ADHD PROBLEM

65

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1998, 31, 65–78 NUMBER 1 (SPRING 1998)

CLASSROOM-BASED FUNCTIONAL AND ADJUNCTIVEASSESSMENTS: PROACTIVE APPROACHES TO

INTERVENTION SELECTION FOR ADOLESCENTS WITHATTENTION DEFICIT HYPERACTIVITY DISORDER

RUTH A. ERVIN

WESTERN MICHIGAN UNIVERSITY

GEORGE J. DUPAUL

LEHIGH UNIVERSITY

LEE KERN

UNIVERSITY OF PENNSYLVANIA

AND

PATRICK C. FRIMAN

FATHER FLANAGAN’S BOYS’ HOME AND

CREIGHTON UNIVERSITY SCHOOL OF MEDICINE

The present investigation evaluated the utility of classroom-based functional and adjunc-tive assessments of problem behaviors for 2 adolescents who met diagnostic criteria forattention deficit hyperactivity disorder (ADHD) and comorbid oppositional defiant dis-order (ODD). For children with ADHD-ODD, environmental classroom variables, whensystematically manipulated by teachers, were related to the occurrence and nonoccurrenceof problem behaviors. Classroom interventions derived from information that was ob-tained during functional and adjunctive assessments and from subsequent analyses re-sulted in substantial reductions in problem behaviors. Teacher and student consumersatisfaction ratings indicated that the interventions were effective and feasible in theclassroom setting.

DESCRIPTORS: behavioral assessment, ADHD, behavior disorders, classroom be-havior modification

Attention deficit hyperactivity disorder(ADHD; American Psychiatric Association,

Portions of this article served as the doctoral disser-tation for the first author. We thank Christine Cole,Linda Bambara, and Edward Shapiro for their timeand valuable feedback as members of the dissertationcommittee. In addition, we acknowledge Peg Millerfor her assistance with data collection as well as Cath-erine DeSalvo and the participants and teachers atBoys’ Town for their contributions to the present in-vestigation.

Correspondence regarding this manuscript shouldbe addressed to Ruth A. Ervin, Department of Coun-selor Education and Counseling Psychology, WesternMichigan University, Kalamazoo, Michigan 49008-5195.

1994) is a disorder characterized by prob-lems with sustained attention, impulsivity,and overactivity. In the classroom, childrenwith ADHD have difficulty sustaining atten-tion to tasks, completing assigned work, fol-lowing instructions, and adhering to generalclassroom rules (e.g., Barkley, 1990; DuPaul& Stoner, 1994). Children with ADHDmay have additional diagnoses characterizedby hostile, disobedient, and defiant behav-iors such as oppositional defiant disorder(ODD; American Psychiatric Association,1994). Research indicates that children withADHD-ODD who display persistent prob-

66 RUTH A. ERVIN et al.

lems in the classroom are at significant riskfor later academic failure, limited education-al attainment, and development of more ex-treme antisocial behavior (Barkley, 1990;Barkley, Fischer, Edelbrock, & Smallish,1990a, 1990b; Weiss & Hechtman, 1993).

The most widely used and effective inter-ventions for ADHD include stimulant med-ication, behavioral interventions, or both(Barkley, 1990), which have been found toreduce undesirable behaviors and enhanceacademic performance (DuPaul & Eckert,1997; Fiore, Becker, & Nero, 1993). How-ever, individual differences in response tovarious interventions have been noted(Whalen & Henker, 1991), and comorbiddisorders such as ODD may adversely affecttreatment success (Barkley, 1990).

The heterogeneity of ADHD-ODDsymptoms, coupled with highly variabletreatment efficacy, underscores the need forindividualized treatment. Unfortunately, lit-tle information is available to provide assis-tance in treatment selection for the individ-ual child with ADHD-ODD. Without asystematic method for intervention selec-tion, practitioners often resort to trial anderror, a process that can deplete scarce timeand resources. Functional assessment is onepromising method for intervention selectionthat has been increasingly touted as a supe-rior alternative to traditional psychiatric di-agnostics that are often not tightly connect-ed to interventions (DuPaul & Ervin, 1996;Kratochwill & McGivern, 1996; Zentall &Javorsky, 1995). Although the functional as-sessment literature has focused primarily onproblem behaviors exhibited by persons withdevelopmental disabilities (e.g., Iwata, Dor-sey, Slifer, Bauman, & Richman, 1982/1994), this approach has been extended tobehavior problems exhibited by childrenwith average intelligence (e.g., Cooper,Wacker, Sasso, Reimers, & Donn, 1990).

One line of research has focused on test-ing variables that may maintain functions of

problem behavior through brief manipula-tions of consequences (e.g., Northup et al.,1995). For example, Northup et al. con-ducted brief functional analyses of the be-haviors (i.e., out of seat, inappropriate vo-calizations) of 3 children (7 to 9 years old)who had been diagnosed with ADHD.These analyses were conducted in an ana-logue classroom setting, and results indicatedthat the frequency of problem behaviors washighest during a peer attention condition.Broussard and Northup (1995) extended thefindings of Northup et al. to integrated pub-lic school settings. Brief functional analysissessions were conducted in the general edu-cation classroom of 3 young students, oneof whom was a 6-year-old boy who had beendiagnosed with ADHD. Brief analyses sug-gested that his problem behavior was main-tained by escape from difficult academictasks. Systematic analogue analyses conduct-ed by Northup et al., together with otherschool case reports (i.e., Broussard & Nor-thup, 1995; Lewis & Sugai, 1996; Umbreit,1995), suggest that functional analysis maybe a useful method of assessment and inter-vention selection in the classroom setting foryoung children with ADHD.

