Identifying the maintaining variables of undesirable behavior

MISSING IMAGE

Material Information

Title:
Identifying the maintaining variables of undesirable behavior
Physical Description:
v, 63 leaves : ; 28 cm.
Language:
English
Creator:
Rodgers, Teresa A., 1957-
Publication Date:

Subjects

Subjects / Keywords:
Self-injurious behavior   ( lcsh )
Psychology thesis Ph. D
Dissertations, Academic -- Psychology -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 60-62).
Statement of Responsibility:
by Teresa A. Rodgers.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001701178
oclc - 25350790
notis - AJB3364
System ID:
AA00002105:00001

Full Text














IDENTIFYING THE MAINTAINING VARIABLES
OF UNDESIRABLE BEHAVIOR





By


TERESA


RODGERS


OF
THE


A DISSERTATION SUBMITTED TO
THE UNIVERSITY OF FLORIDA IN
REQUIREMENTS FOR THE DEGREE


THE GRADUATE SCHOOL
PARTIAL FULFILLMENT OF
OF DOCTOR OF PHILOSOPHY


UNIVERSITY


OF FLORIDA


1991

















ACKNOWLEDGEMENTS


I would


Edward


like


Malagodi,


to thank


Henry


committee


Pennypacker,


members


Carolyn


Marc


Tucker,


Branch,

and


William


Working


shaping


behavioral


repertoire


throughout


years


graduate


school.


would


like


to give


special


thanks


chair


Brian


Iwata,


his


time


and


assistance


in the


preparation


this


manuscript


and


development


career.


Several


others


have


provided


assistance


the


preparation


this


manuscript.


Peter


Andre,


Suneeta


Jagtiani,


Martina


Jonak,


Jodi


Mazaleski,


Richard


Smith,


Timothy


Vollmer


and


Jennifer


Zarcone


served


experimenters


reviewed


during


earlier


the


drafts


course


the


study.


manuscript.


Michael


thanks


Cripe


to each


these


individuals.


husband,


Michael


Stoutimore,


provided


much


more


than


editorial


comments.


His


support,


guidance,


and


interactions


throughout


the


years


have


been


invaluable


me.


Finally


would


like


thank


father


who


encouraged





















TABLE OF CONTENTS





ACKNOWLEDGEMENTS......................................... ii


ABSTRACT.............................................................. iv


GENERAL METHOD..............................
Subjects..................................
Human Subjects Protection...................
Motivational Assessment Scale (MAS).........
Time-referenced assessment.................
Behavior-referenced assessment..............
Response definitions and direct observation.
Experimenters, observers, MAS raters........
Reliability.................................


.............
.............
.............
.............


. . a. .
* a

. . a. ...
.............
.............


EXPERIMENT 1........
Subjects............
Experimental Design.
Results.............
Discussion..........


...........................
...........................
*. .. .. .. . ..
S *S*


.....


..'..
0 0
** *


EXPERIMENT 2..............................................
Subjects..................................................
Experimental Design.....................................
Results...................................................
Discussion.............. .......... ........................


GENERAL DISCUSSION................. .... .......... ..... 54


REFERENCES...... ........ ........ ............ ... ........... 60


BIOGRAPHICAL SKETCH.............................. ........ 63


INTRODUCTION, .,....,..,,.,.... ................... ......
















Abstract


School


of Dissertation


University


Fulfillment
Degree


Presented


of Florida


Requirements


Doctor


the


in
for


Graduate


Partial
the


Philosophy


IDENTIFYING


THE


MAINTAINING


VARIABLES


OF UNDESIRABLE


BEHAVIOR

By


Teresa


August,


Rodgers

1991


Chairman:


Major


Brian


Department:


Iwata


Psychology


Three


methods


are


use


commonly


identify


the


maintaining


variables


aberrant


behavior:


indirect


assessment


through


interviews,


questionnaires,


and


rating


sca


descriptive


observational


analysis


of events


they


occur,


natural


observation


and


field


studies);


and


experimental


analysis


exposure


analogue


situations.


the


present


study,


within


-subject


comparisons


were


made


between


indirect


experimental


assessment


methods.


Experiment


compared


results


from


the


Motivational


Assessment


Scale


(MAS)


those


from


a brief


(time-


referenced)


experimental


analysis


, with


34 developmentally










analysis


with


those


from


a more


extensive


behavior-referenced


analysis


10 of


the


subjects


from


the


first


experiment.


The


MAS


was


found


to be


an invalid


assessment


tool


, in


that


the


results


two


assessments


(MAS


and


time-


referenced


analysis


matched


only


8 of


the


subjects.


results


the


two


experimental


analyses


matched


8 subjects,


analyst


provides


indicating

a better


that


a time-limited


indication


the


experimental


maintaining


variables


of SIB


than


an indirect


assessment


such


as the


MAS.















INTRODUCTION


Identification


the


maintaining


variable


undesirable


thorough


effective

Durand &

Richman,

Crisp & N


behavior


prior


understanding


treatment


Crimmins


1982


ewton,


to treatment


the


(Carr,


, 1988;


Johnson

1988).


behavior


1977;


Iwata


Demchak


, Dorsey


& Baumeister,

An increased


will


and


result


more


& Halle,


Slifer


1978; St

interest


in a more


efficient


1985;


, Bauman,


urmey


and


Carlsen


emphasis


on assessment


of problem


behaviors


have


resulted


varied


technology


asses


sment


and


a journal,


Behavioral


Assessment,


devoted


solely


the


assessment


of behavior


problems.


In a recent


review


the


literature,


Iwata


, Vollmer


and


Zarcone


(1990)


proposed


ass


ification


of assessment


methods

control


based

exerte


on the

d over


type


of data


behaviors


collected


during


asses


and


the


sment.


degree


The


three


classifications


are:


indirect


assessment


through


interviews


, questionnaires,


rating


scal


uncontrolled


observation


of naturally


occurring


events


naturalistic


observation


and


field


studies


: and 31


.g.,












The


most


time-efficient


method


the


indirect


assessment.


Indirect


assessments


are


based


on verbal


reports


from


sources


other


than


the


subject.


Informants


may


include


family


members


, teachers,


direct


-care


staff,


and


counselors/therapists.


Chamberlain


Reid


(1987)


described


a checklist


called


the


Parent


Daily


Report


(PDR)


and


proposed


use


as a compromi


between


other


more


general


checkli


and


direct


observation.


They


suggested


that


the


problems


usually


associated


with


indirect


methods


obtaining


information


to aggregate


such


as systematic


perceptions


over


bias


long


, requiring


periods


the


time


informant


(since


problem

limited


began,

utility


or since


working


assessing


with


short


the


-term


individual)


changes


and


were


eliminated


PDR.


Furthermore


, Chamberlain


and


Reid


said


that


the


PDR


eliminated


these


problems


asking


the


parent


over


phone


to note


whether


or not


one


behaviors


had


occurred


over


the


past


24 hours


However,


the


PDR


assessment


indicates


only


occurrence


or non-


occurrence


behavior


, rather


than


information


with


respect


frequency


or environmental


correlates.


Durand


and


Crimmins


(1988)


argued


that


assessment


methods


involving


direct


observation


and


manipulation


~inS1onob~nnne


nrtrl


i-n a nrani 1ana


a aa a


*I*** g I U *


J- a


r1 imuihoranm


r n


IrlU rEl rn












the


Motivation


Assessment


Scale


(MAS).


The


sixteen


questions


on thi


scale


focus


on four


possible


maintaining


variables


for an

form of


individual'


attention,


positive


positive


reinforcement


reinforcement


the


form


tangible


items


.e.,


food


, toys,


or activities),


negative


reinforcement


.e.,


escape


from


demands),


and


sensory


reinforcement


, self-stimulation


or automatic


reinforcement)


Durand


and


Crimmins


administered


the


MAS


a teacher


and


a teacher'


aide


each


50 students


who


exhibited


SIB.


They


presented


inter-rater


and


test


-retest


reliability


experiment


MAS,


which


as well


behavior


as results


of subjects


from


various


analogue


conditions


was


compared


the


subjects


' rating


MAS.


They


concluded


that


the


MAS


was


a reliable


and


valid


indicator


behavioral


function,


which


could


used


instead


analogue


assessment


methods.


A problem


with


Durand


and


Crimmins


(1988)


study


was


that


the


experimenters


used


only


highly


trained


informants


who


also


were


familiar


with


subjects,


placing


limitation


on the


observations;


use


that


MAS


, highly


similar


trained


that


staff


are


direct


required.


ess


well


trained


individual


or individuals


less


familiar


with


the


subjects


were


asked


to complete


questionnaires


. the












to replicate


reliability


results


the


MAS


with


both


institutional


staff


cla


ssroom


teachers


as respondents.


The


sec


method


of assessment


described


Iwata


. (1990)


observations


naturalistic


a subject'


observations;


behavior


these


the


are


direct


environment


where


the


target


behavior


typically


occurs.


The


observer


may


record in

of events


occurrence


formation


in a behavior


the


a variety

r log, re


target


ways


cording


behaviors


noting


the


, or using


the


occurrence


a behavior


sequence

or non-


code


to record


the


frequency


of events.


The


recording


may


occur


continuously


or on a time-sampling


schedule.


Each


these


recording


methods


some


advantages


and


limitations


The


information


provides


a descriptive


account


of events


and


their


sequence;


however


does


necessarily


provide


quantifiable


data.


This


method


cumber


some


and


allows


the


recording


only


a few


events


one


time.


