VoL 73, No. 3, pp. 288-310.©2007 Council JOT Exccplional Childrrn.

^Exceptional Children

Response to Intervention:Examining Classroom BehaviorSupport in Second Grade

5ARAH FAIRBANKS

GEORGE SUGAI

University of Connecticut

DAVID GUARDINOUniversity of Oregon

MARGARET LATHROPBethelSibtiolDistrict. Eugene, OR

ABSTRACT:r: This article reports on 2 studies investigating a response-to-intervention (RTI)

approach to behavior support in 2 second-grade classrooms. The results suggest that a slightly more

intensive but efficient targeted intervention ("check in and check out") was effective in supporting

the social behavior success of 4 students whose problem behaviors were unresponsive to general

classroom management practices. For 4 other students whose problem behaviors continued to be

unresponsive to the "check-in and check-out" intervention, more individualized and function-

based interventions were indicated and proved to be effective. The results from this research

suggest that RTI logic can be applied to the social behavior support of students who present inter-

fering problem behaviors in the classroom. Implications and recommendations for research and

practice are discussed. r

Schools are increasingly heldaccountable for their efforts toimprove the academic and so-cial behavior of their students,despite diminishing resources

to support those efforts (Eber, Sugai, Smith, &Scott, 2002; Sugai et al., 2000). In addition,many schools lack the expertise to define and usepractices and systems that meet the needs of theirstudents with both efficiency and effectiveness(Sugai et al., 2000; U.S. General Accounting Of-

fice, 2001). Further, with the advent of legislationrequiring more proactive strategies to identify andserve students with academic and social behaviorconcerns, schools may be unprepared and ill-ad-vised as to how to best implement such practices.

R E S P O N S E T O I N T E R V E N T I O N

Language in the Individuals With Disabilities Ed-ucation Improvement Act of 2004 (IDEA) aboutspecial education eligibility and assessment proce-

288 Spring 2007

dures indicates that a local education agency "mayuse a process that determines if the child respondsto scientific, research-ba.sed intervention as a partof the evaluation procedures" (Pub. L. No. 108-446 § 614, 118 Stat. 2706, 2004). This statementrepresents a considerable departure and alterna-tive to the traditional IQ achievement discrep-ancy model used to determine special educationeligibility under the learning disabilities (LD) cat-egory. The IQ achievement discrepancy modelhas been criticized for both its lack of treatmentutility (Gresham et al., 2005) and inability to ac-curately differentiate low achieving students fromstudents with learning disabilities (Fletcher et al.,1998). The response-to-intervention (RTI) pro-cess, in contrast, incorporates low-inference andfunctional assessment procedures that can link di-rectly to group and individual intervention plan-ning (Christ, Burns, & Ysseldyke, 2005).

RTI models typically are composed of a min-imum of the following components: (a) a contin-utim of evidence-based services available to allstudents, from universal interventions and proce-dures to highly intensive and individualized inter-ventions (Matston, Muyskcns, Lau, & Canter,200.3); (b) decision points to determine if stu-dents are performing significantly below the levelof their peers in academic (Vaughn, Linan-Thompson, & Hickman, 2003} and social behav-ior domains; (c) ongoing monitoring of studentprogress (Gresham, et al., 2005); (d) employmentof more intensive or different interventions whenstudents do not improve in response to other in-terventions; and (e) evaluation for special educa-tion services if students do not respond tointervention instruction (Fuchs, Mock, Morgan,& Young, 2003).

Traditionally, RTI has focused on academicconcerns as a means to identify students underthe LD category for special education services(Gresham et al., 2005). Research, generally, hasevaluated either universal and/or targeted groupinterventions, often referred to as RTl-StandardProtocol (SP; Fuchs et al., 2003) or evaluated ter-tiary level individualized intervention, sometimesreferred to as RTl-Problem Analysis (PA; Christet al., 2005). Research evaluating components ofeither RTl-SP or RTI-PA has been conductedwith elementary students with reading prohlcms(Daly, Martens, Hamler, Dool, &c Eckert, 1999;

Vaughn et al., 2003). In addition, mutticompo-nent and multiple baseline research designs andthe conceptual logic of applying interventions ofincreasing intensity, as indicated by the needs ofthe student, have been used effectively to identifythe most appropriate tertiary leve! academic or so-cial behavior interventions for children (Barnett,Daly, Jones, & Lentz, 2004).

A S O C I A L B E H A V I O R R T I M O D E L

RTI logic has intuitive appeal as a means to serveand identify students with emotional and/or be-havior disorders. Despite the lack of specific em-pirical support for RTI in the social behaviordomain, similar models of behavior support havebeen implemented in schools. Such models havebeen based on principles of wraparound behaviorsupport (Eber et al., 2002) and/or the inclusionand integration of graduated systems of behaviorsupport (Sugai et al., 2000). A social behaviormodel of RTI promises to be an extension andnew application of the already substantial researchbase regarding positive behavioral interventions,functional behavior assessment (FBA), and earlyintervention (Sugai et al., 2000; Vatighn et ai.,2003).

UNIVERSAL INTERVENTION

Applying RTI logic to social behavior supportcould require a standard-protocol approach foruniversal and targeted group level interventions(sometimes referred to as Tier 1 and Tier 2 inter-ventions). The universal system, implementedschoolwide for ali sttidents, might require schoolsto identify and explicitly teach schoolwide expec-tations; implement a system to acknowledgeexpectation-compliant behavior; define and con-sistently apply consequences for inappropriate be-havior; and regularly review progress towardsschooiwidc goals. Such a universal system reflectsthe features of schoolwide positive behavior sup-port (SW-PBS; Lewis & Sugai. 1999). Numerousstudies indicate that teaching expectations acrosssettings and providing incentives fot appropriatebehavior (within the SW-PBS framework) can ef-fectively reduce student problem behavior (Kar-tub, Taylor-Green, March, & Horner, 2000;Lewis, Powers, Kelk, & Newcomer, 2002;

lixceptional Children 2 8 9

Metzler, Biglan, Rusby, & Sprague, 2001; Satran& Oswald, 2003).

Employing evidence-based classroom man-agement strategies may also serve as a universallevel preventative intervention. For example, con-sistently implementing an acknowledgment sys-tem to recognize appropriate behavior in class,providing multiple and varied opportunities forstudents to respond during instruction, minimiz-ing transition time between classrootn activities,and ptoviding direct and immediate correctivefeedback for social or academic behavior errorsmay provide an excellent universal level founda-tion from which to identify students who may re-quire more specific intervention supports.

TARGETED INTERVENTION

Students who do not respond as expected to theuniversal level or Tier 1 intervention may receivetargeted or Tier 2 interventions. Tier 2 interven-tions typically provide targeted instruction focus-ing on the development of specific skills for agroup of individuals engaging in similar error pat-terns. For example, students who are accurate butslow readers might receive a targeted fluency-based intervention. Within a social behavior RTIlogic, corollaries to such targeted reading inter-ventions have been examined, for example, "checkin and check out" (CICO) and behavior educa-tion program (BEP). The CICO interventionprovides additional structure, prompts, instruc-tion, feedback, and acknowledgment for studentsengaging in similar low-level social behavior er-rors (Filter er al., in press; Hawken & Horner,2003; Todd, Kauffman, Meyer, & Horner, inpress). There are numerous variations of CICOinterventions, such as Daily Behavior ReportCards (Chafouleas, McDougal, Riley-Tillman,Panahon, & Hilt, 2005), but what is generallyconsistent across most variations is the efficiencywith which an intervention can be implementedand its application to groups of students. Incor-porating a targeted social behavior interventioninto a social behavior RTI logic also fits seam-lessly within the SW-PBS framework.