Another line of functional assessment re-search involves identifying functional rela-tions between curricular variables and desir-able and undesirable student behavior (forreview, see Dunlap & Kern, 1996). Curric-ular variables that have been shown to sys-tematically relate to target behaviors can bemodified to decrease or eliminate problembehaviors. This approach to classroom inter-vention selection has been successfully dem-onstrated with students with various disabil-ities (e.g., Dunlap et al., 1993; Kern, Childs,Dunlap, Clarke, & Falk, 1994). For exam-ple, Kern et al. evaluated the effectiveness ofteacher implementation of five curricularmanipulations (e.g., short vs. long tasks,written vs. nonwritten work) on the on-task

67FUNCTIONAL AND ADJUNCTIVE ASSESSMENTS

behavior of an 11-year-old boy with averageintelligence.

As functional assessment moves from an-alogue to applied settings, issues of assess-ment, treatment integrity, and acceptabilityare of increasing importance and may im-pede the degree to which traditional func-tional assessment procedures are feasible. Forreasons such as teacher preference, ease ofimplementation, and matching interventionsto existing classroom routines or structures,teachers may be reluctant to manipulate cer-tain events contingent on the occurrence ofproblem behavior. Specifically, althoughteachers may inadvertently or unintention-ally reinforce problem behaviors (e.g., send-ing a student to the office for misbehavingduring an academic activity), they may bereluctant to intentionally implement ana-logue procedures to increase problem behav-ior, even for the purpose of determining be-havioral function (Repp & Karsh, 1994). Inaddition, determination of behavioral func-tion may not be feasible in naturalistic set-tings when direct control over the presumedreinforcer is difficult to obtain (e.g., peer at-tention). Alternatively, teachers may be morewilling to systematically manipulate antece-dent events (Martens & Kelly, 1993). Al-though not a direct examination of behav-ioral function, this approach has the advan-tage of providing the teacher with informa-tion concerning the efficacy of a potentialintervention strategy. In addition, in cases inwhich systematic control of potential rein-forcing stimuli is difficult, it may be neces-sary to explore assessment-based adjunctiveintervention strategies that support the de-velopment and systematic reinforcement ofincompatible behavior.

The purpose of the current research wasto address the above concerns and extend theprevious literature by assessing the applica-bility of a model of school-based assessmentthat included functional assessment proce-dures as well as adjunctive procedures with

adolescents having a long history of problembehaviors (i.e., ADHD and related difficul-ties) and for whom previous interventionshad been unsuccessful. The assessment mod-el evaluated in the present investigation ex-tends previous school-based functional as-sessment studies (e.g., Cooper et al., 1992;Dunlap et al., 1993; Kern et al., 1994) byincluding adjunctive curricular manipula-tions to explore whether determination ofbehavioral function (i.e., through the directmanipulation of maintaining or antecedentvariables) is necessary for the identificationof effective and socially acceptable interven-tions in the natural context. Another pur-pose was to explore the utility of a collabo-rative consultation model in the develop-ment and selection of hypotheses and inter-vention strategies. In short, we wereinterested in the involvement of teachersthroughout the functional assessment pro-cess, with an emphasis on providing themwith choices among several interventionstrategies. A final goal was to evaluate theeffectiveness and the acceptability of inter-ventions derived from the assessment infor-mation for 2 individuals who had been di-agnosed with ADHD and ODD.

METHOD

Participants

Both participants met the criteria forADHD of the Diagnostic and StatisticalManual of Mental Disorders (4th ed., DSM-IV; American Psychiatric Association, 1994)based on the ADHD Rating Scale-IV(DuPaul, Anastopoulos, Power, Murphy, &Barkley, 1994), information obtained duringa diagnostic interview, and guardian andteacher ratings on the attention problemsfactor of the Child Behavior Checklist (Ach-enbach & Edelbrock, 1991). Both partici-pants also met the DSM-IV criteria forODD, and neither participant met criteria


for conduct disorder based on diagnostic in-terviews.

Joey. Joey was a 13-year-old Caucasian boywho obtained a full-scale IQ score of 98 onthe Wechsler Intelligence Scale for Children(3rd ed., WISC-III; Wechsler, 1991). Joeyattended the seventh grade and was referredby his teacher for concerns with excessiveoff-task behavior. Throughout the investi-gation, Joey took 20 mg of methylphenidatetwice daily. Documentation indicated thatmedication was taken 100% of the timethroughout the study.

Carl. Carl was a 14-year-old Hispanic boywho obtained a full-scale IQ score of 91 onthe WISC-III (Wechsler, 1991). Carl attend-ed the eighth grade and was referred by hiswriting and math teachers for concerns withbehaviors that were disruptive (i.e., talking,making funny noises, making faces, gestures,or writing notes). Throughout the investi-gation, Carl took 10 mg of methylphenidatefour times daily. Documentation indicatedthat medication was taken 100% of thetime.