The


occurrence/non-occurrence


method


is more


precise


and


provides


some


estimate


events


distortion


frequency,


however,


depending


the


as well


estimate


on the


as a time-based


of frequency


duration


rate


account


subject


the


behavior


behavior


code


groups


respon


ses


into


classes,


and


so allows


the


recording


several


events


rocnnncac


0i n 1 -i 1 s


nf coiro a


I ~ n e-i n M nooo


a a- ra


*-4imo


LL ~


Ilnw~ttar


ar












There


are


observations


at least


as a whole:


problems


one


is that


with


the


naturalistic


controlling


variable(


may


not


be readily


apparent


the


environment;


another


that


the


method


does


not


demonstrate


functional


relations


(i.e.


, the


information


correlational).


The


first


problem


difficult


to detect


and


overcome


because


the


conditioning


that


resulted


the


initial


control,


either


antecedent


or consequent,


probably


occurred


an earlier


time.


However,


some


pairing


of stimuli


and


response


must


continue


occur


at least


occasionally,


or control


variable


would


have


been


extinguished;


the


variable


and


effects


might


environment.


intermittently


example,


observed


may


that


the


the


natural


subject


initially


deal


emitted


attention


subject,


almost


some


from


aberrant


staff.


every


response


response


this


may


and


response


have


received


was


a great


new


occasioned


more


concern


from


staff.


time


passed


and


staff


habituated


response,


they


might


have


provided


attention


intermittently.


Finally,


behavior


would


have


been


maintained


on a lean


schedule


of reinforcement.


Natural


observations


would


reveal,


this


case,


that


staff


rarely


attend


the


response.


However


, it


this


limited


amount


of attention


that


maintains


the


rePsnonse-


This


tvne


JJ I












respect


than


to functional


pursue


relations


an experimental


there


analy


no solution


the


other


behavior


Bijou,


Peterson,


and


Ault


(1968)


described


a system


that


differed


from


previous


descriptive


studies


that


frequency


data


were


collected


rather


than


data


the


form


of rating


scales


, running


descriptions


of situations


, or some


other


indirect,


subjective


measurement.


The


collection


frequency


data


natural


environment


allowed


the


comparison


data


from


naturalistic


experimental


studies.


events


Bijou


et al. specified


to be observed,


recorded


behavioral


data


and


these


stimulus


events


two


ways.


One


method


consisted


logging


the


occurrence


responses,


and


sometimes


duration


of each


response,


and


the


other


consisted


registering


the


frequencies


occurrences


and


non-occurrences


within


time


intervals


The


observation


system


outlined


Bijou


et al.


(1968)


eliminated


problems


inadequate


recording


behavior


limited


response


definitions,


problems


common


descriptive


or field


collected


studi


are


(Johnson


quantified,


Bolstad,


1973).


responses


are


The


data


observable


and


measurable,


and


methodology


produces


replicable


results.


The


limitation


remains,


because


only


naturalistic


observation


nr~rnirroA


4-lin


th im


Thai n*~ in't ncr


vari ahl l


may


not


be readily


.












developed


Touchette,


MacDonald,


and


Langer


(1985)


Frequenci


some


interval


time


were


recorded


on a


scatterplot


grid.


Different


symbols


were


used


occurrence


the


response


during


the


interval


, for


rates


the


response,


high


rates


the


response.


The


grid


sheet


formed


a visual


display


the


data


without


formal


calculations,


time


interval


containing


high


rates


behaviors


were


identified


grouping


high


rate


symbols


within


across


days.


The


criteria


for


and


high


rates


were


determined


individually


based


on the


frequency


situational


the


response.


variable


were


Activities


identified


, staff,


the


and


high


other


rate


periods.


Touchette


et al.


(1985)


described


the


scatterplot


identifying


patterns


of responding


occurring


the


natural


setting


, and


these


patterns


responding


suggested


sources


of environmental


control.


Unlike


line-graph


display


data


that


reveal


only


average


rate


and


necessarily


condenses


pattern

claimed


the


that


response


the


into


scatterplot


larger

keeps


units,


the


Touchette


pattern


responding


useful


intact.


when


They


target


proposed


behavior


that


scatterplot


frequent


and


would


there


nf- nhuvi ni11


flflrrc acnnflnti nfl


r-l..


uni rnnman+t I


rnnnti 4 i I nn


haf wapn












specific


stimuli


or contingencies


operating


on the


behavior.

scatterplot


Touchette'

analysis,


suggested


treatment,


rearrange


the


based


subject'


on a


schedule


so that


condition


associated


with


high


rates


either


longer


occurs,


or occurs


at a lower


frequency


or different


time.


This


treatment


does


not


take


into


account


the


specific


aspects


identified


problem


situations


that


maintain


the


behavior.


Re-arranging


schedule


may


result


initial


decrease


the


response;


however,


when


the


subject


been


the


new


situation


and


exposed


to similar


stimuli


or contingencies,


problem


behavior


may


reoccur.


Another


problem


with


the


scatterplot


method


that


grouping


frequencies


may


distort


response


pattern.


The


actual


frequency


response


lost,


and


the


condensation


criteria


are


inappropriate,


important


periods


time

seems


may b

that


e excluded


would


from


cons


involve


ideration


little


effor


as problem

t to record


times.

the


actual


frequency


each


interval,


and


then


determine


the


high


rate


rate


times.


The


third


approach


ass


essment


described


Iwata


(1990)


experimental


analysis.


the


most


thorough


vari abl es -


method


and


obtaining


invnlv


information


aranm ncr


about


flnn~prsin hr snf


maintaining


tn tesCt+


1 t JI


,,












maintaining


variable


based


on the


treatment


that


was


the


most


effective,


other


to provide


special


arrangements


consequences


the


behavior


controlled


situations


that


recreate


possible


reinforcement


available


the


typical


consequences


environment.


are


Special


standard


the


arrangements


reinforcer


classifications


.e.


, POS


itive


or negative


reinforcement);


additionally


, arrangements


of particular


stimuli


can


individualized


based


on the


subject'


specific


history


The


first


type


test


is a post-hoc


analysis


that


may


not


identify


the


maintaining


variable


the


behavior


may


simply


come


under


the


control


the


new


arrangement


contingencies.


An example


of such


an analyst


the


study


Repp,


, and


Barton


(1988),


who


tested


hypotheses


regarding


the


maintaining


variables


predicted


from


baseline


observations


exposing


subjects


two


three


different


treatments

behavior,


either


extinction


extinction


of positively


negatively


reinforced


reinforced

behavior.


increased


stimulation.


Based


on effects


the


treatments


they


concluded


that


their


hypotheses


had


been


correct


three


their


subjects.


Repp


et al.


(1988)


did


not


provide


baseline


data,


thereby


preventing


independent


confirmation


hvnothes


es. A


difficulty


this


uL LIIG


U












analysis


they


collected


data


specific


the


possible


maintaining


variable


each


subject.


That


, to


determine


they


the


collected


behavior


data


might


be positively


on consequences


behavior


reinforced,


including


teacher


proximity,


negatively


verbal


reinforced


attention

behavior,


from


they


staff,

noted


and

what


touching.

was


removed


following


the


behavior


self-stimulation


maintaining


variable


, they


noted


whether


the


subject


had


environmental


stimulation


available


prior


occurrence


behavior.


Several


examples


second


type


, (i


.e.,


arranging


analogue


conditions


to test


maintaining


variables


have


been


reported.


Carr,


Newsom,


Binkoff


(1980),


after


observing


subjects


their


typical


environment,


hypothesized


that


aggressive


responses


were


motivated


escape


from


demands.


They


then


systematically


manipulated


environmental


variables


test


escape-maintained


aggression.


The


variable


manipulated


were


demands


, reinforcement


correct


responses,


stimuli


associated


with


sessions


(safety


signals),


escape


from


session.


Sess


ions


were


5-min


long


one


subject


and


10-min


long


the


other


Frequency


data


were


collected


1-min


interval


A reversal


rie filnn


nmnl nvuri


wi th


di ffrsnt


ricimztnris


I t


hnth


fillh~ P~t c:


WF~ S















a decrease in

reinforcement


decrease


a targeted

aggressive


aggressive

contingency


in aggressive


non-aggressive


responses.


responses


correct


responses


response

increase


both


subjects.


responses


escape


resulted


in aggressive


resulted


contingency


a decrease

e responses


occurred


with


contingent


esc


from


the


demand


situation.


Extinction


through


removal


the


escape


contingency


resulted


a decrease


in aggressive


responses


They


concluded


that


aggre


ssive


reinforced


responses


escape


from


of both


demand


subjects


were


situations,


negatively


and


they


designed


treatment


conditions


based


on their


results.


Iwata


et al.


(198


also


attempted


to identify


functional


relations


between


self


-injury


and


specific


environmental


series


events


of standard


manipulated


through


1:


conditions.

conditions


,rief,

The

were


repeated


exposures


environmental

availability o


variable


f play


material


, attention


, and


demands.


Nine


subjects


were


exposed


to 4


experimental


conditions


in a multielement


design.


An observer


recorded


the


occurrence


and


nonoccurrence


of self-injurious


behavior


continuous,


10-s


interval


The


dependent


variable


was


the


percentage


interval


during


which


one


or more


self-injurious


responses


crnrct -o


WP~P












staff


attention


to self-injurious


behavior


The


experimenter


did


not


interrupt


the


SIB


, although


physical


contact


was


paired


with


social


approval


The


academic


demand


condition


was


arranged


test


negative


reinforcement


removal


of demand


situations


when


SIB

were


occurred.

presented


During


thi


subject


condition,


every


educational


activities


Contingent


upon


the


experimenter


turned


away


from


the


subject


the


trial


was


terminated.


A change-over


delay


was


implemented


each


repeated


SIB.


The


alone


condition


tested


the


effects


impoverished


environment


The


subject


was


left


alone


room


with


observer


behind


a one-way


mirror


another


room.