INDIVIDUALIZED INTERVENTION

Students who are unsuccessful in response to atargeted intervention may experience Tier 3 or

RTI-problem analysis (Christ et aL, 2005), spe-cific and time-intensive as.sessments to determineindividual skill deficits and to assist in the designof an individual intervention. Depending on a va-riety of factors, evaluation for special educationeligibility may also ensue. Functional behavior as-sessment (FBA) might be deemed comparable toa Tier 3 assessment for reading. FBA is a processused to determine events that reliably predict andmaintain behaviors of concern (Horner, 1994;Sugai, I-ewis-Palmer, & Hagan, 1998). The pro-cess is a widely supported assessment procedurespecifically mentioned in IDEA, used to informbehavioral intervention, and theoretically an-chored to applied behavior analysis and PBS(Sugai et al., 2000). FBAs require collecting andanalyzing various forms of indirect, descriptive,and experimental assessments to deduce a plausi-ble hypothesis which can be tested and whicbidentifies the likely conditions under which be-haviors of concern occur.

The FBA is of little value by irself, unless it isused to inform an intervention plan. Because it isused to design a function-based intervention, theFBA-based hypothesis statement—which include.sboth desired behaviors and acceptable alternativebehaviors or benchmarks toward desired behav-iors—is usually stated clearly within the plan. Themajority of the function-based behavior interven-tion plan is then devoted to outlining strategies to(a) change antecedent conditions likeiy to precedeproblem behavior, (b) teach prosocial behaviorseffective in accessing the same consequences asproblem behaviors, (c) decrease access to desiredconsequences following problem behavior, and(d) increase access to desired consequences follow-ing appropriate behavior (Crone & Homer,2003). The plan also includes implementationtasks, deci.sion rules to modify implementation,and demographic information.

The FBA is of little value by itself unless itis used to inform an intervention plan.

Function-based support—the practice oflinking FBA information to the design and imple-mentation of behavior intervention plans—hasresulted in positive outcomes for individuals with

2 9 0 Spring 2007

developmental and intellectual disabilities (Carr,1977; Iwata, Dorsey, Slifet, Bauman, & Richman,1982) and individuals with emotional and behav-ior disorders (Lewis & Sugai, 1993; Vollmer &Northup, 1996).

RTI activities fit easily within SW-PBS'sthree-tiered prevention logic (Sugai et al., 2000).RTI represents a process that fiicilitates access toappropriate levels of both academic and social be-havior intervention support. Althotigh intensify-ing levels of behavior support can be defined, thedependent measures and decision rules schoolsuse to identify "low responders" to social behaviorinterventions are not as well established. Further-more, eligibility for special education services andmore intensive, wraparound behavior support ser-vices must also be considered (Eber et al., 2002;Epstein et al., 2005).

R E S E A R C H Q U E S T I O N S

Study 1 evaluated a CICO targeted interventiondesigned with minimal researcher consultationand implemented by two second-grade teachersfor 10 students. Study 2 consisted of an experi-mental evaluation of the effects of individualizedfunction-based support for 4 participants whosebehaviors did not respond as expected to theCICO intervention. Study 2 interventions weredeveloped collaboratively by school personnel andresearchers and were implemented by one second-grade teacher. The current study addresses twooverarching research questions:

Study 1: Does a relationship exist between im-plementation of a CICO targeted interven-tion and (a) percentage of intervalsparticipants were observed to be engaged inproblem behavior, (b) frequency of office dis-cipline referrals, and (c) teacher perceptionsof problem behavior intensity and frequency?

Study 2: Does a functional relationship exist be-tween implementing function-based behaviorintervention plans and reductions in (a) per-centage of intervals participants were ob-served to be engaged in problem behavior, (b)freqtiency of office discipline referrals, and (c)teacher perceptions of problem behavior in-tensity and frequency?

S T U D Y 1 : M E T H O D

SETT INC

The study took place at a public elementaryschool within a suburban school district servingapproximately 5,500 students in a small city inthe northwestern United States. The elementaryschool was implementing SW-PBS (Sugai et al.,2000) effectively (100% of components) based onscores obtained from the School-wide EvaluationTool (SET), a reliable and valid assessment toolmeasuring PBS implementation accuracy and fi-delity (Horner et al., 2004).

PARTICIPANTS

The principal and second-grade teachers at theschool contacted researchers because of concernsrelated to increasing numbers of office disciplinereferrals and overall disruptive behavior in theirtwo classrooms. The principal, two second-gradeteachers, and the coun.selor met with researchersand discussed concerns and possible solutions tothe reported increases in problem behavior intheir classrooms. Hie teachers then nominatedstudents in each classroom to take part in thestudy. Parents of the participants were sent lettersof consent regarding the study and their child'spotential involvement in the study. Students alsoprovided their assent to be involved in the study.

Ten children between the ages of 7 and 8 par-ticipated in the study. All participants received in-struction in a general education setting in one oftwo second-grade classrooms. Table 1 provides asummary of student characteristics (pseudonymsused) and their Dynamic Indicators of Basic EarlyLiteracy Skills Oral Reading Fluency (DIBELSORF; Good & Kaminski, 2002) rneasures in falland spring. All the participants are described asone group across the two classrooms becatise bothteachers worked collaboratively and observationand intervention procedures were the same. An-other male initially participated in the study butmoved away during the beginning of the study;his data are not included. Helena also moved outof the school neighborhood toward the end of thestudy, and her data are included. Farrell and Mar-cellus were receiving behavioral interventions atthe time of the study's implementation. Six of 10participants met DIBELS standard benchmarks

Exceptiotml Children

TABLE 1

Participant Demographic Information

Participants

Classroom A

Chase

Randy

Isabel

Classroom B

Heletia

Jade

Farrell

Marcel 1 us

Blair

Ben

Olivia

Gender

Male

Male

Female

Female

Female

Male

Male

Female

Maie

Female

Ethnicity

Caucasian

Caucasian

Native Ameticati

Caucasian

Caucasian

Africati American

Caucasian

Caucasian

Caucasian

Caucasian

Medication

Concerta

No

No

NoNoAddetall

NoNoNoNo

Special

Education

Eligibility

No

NoNo

NoLDNo

No

NoNo

LD

DIBELS ORF Scores

Eall

27

52

9

492965

81

80

7418

(wrc)

Spring

79

100

. 0 (Jan.)

80 (Jan.)

59158

103

110

116

60

Note. DIBELS ORF = Dynamic Indicators of Basic Early Literacy Skills Oral Reading Fluency; wtc = wotds read

correctly pet min; LD = learning disability.

for fall (i.e., 44 words read correctly = low risk)and 5 of 8 participants (2 participants hadmoved) met benchmarks for spring (I.e., 90 wordsread correctly - low risk).

Due to classroom and parental factors. Chaseand Isabel received only the CICO intervention,even though the two were considered candidatesfor function-based behavior support. Chase andIsabels data provide an uiditect peer comparisonbetween participants who successfully respondedto CICO and participants who were not respon-sive to CICO, but were responsive to function-based behavior support.

Durmg each observation, two randomly cho-sen peers (who served as composite peers) werealso observed to gauge typical problem behaviorlevels in the classroom beyond primary partici-pants and to discern what acceptable levels ofproblem behavior in the classroom were. One in-dicator of participants' responsiveness to interven-tions was how their data compared to compositepeer data.

DESIGN

A descriptive quasi-experimental design was usedto study CICO implementation and effectiveness.

Time-series data were collected on each studentacross five phases: baseline, CICO 70%, CICO75%, CICO 80%, and CICO 90% of points.These phases represent an increase in the percent-age of paints required of students on the CICOplan to earn a classroom reward. For example, inthe 70% phase student participants had to earn70% of possible points to earn the classroom re-ward. Because CICO is presented and managed asa group-contingency intervention, all students re-ceived the intervention at the same time, prevent-ing staggered implementation across students.