Setting

The study was conducted at a school thatserves approximately 250 students, Grades 1through 8, at Boys’ Town. The program uti-lizes the Teaching Family Model (e.g., Phil-lips, Phillips, Fixen, & Wolf, 1971) that em-ploys a comprehensive token economy. Thiswas maintained throughout the study. Hy-pothesis testing and intervention conditionswere conducted by the teacher within thenatural classroom setting. Each classroom inthe present investigation, with one excep-tion, consisted of 7 to 12 students and oneteacher. For Carl, two writing classes werecombined, consisting of 14 to 24 students.

Dependent Measures and InterobserverAgreement

On-task behavior. On-task behavior con-sisted of attending to assigned class work in

an active (i.e., writing, reading aloud, raisinghand, verbal behavior relevant to the task)or passive (i.e., eyes directed toward the taskor teacher during lecture, looking at work,silent reading) manner. Off-task behaviorwas coded if any problem behaviors occurredwithin the interval (i.e., calling out, gestur-ing, talking to peers, playing with objects,making funny faces) or if the student wasnot attending to the task for 3 s or more.Data were collected using the Problem Be-havior Observation Form (available from thefirst author). Students were observed for a15-s interval followed by a 5-s period forrecording.

Teacher satisfaction ratings. This question-naire assessed teacher satisfaction and con-sisted of 10 items rated on an 8-point Likertscale ranging from 0 (not at all) to 7 (verymuch). Items focused on intervention effec-tiveness (e.g., ‘‘problem behavior improvedduring the intervention’’), feasibility (e.g.,‘‘the intervention was time consuming toimplement’’), and social validity (e.g., ‘‘thestudent seemed embarrassed by the interven-tion’’). The questionnaire was completed af-ter the intervention phase and is availablefrom the first author.

Student satisfaction ratings. A consumersatisfaction questionnaire (available from thefirst author) assessed the students’ opinionsby asking them to rate the success of theintervention (e.g., ‘‘I got more done duringthe intervention’’) and social validity (e.g.,‘‘the intervention was embarrassing’’). Joeyand Carl completed the questionnaire afterthe intervention phase.

Procedural integrity. To assess proceduralfidelity, implementation of the manipulatedvariables during hypothesis testing and in-tervention was recorded on the Problem Be-havior Observation Form. If the manipulat-ed variable was a type of task or task struc-ture, the data collector wrote a descriptionof the activity (e.g., math word problems,correcting homework, class discussion) and


the way in which it was carried out (e.g.,lecture, independent task, small groups,working on the computer). If teacher ma-nipulations involved a series of steps (e.g.,steps involved in a self-evaluation proce-dure), a treatment integrity checklist wascompleted.

Interobserver agreement. An independentobserver (unaware of the nature of the study)collected data during a minimum of 25% ofobservations across participants and condi-tions. This observer was trained on the data-collection procedures to 80% agreement pri-or to the beginning of the investigation. In-terobserver agreement for on-task behaviorwas calculated by dividing the number ofagreements by the number of agreementsand disagreements and multiplying by100%.

During hypothesis testing, interobserveragreement data were collected during 33%of the observations across conditions andparticipants. For Joey, average total, occur-rence, and nonoccurrence agreements foron-task behaviors were 94%, 70%, and91%, respectively. For Carl, average total,occurrence, and nonoccurrence agreementson on-task behaviors were 90%, 67%, and86%, respectively.

During intervention evaluation for Joey,interobserver agreement was taken on 30.8%of sessions during baseline and interventionconditions. Average total, occurrence, andnonoccurrence agreements were 92%, 75%,and 90%, respectively. For Carl, interobserv-er agreement was taken on 33% of sessionsin baseline and intervention conditions dur-ing math, writing, and science classes. Av-erage total, occurrence, and nonoccurrenceagreements were 96%, 81%, and 96%, re-spectively.

General Functional Assessment Procedures

A collaborative consultation model wasused wherein teachers participated in allphases of the functional assessment. The first

author served as the consultant. Procedureswere adapted from previous literature onschool-based functional assessment for thepurpose of curricular revisions (e.g., Dunlapet al., 1993; Dunlap, Kern-Dunlap, Clarke,& Robbins, 1991; Kern et al., 1994).

Descriptive assessment and hypothesis devel-opment. Hypotheses regarding potential in-tervention options were generated collabor-atively and were based on (a) interviews withteachers using the Preliminary FunctionalAssessment Survey (Dunlap et al., 1991), (b)interviews with participants using the Stu-dent Assisted Functional Assessment Inter-view (Kern, Dunlap, Clarke, & Childs,1995), (c) direct observations of teachers us-ing the Behavioral Tracking Form (adaptedfrom O’Neill, Horner, Albin, Storey, &Sprague, 1990), and (d) direct observationsby the consultant during times and situa-tions the teachers had identified as problem-atic.

Each hypothesis statement was based ontwo convergent sources of information (e.g.,interviews and direct observations), identi-fied specific variables that could be observedand manipulated by the teachers within theclassroom context, and were agreed upon byboth the teacher and consultant as reason-able, given the accumulated information(Dunlap et al., 1993). Two hypotheses weredeveloped for each participant and are de-scribed in the Results.

