was a Caucasian female with a significant visual classroom was 83% for item agreement and 80% for classification agreement. These percentages are generally considered t o be high congruence between parents and ECEs. Examining item endorsed endorsed this item, in addition to five other items which her mother had indicated she was not yet demonstrating. The ECE indicated that Child A was able to push a toy on whee ls (e.g., wagon) and could sometimes point to seven body parts when asked, demonstrate receptive understanding not yet and sometimes may indicate that Child A is begin ning to develop these skills and her mother may not provide opportunities or activities in which to observe these developing skills or may need some support in recognizing these discrete skills or how to teach these skills to her young child with a visual impairment. Although her mother and ECE have similar perspectives about her overall development (i.e., agreed that she was not yet exhibiting skills reflected on 24 of 30

PAGE 152

152 ASQ 3 items), regular use of developmental screening instruments by both informants c guidance for shared conversations. Systematic developmental screening programs afford early education and care centers the opportunity to use program data as a need s assessment for professional development of staff on the observation, documentation, or support of developmental skills. This need became evident when comparing the ASQ 3 results from one infant classroom to the other infant classrooms involved in the pre sent study. Of the four infant classrooms in the present study, one of the infant classrooms had 6 of 7 children identified by ECEs as being below the cutoff on the ASQ 3 personal social domain and therefore eligible for further evaluation. Of these childr en, the parents of 2 children also identified their child as eligible for further evaluation, while the remaining 4 children were only identified by ECEs. The frequency of below the cutoff classifications was particularly high in this one classroom compare d to the other infant classrooms, signifying a possible need for professional development on the observation, documentation, or support of personal social skills for the ECEs assigned to that classroom. Future research might replicate the present study i n different early childhood settings or with other populations to explore generalizability of results, particularly, with a larger sample of infant toddler parent ECE pairs. A larger sample of parent ECE pairs might permit the application of other explorat ory analyses that would provide additional or different information about predictors of parent ECE congruence. Additional studies might be conducted in which multiple and periodic ASQ 3 forms are collected from parent ECE pairs to explore whether there are changes in congruence over time, considering the promising results of Squires and Bricker

PAGE 153

153 (1991) in which serial completions of the ASQ 3 was associated with higher correlations Based on findings of the present study, researchers might want to investigate further the ecological validity of several ASQ 3 (Squires & Bricker, 2009) items. According to the ASQ 3 i tems were developed for the ASQ after a review of standardized developmental tests availab le at the time the first version of the instrument (IMQ ; Bricker & Squires, 1989a ) was published, nonstandardized tests focused on early development, textbooks, and literature about early developmental milestones (Squires et al., 2009). ASQ items were deve loped based on the following criteria: (a) skills could be easily observed or elicited by parents, (b) skills were highly likely to occur within homes and child care settings, and (c) skills reflected important developmental milestones. In the present stud y, half of the omitted items occurred in the fine motor domain, which raises questions about the ecological validity of some fine motor items The noodles. With resp ect to another ASQ 3 fine motor item, a parent noted she did not rate an item safe scissors were not available in the home. These examples raise questions about the extent to which some materials associat ed with fine motor items on the ASQ findings suggest additional studies about the ecological validity of some ASQ 3 items might be warranted, particularly given the consequential validity of parents either omitting or scoring these items incorrectly (e.g., false positive or false negative screening decisions). Rewriting several fine motor items to reflect skills more likely to be observed in the home setting that sample similar developmenta l skills and conducting studies to validate these items might also be warranted.

PAGE 154

154 Results of the present study provides information that may be beneficial for enacting strategies related to supporting the formation of multiply linked mesosystems (Bronfenbre nner, 1979) by showing the potential of adjusting existing program practices to serve multiple goals of early identification, parent involvement in the assessment process, and most importantly, shared sion across the multiple contexts in which children learn and grow. While parents and ECEs typically had high congruence in rating discrete developmental skills and recommending whether to refer children for further evaluation, examination of the variation s in congruence by developmental domain and age cohort show that the nuances of these perspectives can be complex. Rather than viewing differences in perspectives as a concern or an indication of inaccuracies in judgment or scoring, findings of the present study suggest explanations for and discussions about disagreements between informants should be explored. In the present study, disagreements between parents and ECEs occurred primarily when rating emerging skills, and thus, variations in congruence might largely be a cohort and developmental domain suggest that differences might be due to complexities in the sequence of development, with the timing of importa nt developmental milestones heightening the awareness of informants about particular developmental skills. When exploring explanations for congruence, it is also important to consider the differing expectations or opportunities provided to young children a cross contexts. Most notably, results from the present study show developmental skills, research and practice should continue to preserve the richness of incorp orating multiple perspectives and uphold parents as essential informants and team members.