MEASUREMENT

Primary Dependent Variables. The observa-tional dependent variables for the study were per-centage of intervals engaged in (a) inappropriatephysical contact, (b) talk-outs, (c) inappropriateplacement, (d) noncompliance, (e) nondisruptiveoff-task behavior, and (f) academic engagement.With the exception of academic engagement, allof the behaviors were used as an index of overallproblem behavior. Tn other words, an interval wasreported as a problem behavior interval if one ormore of the problem behaviors occurred duringan interval; otherwise the participant was consid-

2 9 2 Spring 2007

ered academically engaged. The nondisruptiveoff-task code assisted data collectors in differenti-ating participants who were not engaged in obvi-ous overt problem behavior, but who wereoriented away from the task at hand, from partici-pants who were academically engaged. Oniyptoblem behavior (a composite of all the problembehavior codes) is reported in the results seaion.

Inappropriate physical contact was separatedinto three categories and was defmed as any vol-untary action, attempted or actual behavior that(a) could cause injury to oneself or another indi-vidual (physical assault), (b) could cause damageto an object and was not in compliance withteacher directions (inappropriate physical conractwith an object), (c) or involved physical contactwith another person and was not in compliancewith teacher directions (inappropriate physicalcontact with a peer). Talk-outs were defined as anyverbal utterance that interrupted teacher or stu-dent directions, comments, or questions withoutthe student being called on or asked a questiondirectly. Talk-outs could be directed towards apcer(s), a teacher, or oneself Inappropriate place-ment was defined as losing contact with the seatsurface, all tour legs of a chair off the floor, orstanding when expected to sit. Noncompliance wasdefined as not following classroom rules or expec-tations or doing something other than complyingwith an adult's directive within 5 s. Nondisruptiveoff-task behavior was defined as being orientedaway from the task ar hand for more than 3 s ofan interval. Academic engagement was defined asorientation toward the task at hand, compliancewith all directions, and workmg with appropriatematerials.

Observers additionally recorded contextcodes denoting concurrent activities within theclassroom, including (a) teacher-led instruction(TLI), (b) independent seat work (ISW), (c) co-operating work group (CWG), (d) free-choicetime (FCT), and (e) transition time (TT).

Data Collection. The first and third authors,who had extensive experience observing problembehavior, collected observational data. Initially,they spent time in the two classrooms identifyingthe most frequent and intensive occurring prob-lem behaviors and testing different observationsystems. Then, operational definitions of behav-iors of interest and an observation system were

developed and practiced. Together, the observersreviewed operational definitions and conductedobservations to ensure initial observations werereliable. The observers also frequently consultedwith the second and fourth authors to clarify andimprove the consistency of definitions and obser-vation procedures across both classrooms.

The primary observational dependent vari-ables were measured using a 10-s partial intervalrecording system for 40-min observations. Allparticipants in the CICO progratn and two ran-domly selected peers were observed for 1-min in-crements. Before conducting an observation,observers selected a code to correspond to eachparticipant. The corresponding participant wouldbe observed for 1 min every time tbe preselectedcode occurred. Digital voice recorders with singleearphones were used to cue observers for the be-ginning of intervals. FLach classroom was observed2 to 4 times a week, at times during which theteacher had indicated problem behavior to bemost intense and frequent.

Observer Agreement for Primary DependentVariables. Interobservet ^reement data were col-lected during approximately 25% of observationsand at least once per phase. The primary observerwas the individual who regularly observed in theclassroom. Agreement was determined by com-paring each interval of the primary and secondaryobservers' data sheets. Two intervals were consid-ered in agreement if, and only if, the same prob-lem behaviors or academic engagement weremarked by both observers. The percentage of totalagreement was calculated by dividing the totalnumber of agreements by the total number ofagreements plus disagreements multiplied by100%. Interobserver agreement averaged 90%(range, 76%-100%),

Secondary Dependent Variables. To evaluatethe teacher's perception of the participant's behav-ior, a 5-point, 5-item rating scale was developedbased on the context and specific interventionsthat were implemented. The rating scale was ad-ministered once before, twice during, and onceafter the study's implementation. The classroomteacher was asked to rate the (a) overall intensityand frequency of problem behavior in the class-room (2 items), (b) intensity and frequency of theproblem behavior of participants in the CICOprogram (1 item), and (c) extent to which the

Exceptional Children 2 9 3

F I G U R E 1

Sample Check-In and Check-Out Card

Name.

Date

POINT SHEET

Rating Scale

2 = Great1 =0K0 = Goal not met

Points Possible

Points Received

Percentage of Points.

Goal Met? Y N

GOALS:

1. RESPECTOTHERS

2. MANAGESELF

3. SOLVEPROBLEMSRESPONSIBLY

8:30-9:30

2

2

2

1

1

1

0

0

0

9:30

2

2

2

•10:30

1

1

1

0

0

0

10:30-11

2

2

2

1

1

1

:30

0

0

0

11:30-12:30

2

2

2

1

1

1

0

0

0

12:30-1:30

2

2

2

1

1

1

0

0

0

1:30-2:30

2

2

2

1

1

1

0

0

0

CICO program was considered to reduce prob-lem behavior of student participants and theirpeers in the classroom (2 items). When the ratingscale was administered before the interventionwas implemented, items addressed the teacher'sconfidence that CICO would positively impactstudent behavior.

Office discipline referral information was an-alyzed before, during, and after the study's imple-mentation. Based on the schools procedures andpolicies, office discipline referrals were categorizedas major or minor behavior violations. Major be-havior violations were given for abusive language,fighting, harassment, and defiance. Minor behav-ior violations were given for disruption, propertymisuse, inappropriate language, and noncompli-

ance.

PROCEDURES

Baseline: SW-PBS and Classroom Manage-ment. In the second-grade classrooms, SW-PBSwas linked to the schoolwide system by acknowl-edging appropriate behavior using the schoolwidepositive behavior "respect" tickets, teaching andreferring frequently to schoolwide expectations,and providing predictable consequences for ruleinfractions. The PBS system was maintainedthrough all phases of the study. Six observationswere conducted during baseline. Office discipline

referral data were collected throughout the studyand teacher perception ratings were collected onceduring baseline.

Check-In and Check-Out Intervention. A tar-geted CICO group intervention was designed andimplemented by the two second-grade classroomteachers. Based on Hawken and Horner's (2003)guidelines, the CICO program provided studentswith (a) increased structure and prompts, (b) ad-ditional instruction on specific skills, and (c) in-creased regular feedback. The researchers andother teachers who had experience with imple-menting CICO interventions provided the sec-ond-grade teachers with suggestions about criticalfeatures and implementation guidelines.

The CICO cards (see Figure 1) were thesame for all participants and restated schoolwiderules as individual "goals" ("respects others,""manages self," and "solves problems responsi-bly"). Students had the opportunity to receive 36points throughout a school day based on their be-havior during six discrete 60-min time periods.Teachers rated each participant's behavior at theend of each designated time period (0 = goal notmet, 1 = okay, and 2 - great). Students were alsogiven feedback about their behavior, usually inthe form of specific praise or corrective feedback.Participants carried their CICO card to all classeson a small clipboard.