Hypothesis testing and intervention devel-opment. Hypotheses generated in the de-scriptive analysis were tested empiricallythrough teacher manipulations using a briefreversal design (Kazdin, 1982). Baseline con-ditions (e.g., typical conditions in the class-room) were alternated with conditions hy-pothesized to produce low levels of problembehavior (potential interventions). After hy-potheses were tested, the teacher and con-sultant jointly selected intervention compo-nents to implement on an ongoing basis. Se-lection was based on efficacy and teacher


preference. The process of intervention se-lection for each participant is described inthe Results.

Intervention evaluation. Prior to imple-menting interventions, teachers agreed to re-turn to their usual teaching procedures toestablish an initial baseline. Joey’s interven-tion was tested during writing class throughan ABAB design. Consistent with a multiplebaseline design, intervention was staggeredacross three academic subjects for Carl. Dur-ing math and writing classes, the interven-tion was briefly withdrawn.

RESULTS

Descriptive Assessment and HypothesisDevelopment

Joey. Information obtained from theteacher’s interview and direct observationsindicated that Joey was most likely to engagein off-task behaviors when he was presentedwith a pencil-and-paper writing task. Duringthe student interview, Joey indicated thatone of his least favorite classes was writing.Brief journal writing (i.e., 5 to 7 min) andlong story writing (i.e., 20 to 25 min) wererequired daily activities during writing class.According to interviews and observations,teacher prompts to begin writing were un-successful, and Joey’s off-task behaviors fre-quently (once or twice per week) led to of-fice referrals. When asked what might im-prove his behavior, Joey stated that he woulddo better if he were given more time tothink about what he had to write. Whenpresented with this information, Joey’steacher agreed that he was more likely to beactively engaged in journal writing if he par-ticipated in a discussion about the topic pri-or to writing.

It was hypothesized that Joey’s off-task be-haviors might be maintained by escape frompaper-and-pencil writing tasks. Thus, poten-tial intervention strategies included (a) pro-viding an alternative means for accomplish-

ing the writing tasks (e.g., talking into a taperecorder, writing on the computer), (b) giv-ing Joey extra time to think about what hehad to write prior to the journal-writing ac-tivity, (c) preventing escape from writingtasks by requiring Joey to complete his writ-ing task in the office when he received a re-ferral, or (d) allowing escape (e.g., briefbreaks) from writing tasks contingent on anappropriate request.

Joey’s teacher opted to provide Joey witha computer as an alternative writing methodduring long writing tasks, because this in-tervention could be implemented with sev-eral other classmates so Joey would not besingled out. Thus, the first hypothesis statedthat ‘‘Joey’s on-task behavior will be in-creased when he is given the opportunity tocomplete long (20 min) writing tasks on thecomputer rather than by hand.’’

Because journal-writing activities werebrief (and potentially less aversive), and be-cause Joey’s teacher was interested in obtain-ing some handwritten samples of his work,a second hypothesis was developed to ad-dress problem behavior during the shortjournal-writing activity. The second hypoth-esis stated, ‘‘Joey’s on-task behavior will beincreased when he is able to brainstorm witha peer prior to a short (5 to 7 min) writtentask.’’ This was accomplished by pairing stu-dents to take turns brainstorming out loudabout their journal topic for a 2-min period.Joey’s writing teacher set an egg timer andcirculated around the room to monitorbrainstorming. Both hypotheses were testedduring the same week in writing class. Hy-pothesis 1 was tested during the long (20 to25 min) portion of writing class, and Hy-pothesis 2 was tested during the brief (5 to7 min) journal-writing activity.

Carl. Descriptive observations and teacherreport indicated that Carl’s off-task behav-iors were followed by peer attention (e.g.,laughter, smiling, returned comments or ges-tures) and were preceded by peer solicitation


(e.g., looking his way, calling his name, mak-ing funny gestures or comments). Carl’s be-haviors occurred at high frequencies duringdescriptive observations, yet consequencesfrom teachers were infrequent and inconsis-tent (e.g., warnings or prompts to discontin-ue disruptive behaviors were sometimes mit-igated with a smile). In addition, teacher re-port and observations suggested that histeachers provided infrequent consequencesto his peers for responding to his attention-seeking behaviors. Teachers reported thatwhen they were able to monitor Carl closely,his problem behaviors were less frequent.This was supported by direct observation.Carl’s disruptive behaviors occurred less fre-quently when teachers were close (i.e., with-in 2 m) or when teachers made eye contactwith him.

This descriptive assessment informationsuggested that Carl’s peer disruptions mightbe maintained by peer attention. Potentialintervention strategies discussed with theteachers included (a) reducing access to peerattention (e.g., separation, reducing peer re-sponsiveness by providing consequences tohis peers), (b) providing contingencies forappropriate behavior that was incompatiblewith peer attention-seeking behavior (e.g.,self-monitoring of on-task behavior), (c)structuring the classroom so that Carl wouldbe less likely to engage in attention-seekingbehaviors (e.g., place Carl in close proximityto the teacher and provide prompts to stayon task and not to disrupt his peers), and(d) providing more consistent and frequentreinforcement (i.e., praise, points) for on-task behavior and punishers (i.e., negativepoints, verbal reprimands) for disruptingpeers.