PAGE 155

155 APPENDIX : INSTRUCTIONS FOR PAR TICIPANTS Instructions for Early Childhood Educators Thank you for taking the time to complete this form. Please complete all of the i tems on this form as best as you can. When you complete this form, please do not discuss your ratings with will indicate how often the child performs the skill by marking one of three items: Select Yes if the child uses the skill consistently at different times and across different people and settings. Select Sometimes if you have occasionally observed the child use this sk ill, the child uses this skill with some settings and some people, or the child uses this skill with some help from an adult. Select Not Yet if you have not observed the child use this skill or if the child can only do this skill with complete adult assi stance. Item: Does your child eat with a fork? When the child is eating food that requires the use of a fork, you have observed the child use a fork all or most of the time, regardless of who he/she When the child is eating food that requires the use o f a fork, he/she switches between using a fork or hands to eat, needs a reminder to use a fork, or only When the child is eating food that requires the use of a fork, he/she only uses hands to eat, o r can only use a fork if you place your hand over his/her hand

PAGE 156

156 If you need any clarification on how to respond to a particular item, please feel free to contact: Cathleen Pasia Cell: (XXX) XXX XXXX Email: c.pasia@ufl.edu Item: Does your baby pick up a small toy with only one hand? When the child is playing with toys slightly larger than the palm of the hand, he/ she can hold it in one hand without dropping it, regardless of whether When the child is playing with toys slightly larger than the palm of the hand, he/she can only hold it for a mom ent before dropping it, can only do it when in certain positions, or can only do it with some small toys, then you would mark When the child is playing with toys slightly larger than the palm of the hand, he/she can only hold onto it if suppo rted by both hands, then you would mark

PAGE 157

157 Instructions for Parents Thank yo u for taking the time to complete this form. Please complete all of the items on this form as best as you can. When you complete this form, please do not discuss your ratings with any staff from [CENTER NAME] during different activities. For each item, you will indicate how often your child performs the skill by marking one of three items: Select Yes if your child uses the skill consistently at different times and across different people and settings. Selec t Sometimes if you have occasionally observed your child use this skill, your child uses this skill with some settings and some people, or your child uses this skill with some help from an adult. Select Not Yet if you have not observed your child use thi s skill or if your child can only do this skill with complete adult assistance. Item: Does your child eat with a fork? When your child is eating food that requires the use of a fork, you have observed your child use a fork all or most of the time, regardless of who he/she is eatin When your child is eating food that requires the use of a fork, he/she switches between using a fork or hands to eat, needs a reminder to use a fork, or only When your child is eating food that requires the use of a fork, he/she only uses hands to eat, or can only use a fork if you place your hand over his/her

PAGE 158

158 If you need any clarification on how to respond to a particular item, please feel free to contact: Cathleen Pasia Cell: (X XX) XXX XXXX Email: c.pasia@ufl.edu Item: Does your baby pick up a small toy with only on e hand? When your child is playing with toys slightly larger than the palm of the hand, he/she can hold it in one hand without dropping it, regardless of whether When your child is p laying with toys slightly larger than the palm of the hand, he/she can only hold it for a moment before dropping it, can only do it when in certain positions, or can only do it with some small toys, then you would mark When your child is play ing with toys slightly larger than the palm of the hand, he/she can only hold onto it if supported by both hands, then you would mark

PAGE 159

159 LIST OF REFERENCES Alpern, G. D., Boll, T. J., & Shearer, M. S. (1980). Alpern Boll Developmental Profile II Manual. Aspen, CO: Psychological Development Publications. American Academy of Pediatrics, Committee on Chi ldren with Disabilities (2001). Developmental surveillance and screening of infants and young children. Pediatrics, 108 192 196. American Educational Research Association, American Psychological Association, and the National Council on Measurement in Edu cation (1999). Standards for educational and psychological testing. Washington DC: American Educational Research Association. American Orthopsychiatric Association (1978). Developmental assessment in EPSDT. American Journal of Orthopsychiatry, 48 (1), 7 21. Bailey, D. B., & Wolery, M. (1992). Teaching infants and preschoolers with disabilities (2 nd ed). New York, NY: Macmillan Publishing Company. Bagnato, S. J. (2007). Authentic assessment for early childhood intervention: Best practices. New York, NY: G uilford Press. B arnett, D. W. (1984). An organizational approach to preschool services: Psychological screening, assessment, and int ervention. In C. Maher, R. Ilbac &J Zins (Eds.), Organizational psychology in schools: A handbook for practitioners (pp. 53 58) Springfield, IL : Charles C Thomas. Bayley, N. (1969). Bayley S cales of I nfant D evelopment. San Antonio, TX; Psychological Corporation. Bayley, N. (1993). Bayley II S cales of I nfant D evelopment San Antonio, TX: Psychological Corporation. Boan, C. H., Aydlett, L., & Multunas, N. (2007). Early childhood screening and readiness assessment. In R. J. Nagle (Ed.), Psychoeducational assessment of preschool children (4th ed., pp. 49 67). Mahwah, NJ: Lawrence Erlbaum. Bradley Johnson, S., & Johnson, C. M. (2007). Infant and toddler cognitive assessment. In B. A. Bracken (Ed.), The psychoeducational assessment of preschool children (4th ed., pp. 325 357). Needham Heights, MA: Allyn & Bacon. Brassard, M. R., & Boehm, A. E. (2007). Preschool assessment: Princi ples and practices. New York, NY: Guilford Press. Bricker, D., & Squires, J. (1989a). Infant Monitoring System. Eugene, OR: Center for Human Development, College of Oregon