Spring 2007

Students with the CICO cards were identi-fied as "leaders" and the rest of the students in theclass were called "coaches." Coaches were asked tohelp the leaders stay on task and follow theschoolwide expectations. At the end of the day,each leader added tip his or her points and sharedthe point total with tbe whole class. If the leaders'cumulative total points met or surpassed a speci-fied percentage of the total points possible (range,70%-90%), the class earned a reward such as aclass game, free choice time, an extra recess, or apiece of candy. At the beginning of tbe CICO im-plementation, leaders were required to earn 70%of possible points. Over the duration of the studythe criterion was increased to 75%, 80%, and 90%of points. Teachers decided when to increase thepercentage of points required to earn the class re-ward. When leaders met criteria consistently (e.g.,5-6 days in a row), teachers raised the criteria.

Participants in Classroom A were observedon approximately 31 occasions {unless they wereabsent) during CICO phases. During CICOphases, in Classrootn B, participants receivingfunction-based support were observed 21 timesand participants receiving CICO throughout thestudy were observed 57 times. Ben, who enteredthe study late, was observed on 8 occasions dur-ing CICO phases. The first CICO phase lasted 1week, the second phase lasted approximately 8.5weeks (including winter break), the third phaselasted 2 weeks, and the fourth phase lasted untilthe end of the study or 16 weeks.

Fidelity of Implementation. Approximatelyonce a week at different points throughout theday, the school counselor would visit each class-room to assess the accuracy ot the CICO imple-mentation. A fidelity checklist was developed sothat each of the nine critical activities of theCICO program could be checked as occurring,not occurring, or unclear. To calculate the per-centage for implementation fidelity, the numberof components that were implemented was di-vided by the number of components imple-mented plus the number of components notImplemented, and multiplied by 100%. Thecomponents that the school counselor identifiedas unclear were not included in the fidelity calcu-lation because it generally meant that the coun-selor was not present to observe whether anactivity occurred during a specified time period.

Fidelity data wete collected on 8 occasionsfot Classroom A participants (Chase, Randy, Is-abel); on average, the CICO program was imple-mented with 88% accuracy (range, 50%-100%).Fidelity data were collected on 14 days for Class-room B participants (Helena, Jade, Farreli, Mar-cellus, Blair, Ben, Olivia); on average, the CICOprogram was implemented with 94% fidelity(range. 82%-100%).

ANALYSIS

To evaluate the effects of the CICO intervention,researchers conducted visual analyses of grapheddata and considered variability, trends (see Figure2), average rates of responding, immediacy of ef-fect, composite peer data, and overlap of datapoints within and across phases. Visual analysiscriteria were based on the guidelines of Horner etal. (2005) and Parsonson and Baer (1986). Partic-ipant behavior was considered responsive to inter-vention if observable, sustained, and socially validreductions in problem behavior were observed, es-pecially in relation to problem behavior rates ofcomposite peers.

STUDY 2: METHOD

PARTICIPANTS AND SETTING

Based on direct observation data and teacher andcounselor nomination, four students (Blair, Ben,Marcellus, and Olivia) in Classroom B were iden-tified as having behaviors unresponsive to theCICO intervention and requiring more individu-alized intervention. "Unresponsiveness" was de-fined as (a) little change in overall rates of problembehavior, (b) increasing trends in rates of problembehavior, and/or (c) continuation of serious dis-ruptive problem behavior. The second-gradeteacher in Classroom B and school counselor com-pleted the Functional Assessment Checklist for'Ibachers and Staff (FACTS; March et. al. 2000),and identified individual student strengths, behav-iors of concern, routines in which problem behav-iors were likely to occur, and conditions in whichthe behaviot of concern were likely to occur (i.e.,setting events, antecedents, and consequences).After completing the FACTS, the second-gradeteachets, school counselor, other school personnel

Exceptional Children

F I G U R E 2 . , . i

Percentage of Intervals Engaged in Problem Behavior: CICO Responders

2

ICO

E-9?o(£

oCDroO)c

LU

o

0O

I

• * • - * • — ^

r. ^

Randy

Farroll |

\>-V;v-%--{School Days

with whom the participant had contact, and oneof the researchers developed a fiinction-based indi-vidualized behavior support plan. The researchers

Each behavior support plan included infor-mation about (a) student strengths, (b) behaviorof interest, (c) setting events and antecedents, (d)

also provided summaries of some observational perceived maintaining consequences, (e) alterna-data and minimal consultation. tive behaviors, and (f) desired behaviors (see

2 9 6 Spring 2007

TABLE 2

Problem Behavior Contingencies Across Students

Student Setting Events Antecedents Behaviors of Interest Consequence

Marcellus N/A (a) Easy unstructuredactivities

(b) Difficult tnath andwriting tasks

(a) Out of seat andmaking faces

(b) Talk-outs, out of seat,poor work completion

(a) Peer attention

(b) Escape work

Blair N/A Independent wotktime

Out of seat and talkingto peers

Peer and adultattetition

Ben N/A Teacher-ledinstruction

Teachet direction

Noncompliance,talk-outs, making jokes

Peer and adultattention

Olivia Thinking about Teacher-ledthe loss of her instructionsibling

Playing with things,looking away trotn theteacher, not followingdirections

Adult attention

Tables 2 and 3). Alternative behaviors were se-lected that (a) were acceptable in the teacher'sclassroom, (b) would allow the participant accept-able access to the same consequence as the prob-lem behaviot, and (c) were comparable to theptoblem behavior with respect to effort and effi-ciency. The selection of the desired behavior wasbased on the teacher's expectations ior other stu-dents. Behavior support plans included interven-tion and iristructional strategies designed toconsider the influence of setting events, an-tecedents, and maintaining cotisequences, and de-tailed teaching and strengthening reinforcingconsequences for alternative and desired behav-iors. Additionally, rhe plan outlined implementa-tion tasks, data collection methods, long- andshort-term goals, and criteria for intervention de-cision making.

DESIGN

Observers collected time-scries data on each stu-dent across five or six phases: (a) baseline, (b)CICO 70%, (c) CICO 75%, (d) CICO 80%, (e)FBA-based plan, and (f) FBA-based plan-ad-justed. Tbe participants in tbis study did not ex-perience the 90% phase. Participants continued at

the 80% phase until their FBA plan was imple-mented. Function-based plan implementationwas staggered, and a multiple baseline designacross four participants was used to examine theeffectiveness of function-based plans.

MEASUREMENT

Primary Dependent Variables. The observa-tional dependent variables were percentage of in-tervals engaged in (a) inappropriate physicalcontact, (b) talk-outs, (c) inappropriate placement,(d) noncompliance. (e) nondisruptive off-task be-havior, and (f) academic engagement. Defmitionswere identical to those provided in Study 1.

PROCEDURES

Data collectors from Study 1 used the same ob-servation forms and procedures with a few excep-tions. A 10-s partial interval recording system wasused fot 30-min obsetvation sessions, conducted3 to 4 times a week. All students receivitig func-tion-based support and a peer were observed dur-ing every observation. Each student whocontinued to receive CICO but was not receivingfunction-based support was observed approxi-mately every third observation.