Carl’s math teacher indicated that she be-lieved self-evaluation procedures would beeffective and feasible to implement. Thus,the first hypothesis stated, ‘‘Carl’s on-taskbehavior will increase when he is instructedto self-evaluate his peer attention-seeking be-

haviors and is awarded points for accuracyand low levels of problem behaviors.’’ Thishypothesis was tested through the use of self-evaluation procedures (adapted from Rhode,Morgan, & Young, 1983) in which Carl wasasked to rate his appropriate behavior on ascale of 0 (unacceptable) to 5 (excellent) atthe end of each class period. On the dayswhen self-evaluation occurred, Carl’s teacherplaced the self-evaluation card, containingspecific criteria for each numbered rating, onhis desk at the beginning of math class. Atthe end of the class period, Carl was askedto rate his behavior while the teacher alsoindependently rated his behavior. If theirratings were within one point of each other,points were awarded. If their ratingsmatched exactly, Carl received bonus points.If ratings were two or more points discrep-ant, no points were awarded. Points earnedduring self-evaluation were converted to val-ues that were comparable to an already ex-isting token economy and were later ex-changed for privileges on the residentialcampus.

Carl’s writing teacher preferred a classwideprocedure. As a result, the second hypothesisstated, ‘‘Carl’s on-task behavior will increasewhen he does not receive social reinforcersfrom his peers for his behavior.’’ To test thishypothesis, the teacher informed all studentsthat on certain days they would be workingon peer relations. This format was chosenbecause Carl’s teacher reported that she of-ten spent time focusing on a selected socialskill with her classes. On days in which thepeer intervention was in place, the teacherdelivered positive points to the students fornot responding and negative points for re-sponding to attention-seeking behaviors.Carl was not singled out as the target of theintervention.

Hypothesis Testing and InterventionDevelopment

Joey. The results of the multielement anal-yses for the two hypotheses are presented in


Figure 1. The percentage of intervals of on-task behavior during hypothesis testing and intervention forJoey.

Figure 1. The percentage of intervals inwhich Joey was on task during long writingtasks was higher when he used a computer(M 5 96.8%) than when he wrote by hand(M 5 64.8%; Hypothesis 1). Similarly, thepercentage of intervals with on-task behav-iors was higher when he brainstormed witha peer prior to journal writing (M 5 91.4%)than when no brainstorming occurred (M 563.2%; Hypothesis 2). Procedural integritydata were collected during all hypothesis-testing sessions; integrity was 100%.

Based on the results of hypothesis testing,Joey’s writing teacher selected both interven-tion components (i.e., brainstorming andcomputer in writing class) as interventionsto be implemented on an ongoing basis.During intervention evaluation, brainstorm-ing occurred during the journal-writing por-tion of writing class, and Joey was instructedto use the computer for longer writing tasks.

Carl. Figure 2 presents the results of thetwo hypotheses. The percentage of intervalsin which Carl was observed to be on taskwas higher when he self-evaluated his peerattention-seeking behaviors (M 5 92.2%)than when he did not self-evaluate (M 563.1%; Hypothesis 1). In addition, the per-centage of intervals with on-task behaviorwas higher when Carl’s peers received con-sequences for responding to his attention-seeking behavior (M 5 78.2%) than whenpeers were not given consequences (M 558.2%; Hypothesis 2).

Procedural integrity data were collectedduring all hypothesis-testing sessions. Self-evaluation procedures were implementedwith 100% integrity (Hypothesis 1). Duringpeer intervention, peers responded to Carl’sinappropriate behavior during 62.7% of theintervals in which Carl displayed problembehavior, in comparison to 85% of the in-


Figure 2. The percentage of intervals of on-task behavior during hypothesis testing and intervention forCarl.


tervals when no intervention was in place.Teachers applied consequences to the peers’reactions in 51.1% of the intervals in whichpeer responses to Carl’s problem behavioroccurred during the peer intervention, as op-posed to 4% of the intervals when no inter-vention was in effect.

Both teachers and the consultant agreedthat the self-evaluation procedure would beimplemented as the sole intervention be-cause (a) it was easier to implement, (b) pro-cedural integrity was higher, and (c) it wasequally effective. Carl’s self-evaluation inter-vention was evaluated across math, writing,and science classes.

Intervention Evaluation

Joey. Data reflecting the percentage of in-tervals on task during writing class are pre-sented in Figure 1. During baseline, on-taskbehavior was variable, with a downwardtrend and a mean of 67.7% (range, 54.2%to 83.3%). Following implementation of theintervention, the percentage of intervals ontask was stable and increased to a mean of96% (range, 93.7% to 98.3%). When theintervention was briefly withdrawn, the per-centage of intervals on task dropped to62.7%. With the reimplementation of theintervention, on-task behavior improved,with a mean of 95.4% (range, 90.7% to98.7%). Procedural integrity data were takenduring all sessions and indicated that bothbrainstorming and computer procedureswere implemented with 100% integrity.