PAGE 160

160 Bricker, D., & Squires, J. (1989b). The effectiveness of parental screening of at ri sk infants: The I nfant M onitoring Q uestionnaires. Topics in Early Childhood Special Education, 9 (3), 67 85. doi:10.1177/027112148900900306 Bricker, D., & Squires, J. (1999). Ages & Stages Questionnaires (ASQ): A parent completed, child monitoring system ( 2 nd ed .). Baltimore, MD: Paul H. Brooks Bricker, D., Squires, J., Kaminski, R., & Mounts, L. (1988). The validity, reliability, and cost of a parent completed questionnaire system to evaluate at risk infants. Journal of Pediatric Psychology, 13 (1), 55 68. Brigance, A. H., & Glascoe, F. P. (2002). Brigance Infant & Toddler Screen North Billerica, MA: Curriculum Associates. Bronfenbrenner, U. (1974). Developmental research, public policy, and the ecology of childhood. Child Development, 45, 1 5. Bronfenbre nner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Brooks Gunn, J., & Lewis, M. (1983). Screening and diagnosing handicapped infants. Topics in Early Childhood Special Education, 3 (1), 14 28. doi:10.1177/027112148300300107 Bureau of Labor Statistics, U.S. Department of Labor. (2013). Occupational outlook handout, 2012 2013 edition, childcare workers. Retrieved from http://www.bls.gov/ooh/personal care and service/childcare workers.h tm Burgess, D. B., Asher, K. N., Doucet, H. J., Reardon, K., & Daste, M. R. (1984). Parent report as a means of administering the prescreening developmental questionnaire: An evaluation study. Journal of Developmental and Behavioral Pediatrics, 5, 195 200 Campos, J. A., Squires, J., & Ponte, J. (2011). Universal developmental screening: Preliminary studies in Galicia, Spain. Early Child Development and Care, 181, 475 485. Child Care Aware of America. (2012). Child care in America: 2012 state fact sheets Retrieved from www.naccrra.org Children's Defense Fund (1978). EPSDT in practice: What's happening in the field? American Journal of Orthopsychiatry, 48 (1), 77 95. (2005). Child c are b asics. Retrieved from http://www. childrensd efense.org/child research data publications/data/child care basics.pdf Chiu, S., & DiMarco, M. A. (2010). A pilot study comparing two developmental screening tools for use with homeless children. Journal of Pediatric Health Care, 24 (2), 73 80. doi:10.1016/ j.pedhc.2009.01.003

PAGE 161

161 Australian day care setting: Developmental concerns of parents and carers. Journal of Paediatric Child Health, 39, 49 54. Colligan, R. C. (19 81). Prediction of reading difficulty from parental preschool report: A 3 year follow up. Learning Disability Quarterly, 4 (1), 31 37. doi:10.2307/1510710 Council on Children with Disabilities, Section on Developmental Behavioral Pediatrics, Bright Futures Steering Committee, & Medical Home initiatives for Children with Special Needs Project Advisory Committee (2006). Identifying infants and young children with developmental disorders in the medical home: An algorithm for developmental surveillance and scre ening. Pediatrics, 118 405 420. doi:10.1542/peds.2006 1231 Diamond, K. E. (1993). The role of parents' observations and concerns in screening for developmental delays in young children. Topics in Early Childhood Special Education, 13 (1), 68 81. doi:10.117 7/027112149301300108 Diamond, K. E., & L e Furgy, W. G. (1988). Screening for developmental handicaps: Outcomes from an early childhood screening program. Physical & Occupational Therapy in Pediatrics, 8 (1), 43 59. doi:10.1300/J006v08n01_04 Diamond, K. E., & Squires, J. (1993). The role of parental report in the screening and assessment of young children. Journal of Early Intervention, 17 (2), 107 115. doi:10.1177/105381519301700203 Dinnebeil, L. A., & Rule, S. (1994). Congruence between parents and professi about the development of young children with disabilities: A review of the literature. Topics in Early Childhood Special Education, 14 (1), 1 25. doi: 10.1177/027112149401400105 Dionne, C., Squires, J., Leclerc, D., Peloquin, J., & McKi nnon S. (2006). Cross cultural comparison of a F rench C anadian and U.S. developmental screening test. Developmental Disabilities Bulletin, 34 (1 2), 43 56. Division for Early Childhood. (1993). DEC recommended practices: Indicators of quality in programs for i nfants and young children with special needs and their families. Pittsburgh, PA: Author. Division for Early Childhood (2007). Promoting p ositive o utcomes for c hildren with d isabilities: Recommendations for c urriculum, a ssessment and p rogram e valuation. R etrieved from www.dec sped.org/.../Prmtg_Pos_ Outcomes _Companion_Paper.pdf Dunbar, S. B., & Reed, C. N. (1999). A developmental screening program in primary health care: Meeting the challenges of children in low income families. Infant Toddler Intervention 9 (2), 195 202.