Exceptional Children 2 9 7

TABLE 3

Individualized Function-Based Intervention Components by Participant

SettingStudent Events Antecedents

Marcellus N/A Cbange seating

Sit by peet of choice

Check-in/out

Continue basic programalready in place

Modify amount and/ordifficulty of work

Can choose to do less work

and take remaining workhome to do with Mom

Precorrectians

Remind Marcellus of(a) choices at beginningoftnath, (b) peer selection,and (c) options at the endof matfi

Behaviors of Interest

'Ieach/rei'iew expectations

Teacher to review choicemaking for math, peer, andend-of-period activity

Consequence

Response to desired behavior

1. Teacher to provide atleast I praise statementper 10 min

Privilege/responsibility:

2. End-of-period aaivitychoice option (e.g..computer time)

3. Usual check in/out at

end of day (goal: 80%of points)

4. At end of day, teacher

reminds Marcellus ofwhat and how muchwork needs to be com-pleted at home

Response to problem behavior

1. Point loss

2. Reminder to return toseat

3. Returti to own seataway from peer

Blair N/A Check-in/out (attachsummary & point sheet)

Basic classroom plan.Modify classroom pointsheet to include under"Manage Self":

1. Staying in seat

2. Talking only when it \s,okay

3. Following directions thefirst time

Teach/review expectations

1. Counselor will reviewwith Biait the expecta-tion that she followdirections the first timegiven

2. If Blair docs not followdirections teacher willsay, "Blair, I gave you adirection and youdidn't do as 1 asked, youneed to put yotir headdown." After 10 s.teachet will follow upwith a second request todo as eatlier instruaed

3. If Blair still does not doas instructed she will geta check on her pointsheet for self manage-ment

4. Teach specific socialskills/scripts (See above)

Response to desired behavior

1. Verbal praise

Privilege/responsibility:

2. Blair earns recess andreward time at end ofday by making herpoints

3. Usual check in/out withclass at end of day (goal:80% of points)

Response to problem behavior

1. Point loss: Points ate to-taled before lunch andagain at the end of theday Blair needs to earn14/18 points to havemorning recess or pat-ticipate in end-of-dayclass reward

continues

298 Spring 2007

T A B L E 3 (Continued)

StudentSettingEvents Antecedents Behaviors of Interest Consequence

Ben N/A Check-in/out (attachsummary & point sheet)

Basic classroom plati.Modify classroom pointsheet to include under"Manage Self":

1. Staying in seat

2. Talkingonly when it is

okay3. Followihg directions the

first time

Olivia

Teach/review expectations

1. Coun.sclor will reviewwith Ben the expecta-tion that he follow di-tectlons the first timegiven.

2. If Ben docs not followdirections, teacher willsay. "Ben. I gave you adirection and you didn'tdo as 1 asked, you needto put your head down."After 10 s. teacher willfollow up witb a secondrequest to do as earlierinstructed.

3. If Ben still docs not doas instructed he will geta check on his pointsheet for self-manage-ment

4. Teach speciBc socialskills/scripts (See above)

Response to desired behavior

1. Vctba! praise

Privilege/responsibility:

2. Ben earns his recess andreward time at the endof the day by makinghis points

3. Usual check in/out with

the class at end of day(goal: 80% of points)

Response to prohlevi behavior

I. Point loss: Points aretotaled before luncband again at the end ottbe day. Ben needs tomake 14/18 pt.s to havemorning recess or par-ticipate in cnd-of-dayclass reward

Grieftherapygroup

Change seating

Increase adult proximity

Check-in/out (attachsummary & point sheet)

Basic classroom plan.Modify cta.ssroom pointsheet to include under"Respect Others":

1. Say nice tbings or nothitigs

2. Look at tbe teacherduring instruction

3. Be a good listener

4. Have empty hand.s

Teach/review expectations

Counselor will review ex-pectations for respecting

others:1. Say nice tbings or no

things2. Ijjok at the teacher

during instruction

3. Be a good listener

4. Have empty hands

Response to desired behavior

1. Verbal praise

2. Usual check in/out witbthe class at end of day(goal: 80% of points)

3. Color spots for beingrespectful to the teacbcron a spot card.

4. Wben Olivias spot cardis ftil! she can choose to;

-Have lunch in theroom with the teacher

-Color with thecounselor

-Take a friend to thecounselor's office toplay

- Get a treat ftom theprincipal or counselor

Response to problem behavior

I. Lose points on pointsheet

2. Does not participate inclass reward if 80% ofpoints are not earned

Exceptional Children 2 9 *

Observer Agreement for Primary DependentVariables. As in Study 1, ititerobserver agreetnentdata were collected and summarized for approxi-mately 25% of observatioti sessions. Interobservetagreement averaged 91% (range, 85%-100%).

Baseline and CICO. Marcellus. Blair, andOlivia received the same baseline and CICO con-ditions described in Study I. Olivia was observedmore sporadically because she was not initiallyconsidered for function-based support, Ben en-tered the classtoom late and became involved inthe study during the 80% CICO phase.

Participants were observed between 19 and35 times during tbe function-based behavior sup-port phases; differences in observation numbersare due to the staggering of intervention imple-mentation. The fiinction-based behavior suppottphase lasted 16 weeks for the first participant.Variations in length of the function-based behav-ior support phase are also due to the staggering ofintervention implementation for different pattici-pants.

Fidelity of Implementation. To examine tbeextent to which function-based support planswere accurately Implemented, researchers devel-oped individualized fidelity checklists that re-flected components for each student's behaviorsupport plan. During regularly schedtiled obser-vations, observers would check whether each ele-ment occurred, did not occur, occurredintermittently or partially, or would not necessar-ily occur (depending on the time of the observa-tion). On average. Marcellus's plan wasimplemented with 84% fidelity (range, 67%-100%), Blair's plan was implemented with 7G%fidelity (range. 60%-100%), Ben's plan was im-plemented with 82% fidelity {range, 60%-100%). and Olivia's plan was implemented with81% fidehty (range, 80%-83%). Across students,fidelity of implementation averaged 80.75%. Toexamine the accuracy of these fidelity checks, asecond observer completed the same checklistduring 28% of Marcellus's observations. 31% ofBlair's observations, 50% of Ben's observations,and 100% of Olivia's observations. Agreement orthe number of agreements divided by the numberof disagreements plus agreements, multiplied by100% was averaged at 92% (range, 87%-100%).

ANALYSIS

Researchers applied the same visual analysis pro-cedures used in Study 1, and examined variability,trends (see Figure 3), average rates of responding,immediacy of effect, composite peer data, andoverlap of data points within and across phases.

RESU LTS

All 10 participating students experienced theCICO intervention. The problem behaviors offour of these students were responsive to CICOand did not requite more intensive, tertiary levelinterventions. Of the 6 students whose behaviorswere not responsive to CICO, 4 students receivedmore individualized function-based interventions.Two students (Chase and Isabel) remained in theCICO condition and served as peer controls;their data (see Figure 4) provide an indirect peercomparison between participants who successfullyresponded to CICO and participants who werenot responsive to CICO. but were responsive tofunction-based behavior support. The results forthese groups of participants are described below.

CICO RESPONDERS

Ten students were nominated as needing moresupport than available through the generalschoolwide and classroom management proce-dures, and CICO procedures were implemented.Observation data for 4 of these students indicatedresponsiveness to CICO (see Figure 2). A visualanalysis of baseline data for Randy (mean inter-vals with problem behavior, 32% range. 0%-63%) and Farrell (mean intervals with problembehavior, 28% range. 5%-38%) revealed increas-ing levels of problem behavior, compared to com-posite peer comparisons (mean, 19% range,6%-58%; see Figure 5). An analysis of baselinedata for Helena (mean. 59% range. 42%-83%)and Jade (mean, 51% range, 24%-67%) indi-cated high levels of problem behavior. WhenCICO was initiated, the problem behaviors ofeach of these students were immediately decreased(level change and decelerating trends), and main-tained at low rates. Mean intervals with problembehavior during CICO for Randy were 20%(range. 0%-33%), Farrell 18% (range, 0%-

3OO Spring 2007

F I G U R E 3

Percentage of Intervals Engaged in Problem Behaviors: CICO Low Responders With Function-Based Supports

o

ICDE

cnCO

cUJ

0Q_

I MarcelluM

School Days

Exceptional Children aoi

F I G U R E 4

Percentage of Intervals Engaged in Problem Behavior: CICO Peer Controls

gICO

En2

X>

cncacnc

LU

"ocuDl

c0)

o

I

\JtJ •

90 -

80

70 •

60

50

4 0 -

30 -

20

1 0 -

BL

•

Chase

100 1

90

CI/CO CI/CO+75%

CI/CO-••80%

CI/CO+ 90%

School Days

61%), Helena 14% (range. 0%-28%), and Jade6% (range, O%-19%). All rates of problem be-havior during CICO phases were lower thancomposite peers (mean, 22% range, 8%-42%).Farrell's rates of problem behavior increasedslightly during the 80% criterion phase, at whichrime he also began taking a new medication.However, rates of problem behavior continued todecrease in association with the introduction of

the 90% criterion phase. In summary, average in-tervals with problem behavior were reduced notonly from baseline to CICO phases for all partici-pants, but were lower than composite peer com-parisons.