The


unstructured


play


condition


was


arranged


as a


control


the


presence


the


experimenter,


lack


attention


SIB,


absence


of demands


, and


the


availability


potentially


stimulating


materials.


The


contingencies


arranged


this


condition


were


the


opposite


the


demand


social


disapproval


conditions.


Iwata


et al. (1982)


were


able


identify


functional


relations


U' I


between


specific


Q srih-iants~


environmental


TPhe


functional


conditions


relations


and


were


.


, r r


* .I -












these


subjects


conditions


were


not


discriminable,


behavior


was


a function


variables


not


controlled


the


study,


or that


the


behavior


was


under


the


control


of multiple


variables


(attention


, escape


from


demands,


and


stimulation


barren


environments)


In contrast


the


Carr


et al.


(1980)


study,


Iwata


al. (198


did


base


their


analogue


conditions


observations


the


subjects


' environment;


they


developed


conditions


test


several


possible


sources


reinforcement.


An advantage


the


standard


conditions


was


that


they


allowed


experimenters


test


more


than


one


controlling


variable,


whereas


specifically


designed


conditions


controlling


of Carr


variable.


allowed


A di


testing


advantage


of only


is that


one


observations


from


the


natural


environment


in Carr


may


have


provided


more


information


with


respect


behavior


subjects


whom


there


was


no discernable


pattern


responding


analogue


conditions.


Mace


and


Knight


(1986)


provided


another


example


experimental


analysis


involving


special


arrangements


consequences


behavior.


They


performed


a functional


analysis


the


pica


exhibited


a profoundly


retarded


U-~~~~ uI.Y~n in ne -


TlTara


et mff


t ~~ Y; ~ ~~1 h A


m a rr ; n~ r 1 ~ Ca


r


e~(


n-


Iiirr A


I 1












treatment


intervention


was


implemented


a day


area


with


other


clients.


Sessions


were


15 minutes


long,


and


a 10-s,


partial-interval


system


was


employed


to collect


data.


The


pica


assessment


response


baseline


or attempted


contingency


response


was


immediate


to follow

ly with a


each


mild


reprimand


and


removal


object


from


the


subject'


mouth.


The


effects


varying


rates


interaction


were


assessed


conditions


when


objects


were


removed


from


subject'


mouth


without


a reprimand.


The


amount


social


interaction


was


manipulated


four


conditions


noncontingent


interaction,


experimenter


delivered


a prompt


to complete


the


task


according


to a variable


time


(VT)


8-min


schedule,


frequent


interaction


with


continuous


eye


contact


and


task


instructions


on a VT 15-s


schedule,


limited


interaction


which


the


experimenter


faced


away


from


the


subject


and


interacted


with


respect


the


task


only


on a FT 3-min


schedule,


interaction


which


experimenter


faced


away


from


subject


at a different


table.


The


effects


helmet


were


analyzed


three


separate


conditions,


within


limited-interaction


contingency:


helmet


face


shield


helmet


only,


and


no helmet.


The


percentage


intervals


with


pica


was


related


to the


rate


nv4orm+-1 I fl .


nt r rflti n f


nrndunaed


hiah


r~~pa


I11 l


iu












the


fewest


percent


intervals


pica.


Finally,


the


experimenters


designed


a treatment


consisting


limited


therapist


interaction


with


the


subject


without


a helmet.


treatment


reduced


the


percentage


intervals


with


pica.


The


study


demonstrates


an interesting


parametric


analysis


variables.


possible


The


antecedent


subject'


behavior


consequent

systematical


controlling

ly changed


with


the


change


in contingencies,


indicating


a functional


relationship


between


the


variables


manipulated


and


pica.


The


rate


the


subject'


behavior


was


lowest


the


frequent


interaction


condition


highest


interaction


condition;


the


helmet


with


the


faceshield


produced


the


highest


rate


of pica


within


the


helmet


manipulations.


explanation


was


provided


including


limited


interaction


rather


than


frequent


interaction


treatment.


Mace


and


Knight


combined


aspects


naturalistic


observation


method


and


collected


data


subject'


home


environment.


These


corroborating


data


from


natural


environment


provide


support


the


conclusions


from


environmental


manipulations.


They


found


rates


of pica


when


there


was


interaction


with


the


subject


without


the


helmet.


They


further


found


that


staff


interacted


less


frequently


with


the


eliH4 antT.T o


no tanro ma


h~1mnt


anti


fare~F


shield -


whan












rocking,


head


rolling


, and


finger


manipulations.


The


contingencies


implemented


each


the


four


conditions


were


the


same


Iwata


et al


alone


, unstructured


play,


demand


, and


social


disapproval.


A multielement


design


was


employed.


subjects,


they


found


that


the


rate


stereotypy


was


highest


alone


condition,


and


conditions


which


any


distractions


were


available


demand,


attention,


play)


The


third


subject


exhibited


essentially


undifferentiated


rates


of stereotypy


in all


conditions.


The


analysis


performed


Sturmey


et al. (1988)


differed


from


Iwata


et al. (1982)


several


ways.


First


, they


used


momentary


time


sampling


to record


occurrence


or nonoccurrence


every


They


proposed


that


this


measure


would


give


unbiased


estimate


proportion


time


spent


behaving


Second,

15-min


Sturmey

sessions


et al. had


Iwata


10-min


et al


sessions


Thi


rather


decreased


than


time


the


period


may


have


reduced


subj ects'


exposure


the


experimental


contingencies


enough


cause


the


undifferentiated


effects


obtained


Sturmey


et al.


The


final


difference


was


that


only


one


experimenter


conducted


the


sessions


in each


condition.


may


have


decreased


discriminability


ni f4 w.P4n+ i at *i nn


roacn 1 +oe


a n -r


each


|


nhnri i C i nn e


Ckn


nnf


uss;


*


I 1 |












consisted


five


conditions:


baseline,


attention,


escape,


tangible,


and


unstructured.


Subj ects


were


exposed


to each


condition


three,


10-min


sess


ions.


The


baseline


condition


consisted


requiring


the


subjects


to perform


an "easy"


task


(matching


sample


items


from


an intelligence


test,


the


Leiter


Scale).


The


task


was


determined


as easy


because


only


items


that


subject


was


able


to match


with


100%


accuracy


were


included.


and


Verbal


attention


praise


form


contingent


of either


upon


correct


a demand,


responses


praise,


or some


other


neutral


comment


were


provided


every


In addition,


contingent


on correct


responses,


tangible


items


were


provided


on a variable


ratio


(VR)


schedule.


There


were


consequences


or other


inappropriate


behaviors.


This


was


a control


condition;


was


hypothesized


that


no aversive


stimuli


were


present


and


that


socially


motivated


SIB


would


due


high


density


attention.


The


attention


condition


involved


the


same


contingencies


as the


baseline


condition


with


the


exception


that


instead


providing


attention


100%


the


intervals


attention


was


provided


only


interval


This


condition


was


designed


to test


adult


attention


as a maintaining


variable


SIB.


The


escane


condition


involved


.I l


same


continaenois












items.


This


condition


was


designed


test


negative


reinforcement


as the


maintaining


variable


SIB.


The


tangible


condition


involved


providing


each


subject'


favorite


tangibles


every


ninth


correct


answer,


while


the


items


were


visible


constantly.


This


condition


was


designed


test


tangible


reinforcement


as the


maintaining


variable


SIB.


The


unstructured


condition


involved


placing


the


subjects'


favorite


foods


toys


within


reach,


having


adult


interacting


with


subject,


and


having


work


materials


available


test


sensory


subject.


This


consequences


condition


as the


was


maintaining


designed


variable


SIB.


Durand


and


Crimmins


(1988)


found


that


each


the


subjects


exhibited


a high


rate


of responding


one


the


experimental


conditions.


Two


the


subjects


had


higher


rates


of responding


unstructured


condition,


had


high


rates


in the


attention


condition.


Two


other


subjects


had


higher


rates


in the


escape


condition,


two


had


high


rates


the


tangible


condition.


There


were


significant


problems


with


the


experimental


conditions


Durand


and


Crimmins


(1988)


study,


as well


r- rtv + nrac nn


rlrmtan


frnm


nlr nri


anti


'r immins


rr; ~k


th pm


1-nn












contingent


upon


SIB.


That


in the


attention


condition,


attention


was


delivered


contingently,


were


the


reinforcing


stimulus


maintaining


the


SIB


Instead


attention


was


delivered


every


, regard


ess


subject'


behavior.


In such


an arrangement


one


might


expect


to reinforce


accidentally;


however,


there


would


arrangement


that


would


determine


SIB


was


sensitive


attention


as a reinforced


reinforcers


the


other


conditions


were


similarly


arranged


on a fixed


-time


schedule


without


a response


requirement.


Therefore,


instead


the


high


rates


that


Durand


and


Crimmins


predicted,


one


might


argue


that


the


frequency


would


be depres


because


reinforcers


were


delivered


regularly


without


requirement.


Because


stimuli


were


delivered


contingent


on SIB,


they


were


not


demonstrated


to be


reinforcers


or maintaining


variable


SIB.


conditions

obtained.


have


room


there

The D


resulted


served


are


urand


high


alternate


and


explanations


Crimmins


rates


as discriminative


the


unstructured


because


stimuli


the


either


results


condition


people


that


SIB


may


the


would


be reinforced


with


escape


or by


attention


from


staff


contrast


, the


Iwata


et al


(198


"alone"


condition


placed


a rnnm


IN ~~wt f- II S N


nlav


nhi 0rflts


ann


with


I l|


sllhi cl~t`


rJ | | u


*












conditions, t

environmental


therefore,


presented


contingencies


and


almost


opposite


a comparison


of responding


during


each


condition


would


provide


additional


information


regarding


sensory


maintained


SIB.