Carl. The percentages of intervals inwhich Carl was on task during math, writ-ing, and science classes are presented in Fig-ure 2. In all three classes, the percentage ofintervals on task increased immediately fol-lowing implementation of the interventionand remained high throughout the interven-tion phases. During baseline phases (includ-ing the brief reversals), on-task behavior av-eraged 69.1% (range, 61.3% to 76%) inmath, 54.2% (range, 36% to 78.7%) in

writing, and 78% (range, 66.7% to 88%) inscience. During self-evaluation phases, on-task behavior increased to 93.1% (range,81% to 100%) in math, 88.2% (range, 72%to 98.7%) in writing, and 95.1% (range,88% to 98.7%) in science.

During baseline and intervention, proce-dural integrity data were taken on 96% ofthe sessions across math, writing, and sci-ence classes. Procedural integrity was 100%in all sessions except two. During writingclass, Carl’s teacher failed to implement oneof the five self-evaluation steps (i.e., teacherrating of the student’s behavior indepen-dently) on two occasions, resulting in amean integrity of 81.8% of the interventionsessions in this class.

Consumer satisfaction ratings. Joey’s ratingsof the combined computer and brainstorm-ing intervention indicated 100% satisfac-tion. Carl’s ratings of the self-evaluation in-tervention indicated 83.3% (35 out of a pos-sible 42 points) satisfaction. Teacher ratingswere also positive (94.3% for Joey’s teacher;M 5 87.1%, range, 81.4% to 97.1%, forCarl’s teachers).

DISCUSSION

The results support the utility of school-based functional assessment for adolescentswith ADHD-ODD. Information gatheredfrom multiple sources led to the collabora-tive development of plausible hypothesesthat were validated during manipulations ofclassroom variables by teachers. Further-more, interventions resulted in improve-ments in students’ behavior. These findingsextend previous research on functional as-sessment for students with ADHD (e.g.,Northup et al., 1995) in several importantways.

First, the present evaluation was conduct-ed in the natural setting without disruptionof the ongoing class routine. Second, thefunctional assessment process was conducted


with 2 students who were in their early ad-olescence and had been diagnosed withADHD and comorbid ODD, whereas pre-vious investigations (e.g., Umbreit, 1995)focused on younger students with ADHDonly. This is important because childrenwith ADHD often continue to exhibit prob-lems during adolescence, and their risk forthe development of more severe problemsincreases with age (Barkley, 1990).

Behaviors exhibited by students withADHD-ODD were found to be systemati-cally related to environmental variables inthe present investigation. The present find-ings, in conjunction those of similar previ-ous studies (Broussard & Northup, 1995;Lewis & Sugai, 1996; Northup et al., 1995;Umbreit, 1995), illustrate the important roleof the environment and its relation to prob-lem behavior displayed by students withADHD. In addition, the current study dem-onstrates a systematic strategy to assist in de-veloping effective individualized interven-tions with this population.

Measures of procedural fidelity indicatedthat teachers manipulated variables asplanned for three of the four hypothesestested. Manipulations involving antecedentprocedures (i.e., writing on the computer;brainstorming) were implemented withhigher integrity than those interventions thatinvolved contingent delivery of conse-quences (i.e., peer intervention). These find-ings are consistent with previous researchsuggesting that consequence-oriented inter-ventions are sometimes viewed by teachersas difficult and time consuming to imple-ment (see Martens & Kelly, 1993). How-ever, it is important to note that significantimprovements in behavior were observedeven when implementation of the peer in-tervention was only moderately consistent.

Students and teachers reported high levelsof satisfaction with the interventions de-signed from the functional assessment infor-mation. In general, teachers rated the inter-

ventions as practical, feasible, and successful.Both students reported that the interven-tions were not embarrassing, helped to im-prove their behavior, and would likely ben-efit other students. These findings are con-sistent with the literature on treatment ac-ceptability, which suggests that interventionsdesigned through collaborative consultationbetween a teacher and school psychologistare rated significantly more acceptable thanare interventions designed by school psy-chologists or teachers alone (Kutsick, Gut-kin, & Witt, 1991).

For both students, a functional relationwas demonstrated between environmentalvariables and problem behavior. However, inmost cases, hypotheses involved antecedentmanipulations (i.e., use of the computerduring writing tasks) or adjunctive manip-ulations (i.e., brainstorming, self-evaluation)and thus provide no formal evaluation ofmaintaining variables. Because the conse-quences that maintained problem behaviorwere not directly manipulated in these cases,behavioral function remains unknown. Oneexception was that peer attention, which washypothesized to maintain Carl’s inappropri-ate behavior, was directly manipulated. Pro-cedural integrity for this assessment was low(i.e., peers continued to attend to Carl’s in-appropriate behavior but at lower levels),however, so it is also difficult to draw firmconclusions regarding the functional role ofpeer attention. Still, the current findingsshow that antecedent assessments may beused to develop effective treatments forproblem behavior even when the conse-quences responsible for problem behaviorare not clearly identified. This was particu-larly important in Carl’s case because the as-sessment results showed that it was difficultto eliminate the consequence that was hy-pothesized to maintain his problem behavior(i.e., the peer intervention reduced but didnot eliminate peer attention for Carl’s in-appropriate behavior). Self-evaluation was


selected (over the peer intervention) becauseintegrity was better, it was preferred byteachers, and hypothesis testing showed thatit was an effective treatment, even though itdid not alter the hypothesized response–re-inforcer relation (i.e., the relation betweendisruptions and peer attention).