CICO + FUNCTION-BASED SUPPORT

Four students whose behaviors were not suffi-ciently responsive to CICO received individually

Spring 2007

F I G U R E 5

Percentage of Intervals Engaged in Problem Behaviors: Peer Composite and CICO Responders

o>COx:CD

E

o

aic

d)O)(0O)cLJJ

rval

.ffic—"o

D)

"cQ)

100 -

0 0 •

eo -TO •

B O

G O

t o

3 O

20 '

1 0

0 •

1CX»

BO -

8 0

7 0

«O -

eo

School Days

Exceptional Childrvn SOS

developed and staggered function-based support(multiple baseline design; see Figure 3). Marcellusand Blair had similar patterns of respondingacross phases. Compared to baseline, their rates ofproblem behavior during CICO were undifferen-tiated. During baseline Marcellus engaged inproblem behavior a mean of 38% of intervals(range, 19%-78%) and Blair a mean of 40%(range, 23%-69%) of intervals. During CICOphases, Marcellus engaged in problem behavior amean of 34% of intervals (range, ll%-64%) andBlair a mean of 30% of intervals (range, 14%-40%). Wirhin the CICO conditions, their ratesof problem behaviors indicated increasing trends.When tunction-based interventions were intro-duced, immediate level changes and relatively flattrend lines were noted in rates of problem behav-ior. During the first function-based phase Marcel-lus engaged in problem behavior a mean of 21%of intervals (range, 6%-39%) and Blair a mean of17% of intervals (range, 0%-33%). Mean inter-vals with problem behavior for composite peerswere 21% (range, 3%-64%; see Figure 6). Afteracceptable rates of problem behavior were ob-served (similar to peers), the plans for both stu-dents were modified to increase self-managementexpectations. Intervals with problem behaviorwere reduced to a mean of 10% (range,O%-25%) for Marcellus and 12% for Blair(range, 0%-39%).

Permission to include Ben in this study wasobrained about halfway through the study. Be-cause of previous displays of problem behaviors,his teacher and the counselor immediately intro-duced him to the CICO intervention, whichproved to be unsuccessful in improving his behav-iors (mean intervals with problem behavior equalto 29% range, 13%-47%). A level change wasimmediately observed and maintained when hisfunction-based plan was implemented (mean in-tervals with problem behavior equal to 10%range, 0%-20%).

Olivia's problem behaviors initially re-sponded to the CICO condition. Mean intervalswith problem behavior during baseline for Oliviawere 41% (range, 36%-50%) and 25% acrossCICO phases (range, 0%-40%). However, overtime her rates of problem behavior increased, ap-proaching rates comparable to baseline. When afiinction-based plan was introduced, a downward

trend was indicated and relatively low rates main-tained until the end of the school year (mean in-tervals with problem behavior equal to 16%range, 0%-47%).

Mean intervals wirh problem behavior werereduced during fiinction-based phases for all par-ticipants as compared to baseline and CICOphases. Additionally, mean intervals with problembehavior were also lower than peer compositecomparisons.

CiCO PEER CONTROLS

Classroom and parental circumstances did notprovide opportunities for function-based inter-ventions to be implemented for Chase and Isabelwhen their behaviors proved to be unresponsiveto CICO. Mean intervals with problem behaviorduring baseline for Chase and Isabel were equal to41% (range, 28%-54%) and 28% (range, 0%-58%) and during CICO phases, 37% (range,7%-80%) and 20% (range, 3%-45%) respec-tively. Tbeir data are included in Figure 4 as anindirect peer comparison. Across the CICO con-dition. Chase's rates of problem behavior weresystematically increasing, and Isabel's rates wereundifferentiated within and across neighboringphases and conditions.

CLASSROOM OEEICE DISCIPLINE

REFERRALS

The number of office discipline referrals (majorsand minors) was averaged to attain the rate of of-fice discipline referrals per instructional day in thesecond-grade classrooms. During September, Oc-tober, and the first half of November the averagenumber of referrals per day eqtialed .85. After theCICO intervention was implemented (during thelast half of November through June), the averagenumber of office discipline referrals per dayequaled .41 (see Figure 7). Additionally, beforeCICO intervention implementation, both teach-ers rated problem behavior intensity and fre-quency across all students as a 4 or higher on ascale of 1 to 5 (5 being extremely intense or ex-tremely frequent and 1 not intense or not frequent).

Following CICO implementation, teachers ratedproblem behavior as a 3 or lower.

Spring 2007

F I G U R E 6

Percentage of Intervals Engaged in Problem Behaviors: Peer Composite and Function-BasedIntervention Participants

9

I

•DOtocn

LU

CO

BS"oCDD)ro

School Days

Exceptional Children

F I G U R E 7

Office Discipline Referrals (Major and Minor) by Month

™ 1.6 n Before-- ^ ^ Interwntionc '-^oZ 1.2

1 1 •

J5 0 . 8 -Q.

"i °' "^ 0.4

B °- '3) 0 -to

CC

1

111After Intervention .

jjllxllJ *CJ^ -^ •v'ff <^ <v fc/S" ^T*

<v --*J O^^ KIIT Y " ^ ^ I " N*^

Month I

D I S C U S S I O N

OVERVIEW OF RESULTS

Given rhe increasing number of demands (e.g.,high-stakes assessments, differentiated instructionfor diverse learners, public outcome accountabil-ity, individualized instruction for students withdisabilities) being experienced by classroom teach-ers and schools, the efficient adoption and accu-rate use of evidence-based practices are especiallyimportant priorities. The RTI logic has beenidentified as a promising approach to improvingthe identification of students who might requiremore intensive instructional support. In this ap-proach, a failure to respond to typically effectiveinterventions is used as a marker for more inten-sive interventions, and may assist in identifyingstudents who might require specially designed, in-dividualized education programs.

In the context of students who display prob-lem behaviors, a similar RTI logic has been sug-gested in efforts to establish continua of positivebehavioral interventions at the school and class-room levels (Sugai er aL, 2000; Walker et al.,1996). The purpose of this research was to inves-tigate the application of an RTI approach to be-havior supporr in two second-grade classrooms.

Although general cla.ssroom behavior and instruc-tional management procedures were in place inthese two classrooms, 10 students continued todisplay high rates of problem behaviors. In re-sponse, a more intensive intervention (CICO)was introduced (Study 1). More intensive func-tion-based interventions were individually devel-oped and implemented for the nonresponders toCICO (Study 2). For each of these students, im-provements in problem behaviors were observedand noted.