Steege,


Wacker


, Berg


, Cigrand,


and


Cooper


(1989)


assessed


potential


reinforcers


and


variables


maintaining


subjects.


Their


study


was


an extension


Iwata


et al.


(1982)


assessment


that


a stimulus


procedure,


preference


was


reinforcers


included


identified


the

were


then


arranged


contingent


upon


adaptive


behavior


during


treatment.


was


determined


interviews


with


the


classroom


teacher


that


(handmouthing


and


biting)


occurred


most


frequently


when


the


first


subject


was


alone.


The


assessment


of maintaining


variables


included


three


conditions.


two


the


conditions


there


was


no contingency


and


the


subject


was


left


alone


except


observer.


The


third


condition


was


a combination


Iwata


et al.


demand


and


attention


conditions.


Data


were


recorded


during


intervals,

considered


sessions


first


lasted


three


15 min.


sessions


Steege


each


et al.


the


(1989)


conditions


as the


assessment.


occur


during


the


Aawv~~~~~~~~~wA~~~~ I ~44a 4 4nnv nf fll% 4 4-4l nn-nl n n Ira 42r nr'imt


~ Am~ n rt I a CC nnC i nn


nhrrrj: C: hn


nn~rrrrra~


unn












maintained


across


conditions.


Steege


et al.


concluded


that


was


maintained


as a self-stimulatory


behavior


The


assessment


second


subject


more


closely


approximated


Iwata


al. (198


assessment.


Three


the


four


conditions--alone


, demand


unstructured


play--were


included,


and


a response


cost


condition


was


added.


The


response


cost


condition


was


said


to approximate


conditions


natural


environment


that


were


reported


the


parents


result

room.


high

the


rates

subject


of SIB.

played


Toy

with


were


a toy


available


the


the


experimenter


removed


the


toy


scolded


the


subject


playing


with


Steege


et al


. (1989)


arranged


subject


be exposed


these


conditions


briefly


10-min


sess


ions


of each


condition)


None


the


data


from


the


assessment


were


presented


study,


although


Steege


reported


that


occurred


exclusively


during


the


demand


condition.


They


concluded


that


the


subject'


SIB


was


maintained


negative


reinforcement


designed


a treatment


which


resulted


the


subject


being


redirected


task,


correct


responses


produced


an identified


reinforcer


The


was


reduced


within


the


first


three


sessions


after


treatment


was


instituted


Xii~a rIra It~1 -


bri 0f


sYnnpsiire


nr ntt r n an


ttanlrar


.I


alEn












maintaining


variables


were


identified


as positive


reinforcement


subject,


automatic


reinforcement


another


subject.


subject,


Although


negative


appeared


reinforcement


that


third


responding


subjects


was


under


control


contingencies


arranged


the


various


conditions,


the


brief


exposure


rai


ses


the


possibility


a transient


trend


that


would


have


changed


had


subject


been


exposed


the


conditions


longer


periods


time.


Steege


et al.


(1990)


also


implemented


brief


functional


analyses


determine


that


the


of 2


subjects


was


maintained


negative


reinforcement.


Subjects


were


exposed


briefly


four


10-min


Iwata


sessions


et al.


and


10-min


conditions.


They


sess


ions)


noted


to each


that


several


sessions


were


terminated


early


and


so lasted


less


than


min.


Even


though


analogue


asses


sments


allow


the


identification


of functional


relations,


might


possible


argue


that


these


prosthetic


assessment


situations


provided


"training"


subject


of SIB


demand


is responsive


the


or attention


contingencies


conditions.


provided,


attention


or escape


from


a task


, then


presenting


stimuli


r'nr4 ea1iS CTnn+nHr4


tt; it


nk~ki liCtr


n9 ~ri~rrrn


Cka nr


nn CtP


r nnra~Ed












variables


are


controlled,


the


reinforcers


available


are


contingent


on SIB.


However,


the


subject


responds


quickly


the


contingencies


and


reports


from


the


subject'


normal


environment


support


that


the


serves


an escape


function,


unlikely


that


the


assessment


develops


"new"


reinforcement


contingencies


the


SIB.


In addition,


behavior


were


come


under


control


new


reinforcement


contingencies


, one


might


expect


rate


of SIB


increase


over


the


course


the


assessment


forming


an acquis


ition


curve.


This


acqui


sition


curve


would


serve


as an


indicator


that


the


had


been


"trained"


rather


than


a maintaining


variable


identified.


summary


, there


are


three


general


methods


assessment:


natural


indirect


environment,


information


and


gathering,


experimental


observation


manipulation


possible


controlling


variables.


often


assumed


that


data


obtained


from


the


different


assessment


techniques


are


correlated,


although


experimental


support


this


assumption


generally


lacking


However


, one


comparison


an indirect


method


of assessment


and


direct


measurement


of subjects'


behavior


assessed


was


the


conducted


validity


Durand


the


MAS


Crimmins


(1988)


identifying


They


the


-.~~~~~~rrr~~ a: tA- a -


eiih4 an+ a


ci ort oei


fr nm


LL LI 1


)I


C' Iv rnr


e


Y












functional


analysis


was


similar


to but


briefer


than


that


arranged


Iwata


(1982)


Durand


Crimmins


(1988)


concluded


that


MAS


provided


a valid


prediction


of results


the


analogue


conditions;


however,


a close


examination


their


data


indicates


that


eight


subjects


the


experiment,


the


results


four


were


inconclusive.


One


subj ect


had


ess


than


intervals


scored


the


highest


condition,


and


another


subject


had


little


difference


between


the


highest


category


and


next


highest


category


in the


percent


intervals


scored


Two


the


subjects


only


exhibited


during


demand


condition


with


percent


interval


scored.


Several


analogue


studi


described


(Sturmey


al.,


1988;


Durand


& Crimmins,


1988;


and


Steege


et al


, 1989)


exposed


subjects


to the


asses


sment


conditions


either


very


sess


ions


, very


brief


sessions,


or both


few


and


brief


sess


ions


The


assessments


were


time-referenced


rather


than


following


a criterion


of stability


of responding


(Sidman,


1960)


The


validity


these


brief-exposure


assessments


has


not


been


determined


and


could


assessed


through


comparison


to behavior-referenced


assessment.


The


present


study


compares


three


assessment


techniques


- 4 n WA 4-aa1 4. 4ta I --1 -n UIf


Ulr


~u: mm: nn


rnaa\


FICll~~t I


~AJL~~


nl I w ~r rr ~


-


r












determine


the


validity


the


MAS


as a predictor


of results


from


the


analogue


analyses.


more


important


comparison


the r

analy


results


of a time-referenced


conditions


and


exposure


a behavior-referen


to functional

ced exposure,


allowing


determination


validity


the


time-referenced


assessment.


The


specific


questions


asked


this


study


are:


the


MAS


a valid


and


useful


tool


determining


the


maintaining


variables


of a subject'


SIB?


and


a time-


referenced


useful


exposure


indicator


the


the


analogue


maintaining


conditions


variables


a valid


of a subject'


SIB?


Two


experiments


were


performed


order


answer


these


questions.


Experiment


compared


results


the


MAS


the


results


of a time-referenced


analogue


assessment.


Because


the


numerous


procedural


problems


previously


specified


and


, therefore,


the


limited


possible


inferences


from


the


Durand


and


Crimmins


(1988)


study,


the


conditions


time-referenced


assessment


had


contingencies


arranged


Iwata


et al.


(1982)


study.


However,


the


MAS


compared


to a time-referenced


behavior-referenced


assessment,


Assessment,

to provide


rather

a more


than


direct


replication


the


Durand


Crimmins


study


In addition,


time-referenced


ass


essment


arrangement


a combination


nf all


nreviousl


described


limited


assessments


(Steeae


,,


.












indicator


of maintaining


variable


identified


behavior-referenced


assessment.


Experiment


compared


the


results


a time-referenced


to a behavior-referenced


assessment


order


to determine


the


time-referenced


procedure


a valid


assessment


methodology












General


Method


Subiects


Subjects


were


drawn


from


a pool


of residents


of a state


institution


the


mentally


retarded


who


had


been


referred


assessment


and/or


treatment


SIB.


subjects


were


reported


to be


functioning


the


severe


to profound


range


mental


retardation


and


had


limited


communication


and


self-


care


repertoires.


Human


Subjects


Protection


In order


assess


differential


effects


the


environment


on self-


injury


, the


study


required


that


subjects


allowed


engage


minor


self-injurious


behavior


brief


periods


time.


procedures


were


reviewed


and


approved


an institutional


review


board,


as well


as the


institutional


behavioral


program


review


committee


and


medical


staff.


The


following


safeguards


were


employed


to reduce


risk


physical


injury


as a result


SIB


exhibited


during


assessment.


First,


each


subject


received


a complete


medical


examination


a physi


cian


, which


assessed


current


physical


status,


ruled


organic


causes


the


SIB,


and


established


a session


termination


criterion


based


on the


severity


the


behavior.


Phy


sicians


occasionally


observed


aacn1 ari


*.~~i *u1 k-i-t I, L


nrt aor


it occurred.


;I C: EP fi ~


rA -


I l l


.


.^ .


, N


u












restraint.


Further


sessions


were


postponed


the


SIB


resulted


an open


wound.


During


the


course


of Experiment


, session


termination


and/or


suspension


only


occurred


one


time


during


time-referenced


assessment


subjects,


and


two


times


with


subject


during


behavior-referenced


assessments.


Assessment


Conditions


Motivational


Assessment


Scale


(NASI.