PAGE 162

162 Dworkin, P. H. (2006). Promoting development through child health services introduction to the help me grow roundtable. Journal of Developmental and Behavioral Pediatrics, 27, S2 S4. doi:10.1097/00004703 200602001 00002 Earls, M. F., Andre ws, J. E., & Hay, S. S. (2009). A longitudinal study of developmental and behavioral screening and referral in N orth C arolina's assuring better child health and development participating practices. Clinical Pediatrics, 48 (8), 824 833. Eddey, G. E., Robey, K. L., Zumoff, P., & Malik, Z. Q. (1995). Developmental delay in children from birth to six years of age. Infant Toddler Intervention: The Transdiscplinary Journal, 5 (3), 233 241. Elbers, J., Macnab, A., McLeod, E., & Gagnon, F. (2008). The Ages and Stage s Questionnaires: Feasibility of use as a screening tool for children in C anada. Canadian Journal of Rural Medicine, 13 (1), 9 12. First, L. R., & Palfrey, J. S. (1994). The infant or young child with developmental delay. The New England Journal of Medicin e, 330 (7), 478 483. doi:10.1056/NEJM199402173300708 Flamant, C., Branger, B., Tich, S. N., Rochebrochard, E., Savagner, C., Berlie, I., & Roze, J. (2011). Public Library of Science, 6 (5) 1 6. Frankenburg, W. K. (1987). Revised Denver Prescreening Devel opmental Questionnaire (PDQ II). Denver, CO: DDM, Inc. Frankenburg, W. K., & Dodd, J. (1975). Denver Developmental Screening Test Revised (DDST). Denver, CO: DDM, Inc. Frankenburg, W. K., Dodd, J., Archer, P., Bresnick, B., Maschka, P., Edelman, N., & Shap iro, H. (1992). The Denver II: Training manual. Denver, CO: Denver Developmental Materials. Frankenburg, W. K., Fandal, A. W., & Thornton, S. M. (1987). Revision of the Denver prescreening developmental questionnaire. The Journal of Pediatrics, 110 653 6 57. Frankenburg, W. K., & North, A. F. (1974). A guide to screening for the EPSDT program under Medicaid. Washington, DC: Social and Rehabilitation Service, HEW Frankenburg, W. K., van Doorninck, W. I., & Liddell, T. (1976). Denver Prescreening Developme ntal Questionnaire (PDQ). Denver, CO: DDM, Inc. Frankenburg, W. K., van Doorninck, W. J., Liddell, T. N., & Dick, N. P. (1976). The D enver P rescreening D evelopmental Q uestionnaire (PDQ). Pediatrics, 57 (5), 744. Frey, A., & Walker, H. (2006). Education pol icy for children, youth, and families. In J. M. J enson & M. W. Fraser (Eds.), Social policy for children and families: A risk and resilience perspective (pp. 67 92). New York, NY: Chatham House/ Seven Bridges Press.

PAGE 163

163 Frisk, V., Montgomery, L., Boychyn, E., Young, R., vanRyn, E., McLachlan, D., & Neufeld, J. (2009). Why screening C anadian preschoolers for language delays is more difficult than it should be. Infants and Young Children, 22 (4), 290 308. Garbarino, J. (1990). The human ecology of early risk. In S. J. Meisels, & J. P. Shonkoff (Eds.), Handbook of early intervention (pp. 78 96). New York, NY: Cambridge University Press. Gasson, N., & Piek, J. P. (2003). A longitudinal investigation of the development of fullterm, preterm and 'at risk' infants from birth to four years using a parent report measure. Australian Journal of Psychology, 78 82. Glascoe, F. P. (1991). Developmental screening: Rationale, methods, and application. Infants & Young Children, 4 (1), 1 10. Glascoe, F. P. (2001). Are overreferra ls on developmental screening tests really a problem? Archives on Pediatric and Adolescent Medicine 155, 54 59. Glascoe, F. P. (2002). The B rigance I nfant and T oddler S creen: Standardization and validation. Journal of Developmental and Behavioral Pediatr ics, 23, 145 150. Glascoe, F. P. (2005). Screening for developmental and behavioral problems. Mental Retardation and Developmental Disabilities Research Reviews, 11 (3), 173 179. doi:10.1002/mrdd.20068 Glascoe, F. P. (2007a). Developmental and behavioral s creening. In J. Rojahn (Ed.), Handbook of intellectual and developmental disabilities. (pp. 353 371). New York, NY: Springer Publishing. doi:10.1007/0 387 32931 5_19 Glascoe, F. P. (2007b). ation of Developmental Status Nolensville, TN: Ells worth & Vandermeer Press Glascoe, F. P. (2010). Developmental and behavioral considerations. In S. Tanski, L. Garfunkel, P. Duncan, & M. Weitzman (Eds.), Performing preventative services: A Bright Futures handbook. Retrieved from http://brightfutures.aap. org/continuing_education.html Glascoe, F. P., Byrne, K. E., Chang, B., Stickland, B., Ashford, L. G., & Johnson, K. L. (1992). Accuracy of the Denver II in developmental screening. Pediatrics, 89, 1221 1224. Glascoe, F. P., & Dworkin, P. H. (1993). Obsta cles to effective developmental surveillance: Errors in clinical reasoning. Journal of Developmental and Behavioral Pediatrics, 14, 344 349. Glascoe, F. P., Martin, E. D., & Humphrey, S. (1990). A comparative review of developmental screening tests. Pedia trics, 86 (4), 547 554.