More specifically, the results suggest that aslightly more intensive but efiflcient intervention(i.e., CICO) was efFecrive in supporting the be-havioral success of four srudents who.sc problembehaviors were initially unresponsive to generalclassroom management practices (e.g., reductionsin mean intervals with problem behavior frombaseline to CICO were at least 38% and greater).For four other students whose problem behaviorscontinued to be unresponsive to the more inten-sive interventions, more individualized, function-based interventions were indicated and proved tobe effective in the classroom setting (e.g., reduc-tions in mean intervals with problem behaviorfrom baseline to function-based phases for all par-ticipants were greater than 60%). Additional

3O6Spring 2007

information on the fidelity of implementation ofCICO and function-based interventions, officediscipline referral patterns, teacher social valida-tion ratings, peer control, and peer compositerates of problem behavior corroborated the use-fulness of an RTI approach to social behaviorscreening and intervention. For example, all par-ticipants engaged in fewer average inrervals witbproblem behavior than composite peers, duringthe intervention phases in which they were mostsuccessful.

This research has a number of notable fea-tures. First, most aspects of the intervention de-velopment and implementation were conductedand informed by the classroom teacher and coun-selor. The fact that school personnel could designand implement interventions with high fidelityreinforces the potential usefulness of the CICOand function-based interventions.

Second, although we had hoped tbat theCICO intervention would be successful with alarger number of studenrs, we learned that specifi-cation and consideration of the foctors that con-tribute to the occurrence of problem behaviorshould be used to predict the likelihood thatCICO would work and the need for more indi-vidualized interventions. For design purposes,nonresponsive students were maintained in theCICO condition for extended numbers of ses-sions. The results from this research would sug-gest that teachers shift to more individualizedinterventions much earlier in CICO implementa-tion. Similarly, because the CICO interventionwas characterized by high rates of teacher atten-tion, students for whom adult attention is not areinforcer or is an aversive should be provided anadapted version of CICO that reduces adult at-tention, increases self-management, and providesmore salient reinforcers.

Finally, teachers reported that the interven-tions were easy to implement and improved thegeneral climate of the classroom, and that beingon CICO or an individualized plan was viewed asa positive experience by students. For example,one teacher reported that when a CICO studentwas absent, another student who was not recom-mended for CICO volunteered to be a "leader"and follow the CICO procedures.

RESEARCH LIMITATIONS

Before discussing the implications and recom-mendations that are associated with the findings,important limitations in this research need to benoted. First, because this research was conductedin two classrooms in one school, we are limited inmaking generalizations about the efficacy of theCICO and function-based approach to RTI toconditions that differ from this study. Researchersshould conduct additional replications across arange of grades and schools and with studentswith varied behavioral profiles and personal de-mographics.

Teachers reported that the interventionswere easy to implement and improved thegeneral climate of the classroom, and thatbeing on CICO or an individualized planwas viewed as a positive experience bystudents.

Second, we conducted this research in actual5econd-grade classrooms where many typical, butunpredictable changes, interruptions, and varia-tions occurred in schedules, routines, and person-nel. Some of the variability in student respondingcould be associated with such variations.

Third, two students who were not successfulunder the CICO condition did not receive moreintensive function-based interventions, in part be-cause of classroom and parental factors. Althoughthese students functioned as opportunistic peercontrols, a test of their responsiveness (rcplica-tion.s) to more intensive interventions would havebeen desirable. Similarly, data collection was dis-continued at the end of March, which did notpermit comparisons with student responding inApril and May.

Fourth, results from secondary data sources,office discipline referrals and teacher social valid-ity questionnaires do not provide experimentalsupport for the CICO and function-based inter-ventions. Although quantitatively both sources ofdata positively favor CICO and function-basedinterventions, improvements in student behaviorcould be associated wirh other plausible explana-tions. For example, teachers may have decreased

Exceptional Children 3O7

their likelihood of referring students to the officeby virtue of their participation in this study.However, these data add to conclusions derivedfrom primary data sources and provide additionalsocial validation information.

Fifth, because observers were also researchersin this study and provided some technical sup-port to classroom teachers, they may not bavebeen objective and consistent in how they identi-fied, rated, and scored observation data (e.g.,reactivity, drift:). However, high levels of interob-server agreement suggesr that observers were ac-curate and consistent in their direct observationresponsibilities.

Sixth, although a multiple baseline designacross four studenrs was possible for the four stu-dents whose behaviors required function-basedinterventions, a similar controlled staggering ofinterventions was nor possible during the CICOconditions. The design limitation is ironically astrength of the CICO intervention, because theobjective is to provide teachers with an efficientmeans of intensifying an intervention for studentswhose behaviors are not responsive to the generalclassroomwide system. In rhis instance, CICOconsisted of a common routine for self- andteacher evaluation against a common set of be-havioral expectations at set times during the day.In addition, the group average was calculated atthe end of the day, and a classroomwide positivereinforcer could be presented. Therefore, all quali-fying students, by design, were introduced toCICO at the same time.

Finally, Helena and Jade had fewer datapoints than other participants—especially acrossCICO phases—-because of their clear and positiveresponse to intervention. In addition, Helenamoved from the school. However, their respond-ing under the general CICO conditions wasclearly improved compared to baseline condi-tions, and levels of problem behavior were clearlycomparable to composite peers. Because he en-tered the study later than the other participants.Ben's data are not as complete or readily compara-ble to the other participants receiving function-based support; however, his data do provideadditional information (one replication) aboutthe impact of function-based support followingCICO intervention.

IMPLICATIONS AND RECOMMENDATIONS

Civen the limitations and the observed results,the findings from this research offer support forthe application of the RTI logic to the implemen-tation of more intensive behavioral interventionsfor students whose behaviors are not responsive tole.ss intensive interventions. Although the CICOand function-based interventions in this researchare evidence-based, guidelines are still needed forknowing when to implement and when to in-crease the intensity of interventions across thethree-tiered logic.

The results from this research were obtainedby using actual school personnel to implement in-terventions in actual classroom contexts. This fea-ture of the research increases the ecologicalvalidity and utility of CICO and fiinction-basedinterventions in real school environmenrs.

Although CICO is an evidence-based prac-tice, the results from this research demonstratethat the behaviors of some students are not im-proved (nonresponders). An important implica-tion is the need to learn more about the contextsin which evidence-based interventions are likelyto have an effect, and the adaptations that mightbe needed to improve outcomes for more stu-dents. Function-based support technologiesmight represent a useful means of guiding suchintervention adaptations.

R E F E R E N C E S

Barnert, D. W., Daly. E. J., Jones. K. M., & Lcnrz, F.E. (2004). Response to intervention: Empirically basedspecial service decisions from single-case designs of in-creasing and decreasing intensity. The Journal of SpecialEducation, 38, 66-79.

Carr, E. G. (1977). The motivation of self-injurious be-havior: A review of some hypotheses. Psychological Bul-letin, 84. 800-816.

Chafouleas, S. M., McDougal, J. L., Riley-Tiliman, T.C , Panahon, C. J., & Hilt, A. M. (2005). What doDaily Behavior Report Cards (DBRCs) measure? Aninitial comparison of DBRCs with direct observationfor off-task behavior. Psychology in tbe Scbools, 42,669-676. I

Spring 2007

Christ, T. J.. Burns, M. K., & Ysseldyke, J. E. (2005).Conceptual confusion within response-to-interventionvernacular: Clarifying meaningful differences. Commu-nique. 34. 1-8.

Crone, D. A., & Horner, R. H. (2003). Building posi-tive behavior support systems in scbools: Functional behav-ioral assessment. New York: Guildford Press.

Daly. E. J. Ill, Martens. B. K., Hamler, K. R., Dool, E.J.. & Eckert, T. L. (1999). A brief experimental analysisfor identifying instructional components needed to im-prove oral reading fluency, journal of Applied BehaviorAnalysis, 32. 83-94.

Eber. I.., Sugai, C , Smith, C. R., & Scott, T. M.(2002).Wraparound and positive behavioral interven-tions and supports in the schools. Journal of Emotionaland Behavioral Disorders. 10.171-180.