The


MAS


(Durand


Crimmins,


1988)


consists


16 questions


about


events


surrounding


self-injurious


behavior


Durand


and


Crimmins


identified


four


categories


of reinforcement:


positive


reinforcement


the


form


adult


attention,


positive


reinforcement


form


of a tangible


item,


escape


from


demand


situations,


and


sensory


reinforcement.


Each


category


represented


four


questions


on the


MAS.


The


answers,


indicating


how


often


occurs


under


the


circumstances


described,


are


specified


on a six-point,


Likert-type


scale


where


= never


, (1)


= almost


never,


= seldom,


= half


time,


= usually


, (5)


= almost


always,


and


= always


The


numbers


associated


with


each


answer


each


four


questions


related


to a category


were


then


added,


and


an average


rating


each


category


was


mb~ra


s.. Ja


nnhmnt *1 4+ ai flt nni E


hi nhocFt-


aAVPriarns


i ni i iatedi


~~ + annrtt


w


ri io












responders


present


study


cons


isted


cottage


supervisors


and


direct-care


staff


, instead


of classroom


teachers.


questionnaire


To avoid


and


problems


possible


such


reading


as timely


return


difficulties


the


staff


, the


questions


were


read


to each


staff


person


a graduate


research


assistant.


The


MAS


was


administered


in as private


setting


as possible,


only


prompts


responders


were


to repeat


the


questions


repeat


the


possible


answers.


The


staff


person


answering


questions


had


the


possible


answers


front


of him/her


on a typed


card,


and


administrator


questionnaire


read


the


question


and


asked


the


staff


to provide


one


the


possible


answers,


ranging


from


zero


("never")


to six


("always


When


a staff


per


son


had


difficulty


answering


a question


the


research


assistant


repeated


question


asked


the


staff


person


give


the


best


answer.


the


staff


person'


reply


did


not


coincide


with


one


possible


answers,


the


administrator


asked


which


six


answers


came


closest


s/her


response.


Time-referenced


assessment.


Three


sets


of contingencies


(demand,


attention,


and


alone)


were


effect


during


the


time-referenced


assessment.


session


consisted


exposure


to the


conti naennai s


associated


with


one


the


conditions.


,,


a


CM












until


three


conditions


had


been


presented.


The


time-


referenced


assessment


was


completed


a private


room


on the


subject'


home


cottage.


demand


condition,


therapist,


subject,


and


observers)


were


present.


Academic


or self-care


materials


were


present


the


room.


Every


30 s the


therapist


presented


an instruction


or task


to the


subject.


Contingent


upon


SIB,


instruction


delivered


or task


contingent


was


terminated


on compliance


with


the


Praise


was


instruction.


other


allow


behaviors


escape


were


from


ignored.


demand


This


condition


contingent


upon


was


SIB;


designed


escape


functioned


as a reinforced,


rate


of SIB


was


expected


higher


this


condition


than


other


conditions.


In the

observers)

potentially


attention


were


condition,


present.


stimulating


items


the


toys

were


subject,


work


present.


therapist,


materials,


or other


Contingent


upon


SIB,


therapist


verbally


attended


to the


subject,


providing


statements


of social


disapproval


or concern.


addition,


attention


consisted


response


blocking


s or


otherwise


touching


subject.


All


other


behaviors


subject,


appropriate


or inappropriate,


were


ignored.


This


condition


was


designed


so that


only


contingency


was


ni-i ont i nfl


* *Ir


qTR~


R*$tnt nflin


fnnnti nnad


aR a rei nfnorcer.


*












the


alone


condition,


subject


was


alone


the


experimental


room


with


no materials


available.


Although


least

with


one

the


observe

subject


was


or with


present,


any


other


observer

observer


did

who


not

might


interact

have


been


present.


eye


contact


was


made


between


the


subject


the


observer


observer


moved


the


away


subject


without


approached


obviously


the


reacting


observer,


any


other


way


the


approach


the


subject


and


attempted


maintain


a distance


of at least


feet.


thi


condition


there


were


no contingencies


SIB


or any


other


behavior


there


were


no distracting


or stimulating


items


available.


was


expected


that


was


automatically


reinforced


, the


rate


would


be highest


condition.


Behavior-referenced


assessment.


The


behavior-referenced


assessment


was


comprised


of four


conditions:


alone,


demand,


attention


and


play.


The


behavior-referenced


asses


sment


was


completed


in a cottage


established


the


treatment


residents


who


engaged


SIB.


There


were


never


more


than


other


subjects


present


on the


cottage


any


one


time,


subjects


spent


less


than


two


hours


per


day


on the


cottage.


Session


length


was


15 min


across


conditions.


least


two


sess


ions


and


as many


four


were


run


each


day


Sessions


sa0flhIpflflr


I .I


a random


order


with


restrict


on that


WPrP


_-**












The


contingencies


implemented


the


demand,


attention


and


alone


conditions


were


the


same


as described


the


time-


referenced


assessment.


play


condition,


various


toys


were


present


the


room


with


the


subject


and


observer(


therapist


approached


the


subject


and


either


presented


a toy


interacted


with


the


subject


on a 30-s,


fixed-time


schedule.


response


occurred


was


ignored


within


s of


and


a self-injurious


of a fixed-time


interval,


therapist


delayed


delivery


of attention


s to avoid


reinforcing


the


SIB.


This


condition


was


designed


to control


effects


of having


staff


and


materials


the


room,


to provide


consequences


appropriate


behavior.


was


expected


to have


the


lowest


rate


this


condition.


Figure


1 presents


conditions


and


controlling


variables


that


correspond


each


the


three


assessments.


example,


MAS


conditions


of Attention


and


Tangible


correspond


Attention


condition


both


the


time-


referenced


controlling


and


behavior-referenced


variable


these


assessments,


corresponding


the


conditions


environmentally


located


positive


reinforcement.


Measurement


Refinnonse


definitions


direct


observation.


Table


uA RL


r


























Figure 1.
variables
assessment
correspond
assessment


The
hat
ar
ng
had


assessment
orrespond
presented
nditions.
control


conditions and controlling
for each of the three
The arrows connect
Only the behavior-referenced
condition.
















CONTROLLING VARIABLE


Positive


Reinforcement


Env Negative
Reinforcement


Automatic


Reinforcement


Control


MAS


Attention


Tangible


Escape


Sensory


Referenced


Behavior-
Referenced


Attention


Attention


Escape


Escape


Alone


Alone


Play


TZmel











Table


Observer


definitions


of subject'


self-iniury


Response


Definition


Number


of Subjects


Armn/hand


banging


Forceful


with


contact


a stationary


the


arm/hand


object


Face


slapping


Forceful


hand


with


contact


the


the


open


face


Hair


pulling


Closure


thumb
motion


on hair


away


fingers


with


from


and


a pulling
e head


Hand


mouthing


Insertion


fingers
touching


into
any


one


mouth,
part c


or more


or tongue
>f hand


Head


banging


Forceful


with


contact


a stationary


the


head


object


Self-biting


Closure


teeth


th


on the


e upper
flesh o


and


any


lower
portion


body


Self-kicking


Forceful
portion


contact


of a foot


any


body


Self-scratching


Gouging
across


movement


skin,


leaving


fingernails


red


mark


wound


Throat-poking


throat


Forceful
or hand


with


finger














During


SIB,


each


compliance,


session,


disruption


an observer


, and


recorded


aggression


occurrences


in continuous


intervals.


Occurrences


significant


experimenter


responses


instruction


or attention)


were


recorded


well.


Observers


recorded


responses


on a hand-held


computer


(Assistant


Model


AST102)


The


dependent


variable


interest


was


the


number


responses


per


minute


each


session.


The


mean


number


responses


minute


was


then


computed


each


condition


the


time


-referenced


behavior-referenced


assessment.


Experimenters


, observers


I MAS


raters


Experimenters


, therapists)


were


eight


graduate


research


assistants.


Each

the


experimenter

assessment co


was


provided


nditions


and


with

served


a written


description


as a reliable


observer


before c

training


conducting

activity


formal


were


sessions.

employed,


In addition

including de


,specific


monstrating


experimental


protocol


informal


sess


ions,


insure


that


an experimenter


would


respond


appropriately


during


each


assessment


condition.


Both


primary


were


graduate


and


enroll


and


reliability

d in or had


undergraduate

observers.

completed a


students


served


Undergraduate


course


students


behavior


...












collected


data


in sessions


with


reliable


observers


until


or better


interobserver


reliability


was


obtained


on three


sess


ions


each


subject.


Each


subject'


was


rated


on the


MAS


two


staff


who


had


worked


with


subject


extensively


(from


several


months


to several


years)


One


supervisor


and


one


direct-care


staff


were


selected


each


subject.


A total


staff


participated


as responders


subjects.


Reliability


Inter-rater


reliability


MAS


was


computed


using


two


methods


point-to-point


agreement


each


question,


and


agreement


on the


category


rankings


each


the


respondents.


Point-to-point


reliability


was


calculated


dividing


the


number


agreements


on the


exact


answer


each


question


over


total


number


of agreements


and


disagreements.


range


inter-rater


reliability


was


from


to 63%


agreement,


with


a mean


of 23


Agreement


highest


ranked


categories


ranged


from


to 100%.


Seventeen


thirty-four


pairs


of raters


agreed


with


respect


to the


highest


ranked


category


No pair


of responders


agreed


on the


rankings


every


category


any


subject.


There


was


special


training


raters,


and


reliability


could


i rnrntioA


rira


4-r *ha


r' 4 lr *1I*u i*u


oC* i r-mnn n i v


n~tilra












first


dividing


the


observation


session


into


consecutive


10-s


intervals.


each


interval,


smaller


rate


was


divided


larger


rate.


These


quotients


were


then


averaged


sess


and


the


total


was


multiplied


100.