PAGE 164

164 Gollenberg, A. L., Lynch, C. D., Jackson, L. W., McGuinness, B. M., & Msall, M. E. (2010). Concurrent validity of the parent completed A ges and S tages Q uestionnaires, 2nd ed. with the B ayley S cales of I nfant D evelopment II in a low r isk sample. Child Care, Health & Development, 36 (4), 485 490. doi:10.1111/j.1365 2214.2009.01041.x Gracey, C. A., Azzara, C. V., & Reinherz, H. (1984). Screening revisited: A survey of U.S. requirements. The Journal of Special Education, 18 (2), 101 107. do i:10.1177/002246698401800203 Green, H., & Scriven, G. (1980). Current practices in the identification of children with handicaps. Early Child Development and Care, 6 (3 4), 125 134. doi:10.1080/0300443800060304 Guiberson, M., & Rodrguez, B. L. (2010). Meas urement properties and classification accuracy of two S panish parent surveys of language development for preschool age children. American Journal of Speech Language Pathology, 19 (3), 225 237. Hall, J. D., Bramlett, R. K., Barnett, D. W., & Cox, F. W. (1994 ). Classification of risk status in kindergarten screening: A comparison of alternative measures. Journal of Psychoeducational Assessment, 12 (2), 154 164. doi:10.1177/073428299401200205 Handal, A. J., Lozoff, B., Breilh, J., & Harlow, S. D. (2007a). Effect of community of residence on neurobehavioral development in infants and young children in a flower growing region of E cuador. Environmental Health Perspectives, 115 (1), 128 133. Handal, A. J., Lozoff, B., Breilh, J., & Harlow, S. D. (2007b). Sociodemograp hic and nutritional correlates of neurobehavioral development: A study of young children in a rural region of E cuador Revista Panamericana De Salud Pblica, 21 (5), 292 300. Haring, N.G., Lovitt, T.C., Eaton, M.D., & Hansen, C.L. (1978). The fourth R: Rese arch in the classroom. Columbus, OH: Charles E. Merrill Publishing Co. School Psychology Review, 13, 363 374. Harris, P. A., Thielke, R., Taylor, R., Payne, J. Gonzalez, N., & Conde, J. G. (2009). Research electronic data capture (REDCap): A metadata driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics 42 377 381. doi:10.1016/j.jbi.2008.08. 010. Heo, K. H., Squires, J., & Yovanoff, P. (2008). Cross cultural adaptation of a pre school screening instrument: Comparison of K orean and US populations. Journal of Intellectual Disability Research, 52 (3), 195 206. Hix Small, H., Marks, K., Squires, J., & Nickel, R. (2007). Impact of implementing developmental screening at 12 and 24 months in a pediatric practice. Pediatrics, 120 (2), 381 389. doi:10.1542/peds.2006 3583

PAGE 165

165 Ikeda, M. J., Neessen, E., & Witt, J. (2008). Best practices in universal screening In A. Thomas, and J. Grimes (Eds.), Best practices in school psychology V. (pp. 103 114). Bethesda, MD: National Association of School Psychologists. Ireton, H. (1990). Developmental screening measures In J. Goldman (Ed.), Developmental assessment in clinical child psychology: A handbook. (pp. 78 99). Elmsford, NY: Pergamon Press. Improving Head Start for School Readiness Act, 42 U.S.C. § 1304.20 2007 Individuals with Disabilities Education Act, 20 U.S.C. 1400 § 602(23), 2004 Individuals with Disabil ities Education Act, 20 U.S.C. 1419 § 300, 1986 Jackson, B. J., & Needelman, H. (2007). Building a system of child find through a 3 tiered model of follow up. Infants & Young Children, 2 0 (3), 255 265. doi:10.1097/01.IYC.0000277756.00938.e2 Janson, H., & Sq uires, J. (2004). Parent completed developmental screening in a N orwegian population sample: A comparison with US normative data. Acta Paediatrica, 93 1525 1529. Jee, S. H., Szilagyi, M., Ovenshire, C., Norton, A., Conn, A., Blumkin, A., & Szilagyi, P. G (2010). Improved detection of developmental delays among young children in foster care. Pediatrics, 125 (2), 282 289. doi:10.1542/peds.2009 0229 Johnson, D., Poteat, G. M., & Kushnick, T. (1986). Comparison of mental age estimates made by pediatricians an d mothers of preschool children. Journal of Pediatric Psychology, 11, 385 392. Kapci, E. G., Kucuker, S., & Uslu, R. I. (2010). How applicable are A ges and S tages Q uestionnaires for use with T urkish children? Topics in Early Childhood Special Education, 30 (3), 176 188. doi:10.1177/0271121410373149 Katoff, L., & Reuter, J. (1980). Review of developmental screening tests for infants. Journal of Clinical Child Psychology, 9 (1), 30 34. doi:10.1080/15374418009532939 Katz, K. S. (1989). Strategies for infant asse ssment: Implications of P.L. 99 457. Topics in Early Childhood Special Education, 9 (3), 99 109. Kenny, T. J., & Culbertson, J. L. (1993). Developmental screening for preschoolers. In D. J. Willis (Ed.), Testing young children: A reference guide for deve lopmental, psychoeducational, and psychosocial assessments. (pp. 73 100). Austin, TX: PRO ED. Kerstjens, J. M., Bos, A. F., ten Vergert, E. M. J., de Meer, G., Butcher, P. R., & Reijneveld, S. A. (2009). Support for the global feasibility of the A ges and S tages Q uestionnaire as developmental screener. Early Human Development, 85 (7), 443 447. doi:10.1016/j.earlhumdev.2009.03.001