Epstein, M. H., Nordness, R D., Gallagher, K., Nelson,J, R.. Lewis. L., & Schrepf, S. (200'i). School as theentry point: Assessing adherence to the basic tenets ofthe wraparound approach. Behavioral Disorders, 30,85-93.

Filter. K. J.. McKenna, M. K., Benedict, E. A.. Horner.R. H., Todd, A. W., & Watson, J. (in press). Checkin/Check out: A post-hoc evaluation of an efficient,secondary-level targeted intervention for reducingproblem behaviors in schools. Education and Treatmentof Children.

Fletcher, J. M., Francis, D. J., Shaywitz, S. E., Lyon, G.R.. Foorman, B. R., Stuebing, K. K.. et al. (1998). In-telligence testing and rhe discrepancy model for chil-dren wilh learning disabilities. Learning DisabilitiesResearch & Practice, 13, 186-203.

Fuchs, D., Mock, D., Morgan, R L., & Young, C. L.(200.3). Responsiveness-to-intervention: Definitions,evidence, and implications for the learning disabilitiesconstruct. Learning Disabilities Researcb & Practice, 18,157-171.

Good, R. H,, & Kaminski. R. A. (Eds.) (2002). Dy-rmmic Indicators of Basic Early Literacy Skills (6th ed.).Eugene. OR: Institute for Development of EducationalAchievement.

Gresham. P. M., Reschly, D. J.. Tilly. W. D., Fletcher,J., Burns, M.. Prassc. D., et al. (2005). A response tointervention perspective. The School Psychologist. 59,26-33.

Hawken. L. S.. & Horner, R. H. (2003). Implement-ing a targeted group intervention within a school-widesystem of behavior supporr. Journal oJ Behavioral Edu-cation, 12. llb-l^Q.

Horner. R. H. (1994). Functional assessment: Contri-burions and future directions. Journal of Applied Bebav-ior Analysis. 27, 40\^04.

Horner, R. H.. Carr, E. G., Halle. J.. McGee, G..Odom, S., & Wolery, M. (2005). The use of singlesubject research to identify e vide need-based practice inspecial education. Exceptional Cbildrcn, 71. 165-179.

Horner. R, H.. Todd. A. W.. Lewis-Palmer, T, Irvin, L.K., Sugai, G., & Boland, J. B. (2004). The school-wideevaluation tool (SET): A research instrument lor assess-ing school-wide positive behavior support. Journal ofPositive Behavioral Inten>entions. 6. 3—12.

Individuals with Disabilities Education ImprovementAct of 2004, 20 U.S.C. § 1400 etseq. (2004).

Iwara. B. A., Dorsey, M., Slifer, K. J., Bauman, K. E..& Richman, G. S. (I982).Tovk'ard a functional analysisof self-injury. Analysis and Intervention in Developmen-tal Disabilities, 2. 1-20.

Kartub, D. T.. 'Faylor-Grecn, S.. March, R. E.. &Horner, R. H. (2000). Reducing hallway noise: A sys-tems approach. Journal of Positive Behavior Interven-tions, 2. 179-182.

Lewis, T. J., Powers. L. J., Kelk, M. J., & Newcomer, L.L. (2002). Reducing problem behaviors on the play-ground: An investigation of the application of school-wide positive behavior supports. Psycbology in theSchools, 39. 181-190.

Lewis, T. J., & Sugai, G. (1993). Teaching communica-tive alternatives to socially withdrawn behavior: An in-vestigation in maintaining treatment effects. Journal ofBehavioral Education, 3, 61-75.

Lewis. T. J., & Sugai, Q. (1999). Effective behaviorsupport: A systems approach to proactive school-widemanagement. Focus on Exceptional Children, 31, 1-24.

March. R. E.. Horner. R. H.. Lewis-Palmer. 1., Brown,D., Crone. D. A., Ibdd. A. W., ct al. (2000). Func-tional assessment checklist for teachers and staff (VAGTS).Eugene, OR: University of Oregon.

Marston, D.. Muyskcns. P. L. Lau, M,, & Canter, A.(2003). Prohiem-solving model for decision makingwith high-incidence disabilities: The Minneapolis expe-rience. Learning Disabilities Research & Practice, 18,187-200.

Metzler, C. W.. Biglan, A., Rusby, J. C , & Sprague, J.R. (2001). Evaluation of a comprehensive behaviormanagement program to improve school-v ide positivebehavior support. Education and Treatment of Children,24, 448-479.

Parsonson, B. S., & Baer, D. M. (1986). The graphicanalysis of data. In A. Poling & R. W. Fuqua (Eds.),

Exceptional Children 3O»

Research methods in applied behavior analysis: Issues and

advances (pp. 157-186). New York: Plenum Press.

Safran. S. R. & Oswald, K. (2003). Positive behaviorsupports: Can schools reshape disciplinary practices?Exceptional Cbildren, 69, 361-373.

Sugai. G., Horner, R. H., Dunlap, G., Hieneman. M..Lewis, T J., Nelson, C. M., et ai. (2000). Applyingpositive behavior support and functional assessment inschools. Journal of Positive Behavior Interventions, 2,131-143.

Sugai, G., Lewis-Palmer, T., & Hagan, S. (1998). Usingfunctional assessments to develop behavior supportplans. Preventing School Failure, 43, 6-13.

Todd, A. W. Kauffman, A.. Meyer, G., & Horner, R.H. (in press). The effects of a targeted intervention toreduce problem behaviors: The check-in check-out pro-gram. Journal of Positive Behavioral Interventions.

U.S. General Accounting Office. (2001). Report to thecommittees on appropriations, U.S. Senate and House ofRepresentatives: Student discipline. Individuals with Dis-abilities Education Act (GAO-01-2I0). Washington,DC: U.S. Government Printing Office.

Vaughn, S.. Linan-Thompson. S.. & Hickman, P

(2003). Response to instruction as a means of identify-

ing students with reading/learning disabiliries. Excep-

tional Children, 69, 391-409.

Vollmer, T. R., & Northup, J. (1996). Some implica-

tions of functional analysis for school psychology.

School Psychology Quarterly, 11. 76-91.

Walker, H. M., Horner. R. H.. Sugai. G., Bullis, M..Sprague, J. R., Bricker, D., et al. (1996). Integrated ap-proaches to preventing antisocial behaviot patternsamong school-age children and youth. Journal of Emo-tional and Behavioral Disorders, 4, 193-256.

A B O U T T H E A U T H O R S

SARAH FAIRBANKS. Graduate Research Assis-tant; and GEORGE SUGAI (CEC CT Federa-tion), Professor and Carole J. Neag EndowedChair, Center for Behavioral Education and Re-search, Neag School of Education, University ofConnecticut. Storrs. DAVID GUARDINO (CECOR Federation). Craduatc Student, College ofEducation, University of Oregon, Eugene. MAR-G A R E T LATHROP, School Counselor, BethelSchool District, Eugene. Oregon.

Address correspondence to Sarah Fairbanks or toGeorge Sugai, Center for Behavioral Educationand Research. Neag School of Education. Univer-sity of Connecticut. 249 Glenbrook Road Unit2064, Storrs, CT 06269 (e-mail: sarah.fair-banks@uconn.edu or george.sugai@uconn.edu).

Preparation of this manuscript was supported inpart by a grant from the Office of Special Educa-tion Programs, U.S. Deparrmcnt of Education(Grant # H324X010015). Opinions expressedherein are those of the authors and do not neces-sarily reflect the position of the U.S. Departmentof Education, and such endorsements should notbe inferred. The authors would like to acknowl-edge Terrie Winkelman, Marsha Barr, and BetsyFernandez for their conrriburions and parricipa-don in this research.

Manuscript received January 2006; accepted June2006.

Spring 2007