Interobserver


reliability


data


were


collected


the


total


number


time-referenced


assessment


sess


ions


subjects.


The


mean


reliability


percentage


was


time-referenced


assessments


Interobserver


reliability


was


collected


the


total


behavior-referenced


assessment


sess


ions.


The


average


percentage


interobserver


agreement


behavior-referenced


assessment


was












Experiment 1


Subjects


Thirty-four


subjects


males


and


females)


, ranging


age


from


to 45


years,


partic


ipated


the


MAS


the


time-referenced


asses


sment


comparison.


Experimental 1


Design


The


MAS


was


administered,


prior


to beginning


the


time-


referenced


assessment,


to two


staff


persons


familiar


with


each


subject.


subj ects


were


exposed


to each


the


three


different


conditions


time-referenced


assessment


in a multielement


experimental


design


Sidman,


1960).


The


order


of presentation


of conditions


each


subject


was


random


with


restriction


that


a condition


could


not


be repeated


unless


each


conditions


had


been


presented


same


number


times.


Results


Figures


each


subject'


show


MAS,


the


mean


mean


ratings


across


responses


per


categories


minute


time-referenced


assessment


conditions.


The


corresponding


points


each


assessment


are


connected


vertical


lines,


indicating


that


MAS


category


time-referenced


assessment


condition


relate


to similar


variables


controlling


a a


- a t2 I U I.. a .4.4 -: C1 a a a a nn a A n rr n r 4 fn n rAn


nh nhnrl:~: hn


cl A FI A A rrmAne


k ~ rl


m.


Ck L' U 111AA












high,


with


means


ranging


from


to 4


The


attention


category


had


the


highest


mean


rating


The


time-


referenced


assessment


results


indicate


that


the


highest


mean


rate


of SIB


occurred


attention


condition


responses


per


minute).


The


other


conditions


had


the


same


rate


of SIB,


responses


per


minute.


Fourteen


subjects


had


highest


mean


rating


the


MAS


the


automatic


reinforcement


category


(sensory


stimulation).


Twenty-two


subjects'


mean


MAS


ratings


were


essentially


undifferentiated


across


conditions


least


three


the


four


ratings


within


point


of each


other).


Comparison


the


MAS


ratings


and


the


time-


referenced


assessment


reveals


that


8 of


the


.5%),


MAS


ratings


matched


the


results


obtained


the


time-


referenced


asses


sment


(i.e.,


the


highest


mean


rating


MAS


corresponded


condition


with


highest


mean


rate


SIB


the


time-referenced


assessment).


The


subjects


whom


results


of both


assessments


matched


were:


S12,


S18,


S23,


S25.


.e.6



























Figur
respo
condi
are p
0 to
the r
by ea


e 2
nse
tio
res
6.
igh
ch


s p
ns
ent
Th
t y
sub


Mean ratings f
er minute for
are presented
ed on the left
e mean response
-axis, with th
ject's rate of


the M
time
Subj
axis,
per m
ange
spond


AS
-re
ect
wi
inu
ind
ing


categories
ferenced a
s 1 17.
th scores
te are pre
ividually
.


and mean
assessment
MAS rating
ranging fr
sented on
determined













MEAN


RAT


a (i


a CR


S. CI


Escape


Attention

Tangible

Automatic


Escape

Attention

Tangible

Automatic


0
0 *

--- --- --o


-0
a


)L~p


0 0 0
b a N)


Escape

Attention

Tangible

Automatic


0-b


a c o


Escape

Attention

Tangible

Automatic



Escape

Attention

Tangible

Automatic



Escape


I~L
I a
0 NiOQ ) a N)
I I I


C-----0


I I I I

I LI I I I


II I |I




-a0
- o


CI


b-0


r


r


*


























Figure 3.
responses
condition
ratings a
ranging f
presented
determine


Mean
per mi
s are p
re press
rom 0 t
on the
d by ea


ratings for
nute for the
resented for
ented on the
o 6. The me
right y-axi
ch subject's


-S


the MAS categories and mean
time-referenced assessment
Subjects 18 34. MAS
left y-axis, with scores
an responses per minute are
s, with the range individually
rate of responding.














MEAN


RAT


A II


Escape

Attention

Tangible

Automatic


Escape

Attention

Tangible

Automatic


Escape

Attention

Tangible

Automatic


Escape

Attention

Tangible

Automatic


0-,


0-


I p
9- Nr h

I I


0-f


SI I I I I





0






oI r
hO
---- ----
b0P


0 0 0
0 -* t


-O


Escape

Attention

Tangible

Automatic


Escape


I-a


I-0


a


I












Spearman


rank


order


correlations


based


on ranked


means


the


MAS


ratings


and


ranked


mean


rates


SIB


the


Demand/Escape,


Attention


Sensory/Alone


conditions


were


completed


subjects.


The


Tangible


condition


MAS


was


not


included


correlations


because


there


was


directly


corresponding


condition


the


time-referenced


assessment.


The


range


was


-0.87 to 1.0


with


a mean


of 0.09


Discussion


Thi


experiment


was


a systematic


replication


the


Durand


and


Crimmins


(1988


comparison


between


a time-


referenced


analogue


asses


sment


indirect


assessment


technique


(the


MAS)


The


data


indicated


that


the


MAS


an accurate


predictor


the


results


an analogue


assessment;


that


, there


was


poor


correspondence


between


MAS


ratings


and


rates


during


analogue


conditions


the


direct


assessments.


Durand


& Crimmins


presented


MAS


as an alternative


method


identifying


variables


controlling


SIB.


They


described


two


major


drawbacks


analogue


method


functional


analy


The


analogue


assessment


requires


extensive


assessment


staff


training


requires


to implement;


an extensive


amount


The


time.


analogue


Although


- -- U t' L~As -I I .. Jaa a fa a S raIn l


,I L


L1- Il~n


rr AY~ r nl r ~


CIA1~AY~


Ml)*Lfi~l ~*LLCI


Ltrnnn


--- -


~AI*A


1












inter-rater


reliability


results


obtained


in the


present


study


and


a prior


study


(Zarcone


et al


in press)


Second


results


the


current


study


were


poor


predictors


results


an analogue


assessment.


Therefore,


validity


the


instrument


identification


of controlling


variables


of SIB


is questionable.


The


poor


interrater


validity


reliability


obtained


may


the


have


current


been


a factor


study


in the


It should


be noted


that


these


results


may


indicative


the


validity


the


MAS


given


better


interrater


reliability.


An additional


point


should


made


with


respect


the


Durand


their


& Crimmins


analogue


study.


comparison


Rather


than


randomly


selecting


, they


selected


subjects


subjects


with


high


ratings


in a single


category


each


the


four


MAS


category


es.


selection


should


have


insured


that


the


controlling


variable


would


be easily


discovered


with


analogue


assessment;


however,


the


subjects


selected

analogue


Durand


conditions


Crimmins,


occurred


differentiation


only


among


subjects


Experiment


Subjects


Eight


subjects


mal


females


participated


at


- - --- - a a.. a S -A na-a 4 -nl -


1~A S~;Ar


k *


rl


,~,,,,,, II


,,,1


..


~~c~LkA


~L


rr












high


rate


or high


degree


severity


their


SIB


targeted


them


as priority


candidates


treatment.


Experimental


Design


The


time-referenced


assessment


(from


Experiment


was


completed


prior


to beginning


the


behavior-referenced


assessment


each


subject.


6 of


the


subjects


, the


four


experimental


conditions


were


presented


a multielement


design.


The


order


of presentation


of conditions


each


subject


was


sequenced


in a random


order


with


the


restriction


that


one


condition


could


not


be repeated


until


three


conditions


had


been


presented.


Three


to four


sess


ions


occurred


each


day


The


behavior-referenced


assessment


was


effect


until


differential


responding


occurred,


or until


least


sessions


of each


condition


had


been


completed


the


remaining


subjects


, (19


and


conditions


demand,


design.


attention,


The


play


sequence


were


presented


presentation


in an A-B


of conditions


differed


these


subjects,


and


each


condition


was


continued


until


there


was


some


stability


in responding


from


sess


sess


ion,


or for


at least


eight


sess


ions


. Stability


was


determined


based


on eyeballing


rates


per


minute


from


individual


sess


ion.


fIant I &r.












each


condition


comparison


are

both


presented


in order


assessments


to allow

The i


a concise


individual


session


data


each


assessment


would


have


resulted


comparison


sess


ions


the


time-referenced


assessment


and


several


subjects,


sessions


the


behavior-referenced


assessment.


Five


subjects


had


their


highest


mean


rates


responding


same


conditions


the


time-referenced


behavior-referenced


assessment.


These


subjects


are


indicated


an asterisk


above


highest


conditions


on each


graph


(see


Figure


The


play


condition


resulted


the


lowest


mean


rate


of SIB


two


subjects


and


whom


a play


condition


was


implemented.


Subject


attention


exhibited


condition


SIB


of both


most


frequently


assessments,


during

mean r


the


ates


responding


the


alone


demand


conditions


of both


asses


sments.


The


rate


of responding


during


play


session


was


slightly


higher


than


either


the


alone


or demand


conditions.


Subject


15 had


highest


mean


rate


of SIB


the


attention


conditions


of both


time-referenced


and


behavior-referenced


assessments.


There


were


rates


nnnA 4 4 ana


ACk hr












the


attention,


alone,


play


conditions


the


behavior-


referenced


assessment.


The


highest


mean


rate


of SIB


was


demand


condition,


with


a difference


four


times


the


rate


between


that


condition


the


condition


with


the


next


highest


mean


rate


of responding.


Subject


19 had


mean


rates


of SIB


conditions


the


time-referenced


assessment.