PAGE 166

166 Kim, M. M., & O'Connor, K. S. (1996). Do parents and professionals agree on the developmental status of high risk infants? Pediatri cs, 97 (5), 676 681. Klamer, A., Lando, A., Pinborg, A., & Greisen, G. (2005). Ages and S tages Q uestionnaire used to measure cognitive deficit in children born extremely preterm. Acta Paediatrica, 94 (9), 1327 1329. doi:10.1080/08035250510042825 Knobloch, H. Stevens, F., & Malone, A. (1980). Manual of development diagnosis: The administration and interpretation of the revised Gesell and Armatruda developmental and neurologic examination. Hagerstown, MD: Harper & Row. Knobloch, H., Stevens, F, Malone, A., El lison, P., & Risemberg, H. (1979). The validity of parental reporting of infant development. Pediatrics, 63, 872 878. Kochanek, T. T. (1993). Enhancing screening procedure s for infants and toddlers: The a pplication of knowledge to public policy and program initiatives. In M. A. Graham (Ed.), Implementing early intervention: From research to effective practice (pp. 46 66). New York, NY: Guilford Press. Lichtenstein, R., & Ireton, H. (1984). Preschool screening: Identifying young children with developmental a nd educational problems. Orlando, FL: Grune & Stratton. Lindsay, N. M., Healy, G. N., Colditz, P. B., & Lingwood, B. E. (2008). Use of the A ges and S tages Q uestionnaire to predict outcome after hypoxic ischaemic encephalopathy in the neonate. Journal of P aediatrics & Child Health, 44 (10), 590 595. doi:10.1111/j.1440 1754.2008.01388.x Macy, M. (2012). The evidence behind developmental screening instruments. Infants & Young Children, 25, 19 61. doi : 10.1097/IYC.0b013e31823d37dd Marks, K., Hix Small, H., Clar k, K., & Newman, J. (2009). Lowering developmental screening t hresholds and raising quality improvement for preterm children. Pediatrics, 123, 1516 1523. McCarthy, D. (1972). New York, NY: Psychological Corp. McCr ae, J. S., Cahalane, H., & Fusco, R. A. (2011). Directions for developmental screening in child welfare based on the Ages and Stages Questionnaires. Child & Youth Services Review, 33 (8), 1412 1418. McLean, M., & McCormick, K. (1993). Assessment and evalua tion in early intervention. In L. F. Pearl (Ed.), Family centered early intervention with infants and toddlers: Innovative cross disciplinary approaches (pp. 43 79). Baltimore, MD: Paul H Brookes Publishing. Meisels, S. J., & Wasik, B. A. (1990). Who shoul d be served? I dentifying children in need of early intervention. In J. P. Shonkoff (Ed.), Handbook of early childhood intervention. (pp. 605 632). New York, NY: Cambridge University Press.

PAGE 167

167 Miller, S. A., Manhal, M., & Mee, L. L. (1991). Parental beliefs, p arental accuracy, and children's cognitive performance: A search for causal relations. Developmental Psychology, 27, 267 276. Moore, B. D. (1978). Implementing the developmental assessment component of the EPSDT program. American Journal of Orthopsychia try, 48 (1), 22 32. Nagle, R. J. (2007). Issues in preschool assessment. In B. A. Bracken (Ed.), The psychoeducational assessment of preschool children (4th ed. pp. 29 48). Needham Heights, MA: Allyn & Bacon. National Association of School Psychologists ( 2009). Early c hildhood a ssessment (Position s tatement). Bethesda, MD: Author. National Association for the Education of Young Children (2003, November). Early c hildhood c urriculum, a ssessment, and p rogram e valuation: Building an e ffective, a ccountable s y stem in p rograms for c hildren b irth through a ge 8. Retrieved from www.naeyc.org/positionstatements/cape National Association for the Education of Young Children (2008). Overview of NAEYC e arly c hildhood p rogram s tandards. Retrieved from http://www.naeyc. org/files/academy/file/OverviewStandards.pdf National Association for the Education of Young Children. (2012, October). All criteria document. Retrieved from http://www.naeyc.org/academy/primary/viewstandards National Research Council (2008). Early c hild hood a ssessment: Why, w hat, and h ow. Committee on Developmental Outcomes and Assessments for Young Children, C. E. Snow, & S. B. Van Hemel (Eds.). Washington, DC: The National Academies Press. Neisworth, J. T., & Bagnato, S. J. (2005). DEC recommended pra ctices: Assessment. In S. Sandall, M. L. Hemmeter, B. J. Smith, & M. E. McLean (Eds.), DEC recommended practices: A comprehensive guide for practical application in early intervention/early childhood special education (pp. 45 61). Longmont, CO: Sopris Wes t. Neisworth, J. T., & Fewell, R. R. (1990). Forward. Topics in Early Childhood Special Education, 10 (3), ix x. doi: 10.1177/027112149001000301 Newborg, J., Stock, J. R., Wnek, L., Guidubaldi, J., & Svinicki, J. (1984). Battelle D evelopmental I nventory. I tasca, IL: Riverside Publishing. Newborg, J., Stock, J. R., Wnek, L., Guidubaldi, J., & Svinicki, J. (2004). Battelle D evelopmental I nventory (2 nd ed.) Chicago, IL: Riverside Publishing. Obama for America (2012). ower Americans with disabilities. Retrieved from www.barackobama.com