The results of the idiosyncratic interven-tions may not generalize to other classroomsettings in which the student is currently in-volved or to future settings. In this study, acomprehensive token economy was in placein each classroom. In addition, the class-room settings included a small number ofstudents (i.e., 7 to 12) in contrast to averageclassroom size (i.e., 25 to 30) in a publicschool setting. Thus, these unique contex-tual features may limit the generality of theeffects of the specific intervention strategiesto other settings and students. Both studentswere taking medication throughout thecourse of the study. This limits the degreeto which we can assume that the interven-tions would have been effective if the med-ications were withdrawn. As noted by Nor-thup et al. (1995), it is important to con-sider the separate and combined effects ofbehavioral interventions and medications inthe treatment of ADHD.

Finally, the dependent variable was on-task behavior rather than academic produc-tivity, and this makes it difficult to deter-mine whether changes in academic perfor-mance occurred. For Joey, both teacher andstudent intervention ratings suggested thatacademic performance and work productiv-ity improved. Similarly, Carl’s science teach-er, writing teacher, and his self-ratings sug-gested that his academic performance andwork productivity improved as a result of theself-evaluation intervention. Carl’s mathteacher reported that, although Carl’s workproductivity appeared to improve somewhat,his academic performance (i.e., quiz grades)did not. This teacher did, however, reportthat the self-evaluation intervention im-

proved the academic performance of otherstudents by reducing Carl’s distracting be-havior.

In general, these findings support the util-ity of functional assessment as a processthrough which classroom intervention strat-egies can be selected and evaluated for ado-lescents with ADHD-ODD. In particular,the present investigation provides supportfor a classroom-based model of functionalassessment that emphasizes the developmentof intervention strategies that are feasibleand acceptable to teachers and students. Al-though more research is warranted, thesepreliminary findings support the use of as-sessment-based approaches to interventionselection for students with ADHD-ODD inschool settings.

REFERENCES

Achenbach, T. M., & Edelbrock, C. S. (1991). Man-ual for the cross-informant program for the child be-havior checklist. Burlington: University of Ver-mont, Department of Psychiatry.

American Psychiatric Association. (1994). Diagnosticand statistical manual of mental disorders (4th ed.).Washington, DC: Author.

Barkley, R. A. (1990). Attention-deficit hyperactivitydisorder: A handbook for diagnosis and treatment.New York: Guilford Press.

Barkley, R. A., Fischer, M., Edelbrock, C. S., & Smal-lish, L. (1990a). The adolescent outcome of hy-peractive children diagnosed by research criteria:I. An 8-year prospective follow-up study. Journalof the American Academy of Child and AdolescentPsychiatry, 29, 546–557.

Barkley, R. A., Fischer, M., Edelbrock, C. S., & Smal-lish, L. (1990b). The adolescent outcome of hy-peractive children diagnosed by research criteria:III. Mother-child interactions, family conflicts,and maternal psychopathology. Journal of ChildPsychology and Psychiatry, 32, 233–256.

Broussard, C. D., & Northup, J. (1995). An ap-proach to functional assessment and analysis ofdisruptive behavior in regular education class-rooms. School Psychology Quarterly, 10, 151–164.

Cooper, L. J., Wacker, D. P., Sasso, G. M., Reimers,T. M., & Donn, L. K. (1990). Using parents astherapists to evaluate appropriate behavior of theirchildren: Application to a tertiary diagnostic clin-ic. Journal of Applied Behavior Analysis, 23, 285–296.


Cooper, L. J., Wacker, D. P., Thursby, D., Plagmann,L. A., Harding, J., Millard, T., & Derby, M.(1992). Analysis of the effects of task preferences,task demands, and adult attention on child be-havior in outpatient and classroom settings. Jour-nal of Applied Behavior Analysis, 25, 823–840.

Dunlap, G., & Kern, L. (1996). Modifying instruc-tional activities to promote desirable behavior: Aconceptual and practical framework. School Psy-chology Quarterly, 11, 297–312.

Dunlap, G., Kern, L., dePerczel, M., Clarke, S., Wil-son, D., Childs, K. E., White, R., & Falk, G. D.(1993). Functional analysis of classroom variablesfor students with emotional and behavioral dis-orders. Behavioral Disorders, 18, 275–291.

Dunlap, G., Kern-Dunlap, L., Clarke, S., & Robbins,F. R. (1991). Functional assessment, curricularrevision, and severe behavior problems. Journal ofApplied Behavior Analysis, 24, 387–397.

DuPaul, G. J., Anastopoulos, A. D., Power, T. J., Mur-phy, K., & Barkley, R. A. (1994). The ADHDRating Scale—IV. Unpublished manuscript, Le-high University.

DuPaul, G. J., & Eckert, T. L. (1997). The effects ofschool-based interventions for attention deficit hy-peractivity disorder: A meta-analysis. School Psy-chology Review, 26, 5–27.

DuPaul, G. J., & Ervin, R. A. (1996). Functionalassessment of behaviors related to attention-defi-cit/hyperactivity disorder: Linking assessment tointervention design. Behavior Therapy, 27, 601–622.