The


rates


the


behavior


-referenced


assessment


were


highest


the


attention


condition


with


rates


the


demand


alone


condition.


The


play


condition


was


included


either


assessment.


Subject


24 had


highest


mean


rate


of SIB


the


alone


condition


time-referenced


assessment.


The


next


highest


rates


occurred


attention


condition


and


demand


conditions,


and


respectively.


behavior-referenced


Results


assessments


the


concurred


time-referenced


across


conditions.


The


alone


condition


resulted


the


highest


rates


in both


assessments,


and


lowest


rate


occurred


control


, behavior-referenced


condition.


Subject


29 exhibited


highest


mean


rate


SIB


attention


condition


behavior-referenced


assessment.


During


behavior-referenced


assessment,


the


highest


mean


rate


of SIB


occurred


during


the


demand


condition


with


the


1 aY,. ~h -t --a~4- 0Th in if 4~I~fl m4an +. 4 nn n nA + n


nnn~ 4 C(nn


1 AI.IA AC


n Ckn


wlrCA


AF CITP


nn



























Figure
refere
assess
axis a
assess
subject
mean r
the ra
of res


Mean


nced
ment
re th
ment
t's r
espon


responses


essment
present
ean resp
h the ra
of resp
for the


nges individually
ponding.


and
ed f
onse
nges
ondi
beh


per


th
or
s
i
ng
av


minute


e behavior-r
each subject
for the time
individually
On the ri
ior-referenc


determined


the


efe
t.
-re
det
ght
ed


each


time-


renced
On the
ferenced
ermined
y-axis
assessme


subj


ect'


left

by e
are
nt w
s ra


ach
the
ith
te













Mean Responses per minute


Time-Referenced Assessment


Demand -

Attention-

Alone

Play


Demand

Attention

Alone

Play


(I
sic',


r *


7(0


b-c

C


(3'
-'Co,


b-a


I r
(JI 0


Demand


Attentions

Alone

Play


-----

p


01

0)


Demand

Attention

Alone


E) A)


0------ '1

p


-o 0


*a.














Subject


32 exhibited


the


highest


mean


rates


demand


condition


time-referenced


assessment


the


alone


condition


behavior-referenced


assessment.


The


difference


between


attention


condition


rate


and


rate


the


alone


condition


was


small


Subject


exhibited


highest


rates


of SIB


demand


condition


of both


assessments.


There


was


little


difference


between


the


rates


the


attention


alone


conditions


of either


the


time-referenced


or the


behavior-


referenced


assessments.


The


rate


of SIB


the


play


condition


was


higher


than


rate


in either


the


alone


attention


condition


behavior-referenced


assessment.


A comparison


between


results


the


MAS


and


the


results


the


behavior


-referenced


assessments


these


subjects


is possible.


Subject


(see


Figure


the


rating


was


highest


the


tangible


category


and


the


mean


rate


of responding


was


highest


attention


category


behavior


-referenced


asse


ssment


see


Figure


, indic


eating


a possible


match


with


positive


reinforcement


as the


identified


maintaining


variable


Subject


MAS


category


with


highest


mean


rating


was


automatic


(see


a


Y-. --


a a


I ,a


JA _^ nn a


-lC1 n1 r% Am Anf nf


nr'nA4+ an A












MAS


results


Subject


were


regarded


as not


matching


behavior-referenced


assessment


because


two


category i


were


rated


equally


high


MAS


(automatic


tangible)


while


responding


the


behavior-referenced


asses


sment


was


highest


the


attention


condition.


Based


on thi


comparison,


the


MAS


identified


maintaining


variable


the


8 subjects.


Spearman


rank


order


correlations


based


on the


ranked


mean


responses


minute


time-referenced


assessment


and


the


behavior


-referenced


assessment


the


Escape,


Attention,


and


Alone


conditions


were


completed


the


subjects


the


participating


behavior-referenced


the


study


assessment


The


was


not


Play


condition


included


because


there


no corresponding


condition


the


time-referenced


assessment.


The


range


was


-0.5


to 1


with


a mean


.12.


In addition


, Spearman


rank


order


correlations


the


ranked


mean


MAS


ratings


and


ranked


mean


responses


per


minute


the


behavior-referenced


assessment


the


Demand/Escape,


Attention


and


Sensory/Alone


conditions


were


completed


each


the


eight


subjects.


range


was


-1.0


to 1


with


a mean


of -0.06


Discussion


S L a rY a a 4. nL aL 4


n* 1r n


a irmnail W a


rACIII1C C1


mt :


rrAllln*~PAd


C~A












provide


results


similar


those


the


behavior


-referenced


assessment


, thereby


providing


a more


efficient


functional


analysis


whom


the


both


variables


assessments


maintaining


were


SIB.


implemented,


the


the


8 subjects


time-


referenced


assessment


condition


with


the


highest


mean


rate


was


same


as the


behavior-referenced


assessment


condition


five


cases


That


highest


mean


rates


were


the


same


approximately


subjects.


Because


the


subjects


partic


ipating


thi


experiment


also


partic


ipated


Experiment


MAS


and


behavior-


referenced


assessment


results


were


compared


Such


comparison


revealed


that


the


MAS


ratings


staff


agree


with


results


an experimental


manipulation


of environmental


variables


only


the


subjects


, a 25%


agreement.


The


MAS


results


do not


agree


with


either


time-referenced


behavior


-referenced


analogue


assessment


results


General


Discus


sion


The


present


study


compared


results


obtained


from


three


methods


assessing


functional


properties


of SIB


interview


procedure


(the


MAS)


and


two


direct


methods


(time-


referenced


and


behavior


-ref


erenced


asse


ssments)