PAGE 168

168 Owen, M. T., Ware, A. M., & Barfoot, B. (2000). Caregiver mother partnership behavior and the quality of caregiver child and mother child interactions. Early Childhood Research Quarterly, 15, 413 428. Paget, K. D., & Nagle, R. J. (1986). A conceptual model of preschool assessment. School Psychology Review, 15 (2), 154 165. Plomgaard, A. M., Hansen, B. M., & Greisen, G. (2006). Measuring developmental deficit in children born at gestational age less than 26 weeks using a parent completed developmental questionnaire. Acta Paediatrica, 95, 1488 1494. doi:10.1080/08035250600684438 Portney, L. G., & Watkins, M. P. (2000). Foundations of clinical research: Applications to practice (2 nd ed.). Upp er Saddle River, NJ: Prentice Hall Health. Richter, J., & Janson, H. (2007). A validation study of the N orwegian version of A ges and S tages Q uestionnaires. Acta Paediatrica, 96, 748 752. doi:10.1111/j.1651.2227.2007.00246.x Ryan Krause, P., Meadows Oli ver, M., Sadler, L., & Swartz, M. K. (2009). Developmental status of children of teen mothers: Contrasting objective assessments with maternal reports. Journal of Pediatric Health Care, 23 (5), 303 309. doi:10.1016/j.pedhc.2008.09.001 Rydz, D., Srour, M., O skoui, M., Marget, N., Shiller, M., Birnbaum, R., Majenmer, A., & Shevell, M. I. (2006). Screening for developmental delay in the setting of a community pediatric clinic: A prospective assessment of parent report questionnaires. Pediatrics, 118, e1178 e118 6. doi:10.1542/peds.2006 0466 Salvia, J., & Ysseldyke, J. E. (1985). Assessment in special and remedial education (3 rd ed.). Boston, MA: Houghton Mifflin Company. ( work? Do they help? Pediatrics, 116, 174 179. Scott, K. G., & Carran, D. T. (1989). Identification and referral of handicapped infants. In H. J. Walberg (Ed.), Hand book of special education: Research and practice : V ol. 3 Low incidence conditions. (pp. 227 241). Elmsford, NY: Pergamon Press. Scott, F. G., Lingaraju, S., Kilgo, J. L., Kregel, J., & Lazzari, A. (1993). A survey of pediatricians on early identification and early intervention services. Journal of Early Intervention, 17, 129 138. doi: 10.1177/105381519301700205 Sheehan, R. (1988). Involvement of parents in early childhood assessment. In T. D. Wachs & R. Sheehan (Eds.), Assessment of young developmentally d isabled children. (pp. 75 90). New York, NY: Plenum Press.

PAGE 169

169 Sheldrick, R. C., & Perrin, E. C. (2009). Surveillance of children's behavior and development: Practical solutions for primary care. Journal of Developmental and Behavioral Pediatrics, 30 (2), 15 1 153. doi:10.1097/DBP.0b013e31819f1bfb Shelton, T. L. (1989). The assessment of cognition/intelligence in infancy. Infants and Young Children, 1, 10 25. Shonkoff, J. P., & Phillips, D. A. (Eds.). (2000). From n eurons to n eighborhoods: The s cience of e arl y c hildhood d evelopment. Washington DC: National Academy Press. Simas, A. B., Barreto Souza, W., & Rocha, A. V. (2010). Improved estimators for a general class of beta regression models. Computational Statistics and Data Analysis, 54, 348 366. Skellern completed developmental questionnaire: Follow up of ex premature infants. Journal of Paediatric Child Health, 37, 125 129. Smith, C., Wallen, M., Walker, K., Bundy, A., Rolinson, R., & Badawi, N. (2012). Validity of the fine motor area of the 12 month Ages and Stages Questionnaire in infants following major surgery. Physical and Occupational Therapy in Pediatrics, 32 (3), 260 271. Smith, R. D. (1978). The use of developmental screening tests by pri mary care pediatricians. Journal of Pediatrics, 93, 524 527. Smithson, M. & Verkuilen, J. (2006). A better lemon squeezer? Maximum likelihood regression with beta distributed dependent variables. Psychological Methods, 11 54 71. Snyder P., McLaughlin T., & McLean M. (in press). Recommended practices in assessment. In M. McLean M.L. Hemmeter, & P. Snyder (Eds.), Essential elements for assessing infants and preschoolers with special needs Boston, MA: Pearson. Snyder, P., Thompson, B., & Sexton, D. ( 1992, January). Congruence in maternal and professional early intervention assessments of young children with disabilities Paper presented at the annual meeting of the Southwest Educational Research Association, Austin, TX. Squires, J., & Bricker, D. (19 91). Impact on completing infant developmental questionnaires on at risk mothers. Journal of Early Intervention, 15 (2), 162 72. Squires, J., & Bricker, D. (2009). Ages and S tages Q uestionnaires (3 rd ed.) : A parent completed child monitoring system. Baltimo re, MD: Brookes. Squires, J., Bricker, D., & Potter, L. (1997). Revision of a parent completed developmental screening tool: Ages and St ages Q uestionnaires. Journal of Pediatric Psychology, 22 (3), 313 328. doi:10.1093/jpepsy/22.3.313