DuPaul, G. J., & Stoner, G. (1994). ADHD in theschools: Assessment and intervention strategies. NewYork: Guilford Press.

Fiore, T. A., Becker, E. A., & Nero, R. C. (1993).Educational interventions for students with atten-tion deficit disorder. Exceptional Children, 60,163–173.

Iwata, B., Dorsey, M., Slifer, K., Bauman, K., & Rich-man, G. (1994). Toward a functional analysis ofself-injury. Journal of Applied Behavior Analysis,27, 197–209. (Reprinted from Analysis and Inter-vention in Developmental Disabilities, 2, 3–20,1982)

Kazdin, A. E. (1982). Single-case research design. NewYork: Oxford University Press.

Kern, L., Childs, K. E., Dunlap, G., Clarke, S., &Falk, G. D. (1994). Using assessment-based cur-ricular intervention to improve the classroom be-havior of a student with emotional and behavioralchallenges. Journal of Applied Behavior Analysis,27, 7–19.

Kern, L., Dunlap, G., Clarke, S., & Childs, K. E.(1995). Student-assisted functional assessment in-terview. Diagnostique, 19, 29–39.

Kratochwill, T. R., & McGivern, J. E. (1996). Clin-ical diagnosis, behavioral assessment, and func-tional analysis: Examining the connection be-

tween assessment and intervention. School Psy-chology Review, 25, 342–355.

Kutsick, K. A., Gutkin, T. B., & Witt, J. C. (1991).The impact of treatment development process, in-tervention type, and problem severity on treat-ment acceptability as judged by classroom teach-ers. Psychology in the Schools, 28, 325–331.

Lewis, T. J., & Sugai, G. (1996). Functional assess-ment of problem behavior: A pilot investigationof the comparative and interactive effects of teach-er and peer social attention on students in generaleducation settings. School Psychology Quarterly, 11,1–19.

Martens, B. K., & Kelly, S. Q. (1993). A behavioranalysis of effective teaching. School PsychologyQuarterly, 8, 10–26.

Northup, J., Broussard, C., Jones, K., George, T., Voll-mer, T. R., & Herring, M. (1995). The differ-ential effects of teacher and peer attention on thedisruptive classroom behavior of three childrenwith a diagnosis of attention deficit hyperactivitydisorder. Journal of Applied Behavior Analysis, 28,227–228.

O’Neill, R. E., Horner, R. H., Albin, R. W., Storey,K., & Sprague, J. R. (1990). Functional analysisof problem behavior: A practical assessment guide.Sycamore, IL: Sycamore Press.

Phillips, E. L., Phillips, E. A., Fixen, D. L., & Wolf,M. M. (1971). Achievement Place: Modificationof the behaviors of pre-delinquent boys with atoken economy. Journal of Applied Behavior Anal-ysis, 4, 45–59.

Repp, A. C., & Karsh, K. G. (1994). Hypothesis-based interventions for tantrum behaviors of per-sons with developmental disabilities in school set-tings. Journal of Applied Behavior Analysis, 27, 21–31.

Rhode, G., Morgan, D. P., & Young, K. R. (1983).Generalization and maintenance of treatmentgains of behaviorally handicapped students fromresource rooms to regular classrooms using self-evaluation procedures. Journal of Applied BehaviorAnalysis, 16, 171–188.

Umbreit, J. (1995). Functional assessment and inter-vention in a regular classroom setting for the dis-ruptive behavior of a student with attention deficithyperactivity disorder. Behavioral Disorders, 20,267–278.

Wechsler, D. (1991). Manual for the Wechsler Intel-ligence Scale for Children (3rd ed.). San Antonio,TX: Psychological Corporation.

Weiss, G., & Hechtman, L. T. (1993). Hyperactivechildren grown up: ADHD in children, adolescents,and adults (2nd ed.). New York: Guilford Press.

Whalen, C. K., & Henker, B. (1991). Therapies forhyperactive children: Comparisons, combinations,and compromises. Journal of Consulting and Clin-ical Psychology, 59, 126–137.

Zentall, S. S., & Javorsky, J. (1995). Functional and


clinical assessment of ADHD: Implications ofDSM-IV in the schools. Journal of Psychoeduca-tional Assessment (Monograph Series: SpecialADHD Issue), 22–41.

Received March 13, 1997Initial editorial decision May 7, 1997Final acceptance October 13, 1997Action Editor, Wayne W. Fisher

STUDY QUESTIONS

1. What are some of the behavioral characteristics and risks associated with the diagnoses ofADHD and ODD?

2. How did the authors develop and test their hypotheses about behavioral function?

3. Briefly describe the treatments used with each participant and the behavior-change mecha-nisms by which these interventions probably influenced behavior.

4. What was the potential limitation of the contingency contained in Carl’s self-evaluationprocedure?

5. Results of the study showed that the interventions produced increases in on-task behavior.Although this finding is significant in light of the participants’ diagnoses, how is it limited,and what additional measures would have strengthened the study?

6. How do the results extend previous research on functional assessment for students withADHD?

7. What unique contextual features of the study may have affected the generality of the find-ings?

Questions prepared by SungWoo Kahng and Eileen Roscoe, The University of Florida