which


- ~~~~~~~~~ ~ ,,c, ---- aaa- i Ji -Va -i1t


,,,e ~ nAl a;


,,,,,, L,, t


,,,~?L


9 VA A












Experiment


was


a systematic


replication


the


Durand


and


Crimmins


between


the


(1988)


Durand


study. There

Crimmins and


were

the


two


major


present


differences


study,


and


these


differences


may


have


been


at least


partially


responsible


results


obtained


the


current


study


First


, the


demand


condition


the


Durand


Crimmins


study


was


aversivee"


along


dimension


of difficulty,


indicated


the


probability


a correct


response


on the


items,


and


possibly


along


the


dimension


time-out


from


the


task/attention.


the


current


study,


other


dimensions


"demand"


that


might


aversive


to subjects


were


included


effort


response,


physical


prompting,


and


demands


other


than


present


academic


study


tasks.


and


The


second


Durand


difference


Crimmins


study


between


was


selection


of subjects.


each


MAS


category,


Durand


Crimmins


that


selected


category


subjects


that


whose


category


MAS


alone.


ratings


The


were


current


high


study


selected


subjects


in a quasi-random


manner


Subjects


were


screened


as they


were


referred


an SIB


intervention


project


were


included


study


as the


assessment


portion


their


treatment


was


completed.


Therefore,


current


study


did


not


specifically


select


subjects


based


on any


rrk~ SlU CA*:~


n. 4 a rr


'rho


nr r (? ~ran~ S


Cka; r


Ctt~t~


r- '













been


less


differentiated


than


others


who


scored


particularly


high


in an MAS


category.


Experiment


involved


a comparison


two


types


functional


assessments


: brief


exposure


to experimental


conditions


the


time-referenced


assessment


and


lengthy


exposure


the


same


conditions


the


behavior-referenced


asses


sment.


The


time-referenced


results


matched


the


behavior-referenced


assessment


results


the


subjects.


These


results


suggest


that


the


accuracy


abbreviated


functional


analysis


may


limited;


correctly


identifying


the


maintaining


variable


62.5%


the


time


is a


success


rate.


The


brevity


of a subject'


contact


with


contingencies,


(through


sessions


brief


exposure)


may

sess


affect


ions


the


and


comparison


s/her


the


normal


subject'


environment


behavior


This


during


is a


familiar


problem


of sampling;


larger


the


sample


, the


more


representative


will


be of


total


population.


case


of a single


measurement,


subject'


in more


behavior


conditions


that


, more


frequent


approximate


possible


instances,


will


yield


more


accurate


data.


The


time-


referenced


assessment


the


current


study,


and


the


brief


analogue


ass


essment


Durand


and


Crimmins


(1988)


, Steege


a~~~~ ~ ~ a'i--- a '--- ** 1 -A


i A A


~


nnh I~A


B % j-%


rr~ 1,


II


A nL


'I


,,













variation


however,


thereof


may


, may


not


be useful


be a substitute


as a "screening"


a more


technique;


complete


functional


analysis.


As previously


discussed


, a possible


criticism


the


analogue


assessment


that


the


subject


trained


to exhibit


the


inappropriate


behavior


contingencies


manipulated


during


the


assessment.


The


training


would


be evident


data


the


form


an acquis


ition


curve.


The


subjects


the


current


study


did


increase


their


rate


SIB


over


the


course


the


analogue


sessions.


Instead


, a high


rate


of SIB


occurred


one


conditions


was


maintained,


rate


of SIB


other


conditions


decreases


discriminated


responding


developed.


Therefore,


there


was


evidence


that


subjects


were


trained


to exhibit


SIB


under


the


control


new


variables.


The


criticism


that


complexity


the


analogue


method


of functional


analyses


prohibits


use


most


settings


has


use


less accurate


methods


identifying


function


of SIB,


or to applying


contingencies


without


regard


the


maintaining


variable


es.


A need


exists


a more


easily


utilized


method


of determining


the


controlling


variables


inappropriate


behaviors.


It must


be pointed


fc -


. a


- ~ r --- --


"~ -an


n 1 t


---5


L a a A. a a i. U i 10At V f l


---3-
r* ~n nfl n r 0


LU 1.~L I,~IL


A


r u


-- ----


n mnna mnra


fc












harmful


the


subject.


example,


a subject'


SIB


was


determined


to be sensitive


(avoidance/escape)


identified


and


variable


to negative


assessment


as positive


reinforcement


inaccurately


reinforcement


(attention)


use


a time


procedure


procedure


would


might


constitute


implemented.


negative


The


reinforcement


and


the


subject'


would


increase


rate


or severity


as a


result


treatment.


There


were


two


possible


limitations


the


current


studi


es.


First,


poor


interrater


reliability


the


MAS


Experiment


may


have


biased


validity


MAS-Time-


referenced


assessment


comparison.


The


interrater


reliability


could


not


be improved


without


significant


complications.


Durand


& Crimmins


(1988)


may


have


had


better


interrater


reliability

behavior,


the


because


or because


questionnaire


their

their


while


responders were

responders were


observing


better

able


subject


observers


to complete


, perhaps


arranging


conditions


to test


some


the


answers.


either


these


possibilities


was


responsible


difference


MAS


longer


has


advantage


of being


ess


cumbersome


than


analogue


studi


and


requiring


trained


observers.


I
I a -
a -


a C r-. a-a 4.


~1 ,~. rcl,4 *1


cr,, c


h ; wk C


c


u


-1












with


more


subjects


may


alter


proportion


of matches


between


assessments.


Another


type


asses


sment


that


eliminates


some


criticisms


associated


with


the


functional


analysis


, and


that


may


also


eliminate


the


problems


inherent


the


indirect


assessments,

advantages c


naturalistic


>f naturalistic


observation


observation


include


method.


ease


The


use


because


there


is no need


to control


the


environment,


the


effi


ciency


with


respect


time


and


personnel


involved


minimal


disruption


subjects'


lives


Many


problems


might


be resolved


a more


standardized


method


observation

A naturalis


the


were

tic o


analogue


developed


observation


method


that m

method


maintained


may


functional


these


advantages.


an important


analyst


that


adjunct


more


idiosyncratic


variables


acting


on the


behavior


interest


might


identified.


Furthermore,


identification


these


variables


develop


during


an effective


assessment


treatment,


would


and


make


would


eas


facilitate


generalization


treatment


effects


the


natural


environment.


Further


research


investigating


the


effects


more


extensive


exposure


during


time-referenced


assessments,


-- S a a -


_ .-


_. --~---L ,,,


I 1


aI
















REFERENCES


S.W., Peterson, R.F., & Ault
to integrate descriptive and
studies at the level of data
Journal of Applied Behavior


, M.H. (1968). A method
experimental field
and empirical concepts.
Analysis, 1, 175-191.


Carr,


E.G. (1977)
behavior:
Bulletin,


The motivation
review of some
800-816.


of self-injurious
hypotheses. Psychological


Carr,


E.G., Newsom, C.D., & Binkoff, J. (1980). Escape as a
factor in the aggressive behavior of two retarded
children. Journal of Applied Behavior Analysis, 13,
101-117.


Chamberlain, P
report
97-109.


& Reid,
child


J.B. (1987). Parent
symptoms. Behavioral


observation and
Assessment, 9,


Demchak, M.A. & Halle, J.W. (1985). Motivational assessment:
A potential means of enhancing treatment success of
self-injurious individuals. Education and Training of
the Mentally Retarded, 20, 25-38.


Durand,


V.M., & Crimmins, D.B. (1988). Identifying t
variables maintaining self-injurious behavior.
Journal of Autism and Developmental Disorders,
117.


Iwata,


B.A., Dorsey,
Richman, G.S.
self-injury.
Developmental


M.F., Slifer, K.J., Bauman,
(1982). Toward a functional
Analysis and Intervention in
Disabilities, 2, 3-20.


K.E., &
analysis


Bijou,










B.A., Vollmer,
experimental (
disorders: Me
limitations.
Perspectives o
interventions
disabilities (
Publishing Co.


T.R.
funct
thodo
In A.
n the
for p
pp. 3


i
1
C

e
0


& z
onal
ogy,
. Re
use
rson
1-33


arcone, J. R. (199
) analysis of behav
applications, and
pp & N.N. Singh (E
of nonaversive and
s with development
0). Sycamore Il:


0).
ior


The


ds.) ,
aversive
1
Sycamore


Johnson, W.L., & E
behavior: A
details of
2, 465-484.


laumeister, A. (197
Review and analyst
published studies.


Self-injurious
of methodological
Behavior Modification,


Johnson


., S.M., & B
in natural
solutions f
L.C. Handy,
Methodology
Champa ign,


ols
sti
or
&


tad,
c obs
field
E.J.
oncep
.: Re


O.D. (
ervati
resea
Mash (
ts, an
search


1973
on:
rch.
Eds.
d Pr
Pre


). Methodological is
Some problems and
In L.A. Hamerlynck,
), Behavior Change:
actice (pp. 7-67).
ss.


Mace,


F.C., & Knight
treatment of
Analysis, 19,


D. (1986).
ever pica.
411-416.


Functional
Journal of


analysis
ADDlied


and
Behavior


Repp,


A.C., Felce, D., & Barton, L.E.
treatments of stereotypic and
on hypotheses of their causes.
Behavior Analysis, 21, 281-290


(1988). Basing the
self-injurious behavior
Journal of Applied


Sidman,


M. (1960).
Basic Books.


Tactics


scientific


research.


New


york:


Steege,


M.V.,
Cooper
to pre
handic
Analvs


Wacker
, L.J.
scribe
apped c
is, 22,


.P.,
89).
eva
dren
-33.


Berg, W.K., Cigrand K.C., &
The use of behavioral assessment
luate treatments for severely
Journal of Applied Behavior


Steege,


M.
C.G
(19
tre
ADD


V., Wacker, D
., Reimers, T
90). Use of
atment of sel
'lied Behavior


.P.,
.M. ,
nega
f-in
Ana


Cigrand, K.C
, Sasso, G.M.
tive reinforc
jurious behav
lysis, 23, 45


., Berg, W.K.
, & DeRaad, A
ement in the
ior. Journal
9-468.


Novak,


of


Sturmey


P., Carlsen, A., Cr
lr ai n3 1 n m 1 4


isp, A.G., & Newton,
nP ml 1 +- n Ti a nMorrnn+


J.T. (1988).
ranan^coc


Iwata,


;sues










Touchette, P.E., MacDonald, R.F., &
scatter plot for identifying
problem behavior. Journal of
Analysis, 8, 343-351.


Langer,
stimulus
Aoplied


S.N. (1985).
control of
Behavior


Wacker,


D.P.,
Reimers
(1990)
communi
severe
Analvsi


Steege, M.W.,
, T., Cooper,
A component
cation traini
behavior prob
s, 23, 417-43


Northup, J.,
L., Cigrand,
analysis of
ng across thr
lems. Journa
0.


Sasso, G.,
K., & Donn,
functional
ee topograph
1 of Applied


Berg,
L.


ies of
Behavior


Zarcone, J.R., Rodgers, T.A., Iwata, B.A., Rourke, D.A., &
Dorsey, M.F. (in press). Reliability analysis of the
Motivation Assessment Scale: A failure to replicate.
Research in Developmental Disabilities.















BIOGRAPHICAL


SKETCH


was


born


on July


, 1957,


Madison,


Wisconsin.


father


was


a technician


the


Air


Force


, and


family


lived


variety


of places


including


Japan,


Montana,


Colorado


Florida


and


New


York.


In 1975


I graduated


from


Peru


High


School


Peru,


New


York.


received


Bachelor


of Science


degree


1979


from


Framingham


State


College


Massachusetts.


I attended


the


University


Central


Florida


from


1979


to 1981,


received


Master


of S


clence


degree


clinical


psychology


I moved


to Gainesville,


Florida,


1981


and


held


a variety


of profe


ssional


positions


including


psychologist


the


Tacachale


institution.


I returned


school


1984


order


pursue


a Ph


. degree.


While


attending


school


was


employed


as a psychologist


Tacachale,


years.


was


a teaching


ass


istant


three


years


have


been


a research


ass


instant


on the


Self-


Injurious


Behavior


Intervention


Project


three


years.


1988


married


Michael


Stoutimore.









I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the Degree of Doctor of Philosophy.


BriAn A. Iwata,'Cha
Professor of Psycho


I certify that
opinion it conforms
presentation and is
a dissertation for


person
ogy


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the Degree of Doctor of Philosophy.


Marc N.
Profess


I certify that
opinion it conforms
presentation and is
a dissertation for


Bra ch
or of Psychology


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the Degy e of Doctor/ff Philosophy.


Edward F. Malagodi
Professor of Psycholo


I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the Degree of Doctor of Philosophy.


/- ^t


Henry/ .
Professor


Pennypack
of Psych


ogy


I certify that
opinion it conforms
presentation and is
a dissertation for


I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the Degrpe of Doct4r of Philosophy.


t -E~u~r rNJ. n


1


r~I~3111










I certify that
opinion it conforms
presentation and is
a dissertation for


I have read t
to acceptable
fully adequate
the Degree of


his study anC
standards of
e, in scope a
Doctor of Phi


^^mjul^


I that in my
scholarly
ind quality,
[losophy.


--F
William D. Working
Professor of Special


This
of the Dep
Arts and S
as partial
Doctor of


cation


dissertation was submitted to the Graduate Faculty
artment of Psychology in the College of Liberal
sciences and to the Graduate School and was accepted
fulfillment of the requirements for the degree of
Philisophy.


August,


1991


Dean,


Graduate


School






















UNIVERSITY OF FLORIDA
H IU I IIIIII Illlllillllll IH 111111Il II
3 1262 08556 9845