PAGE 170

170 Squires, J., Carter, A ., & Kaplan, P. (2003). Developmental monitoring of children conceived by intracytoplasmic sperm injection and in vitro fertilization. Fertility & Sterility, 79 (2), 453. doi:10.1016/S0015 0282(02)04685 X Squires, J., Katzev, A., & Jenkins, F. (2002). Early screening for developmental delays: Use of pare nt completed questionnaires in O regon's healthy start program. Early Child Development and Care, 172 (3), 275 82. Squires, J., Nickel, R. E., & Eisert, D. (1996). Early detection of developmental problems: Str ategies for monitoring young children in the practice setting. Journal of Developmental and Behavioral Pediatrics, 17 (6), 420 427. doi:10.1097/00004703 199612000 00008 Squires, J., Potter, L., Bricker, D. D., & Lamorey, S. (1998). Parent completed developm ental questionnaires: Effectiveness with low and middle income parents. Early Childhood Research Quarterly, 13 (2), 345 54. Squires, J., Twombly, E., Bricker, D., & Potter, L. (2009). ASQ Baltimore, MA: Paul H. Brookes Publishing. Streiner, D. L. (2003). Diagnosing tests: Using and misusing diagnostic and screening tests. Journal of Personality Assessment, 81, 209 219. Suen, H. K., Logan, C. R., Neisworth, J. T., & Bagnato, S. (1995). Parent professional congruence: Is it necessary? Journal of Early Intervention, 19, 243 252. Taylor, R. L. (1993). Instruments for the screening, evaluation, and assessment of infants and toddlers. In M. A. Graham (Ed.), Implementing early intervention: From research to effective practice (pp. 157 182). New Yor k, NY: Guilford Press. Teska, J. A., & Stoneburner, R. L. (1980). The concept and practice of second level screening. Psychology in the Schools, 17 (2), 192 195. doi:10.1002/1520 6807(198004)17:2<192::AID PITS2310170210>3.0.CO;2 B Thompson, L. A., Tuli, S. Y., Saliba, H., DiPietro, M., & Nackashi, J. A. (2010). Improving developmental screening in pediatric resident education. Clinical Pediatrics, 49 (8), 737 742. doi:10.1177/0009922810363818 Thorndike, R. L., Hagen, E. P., & Sattler, J. M. (1985). Stanford B inet I ntelligence S cale (4 th ed.). Chicago, IL: Riverside. Thurlow, M. L., & Gilman, C. J. (1999). Issues and practices in the screening of preschool children. In J. Kalesnik (Ed.), Assessing and screening preschoolers: Psychological and educational dimen sions (2nd ed. pp. 72 93). Needham Heights, MA: Allyn & Bacon. Stages Questionnaires: Feasibility of postal surveys for child follow up. Early Human Development 87 (10), 671 676.

PAGE 171

171 Tsai, H. A., McClelland, M. M., Pratt, C., & Squires, J. (2006). Adaptation of the 36 month A ges and S tages Q uestionnaire in T aiwan: Results from a preliminary study. Journal of Early Intervention, 28 (3), 213 225. United States Departme nt of Education, Offices of Special Education and Rehabilitative Services (2010). Thirty five y ears of p rogress in e ducating c hildren with d isabilities through IDEA. Washington DC: Author. United States Department of Labor, Bureau of Labor Statistics. (2 011 ). Women in t he labor force: A databook (2011 edition). Retrieved from http://www.bls.gov/cps/wlf databook2011 .htm Younger, M. S. (1979). The first course in linear regression North Scituate, MA: Duxbury Press. Ysseldyke, J. E., & O'Sullivan, P. J. (1 987). Predicting preschool screening referral rates from district demographic data Journal of School Psychology, 25 (2), 119 129. doi:10.1016/0022 4405(87)90021 5 Yu, L., Hey, E., Doyle, L. W., Farrell, B., Spark, P., Altman, D. G., & Duley, L. (2007). Eva luation of the A ges and S tages Q uestionnaires in identifying children with neurosensory disability in the M agpie trial follow up study. Acta Paediatrica, 96, 1803 1808. Zehrbach, R. R. (1975). Determining a preschool handicapped population. Exceptional Ch ildren, 42 (2), 76 83

PAGE 172

172 BIOGRAPHICAL SKETCH Cathleen Pasia received her Bachelor of Science in psychology from Michigan State University in the summer of 2007. At the University of Florida, she majored in school psychology with a specialization in early ch ildhood studies. Throughout her graduate studies, her practicum placements reflected a focus on working with young children and their families. While obtaining her doctorate degree, she worked on several research projects, including Institute for Education Sciences Goal 2 Development and Innovation grants, under the direction of Dr. Patricia Snyder. Her research interests include professional development for early childhood practitioners and family engagement in assessment, intervention, and early childhood programs. She completed her predoctoral internship at Cypress Fairbanks Independent School District, which is accredited by the American Psychological Association. Cathleen graduated with her Doctor of Philosophy in August 2013.