xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001670367
007 cr mnu|||uuuuu
008 051118s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001235
Adelinis, John D.
The role of choice versus preference
h [electronic resource] :
b an analysis of why choice interventions work /
by John D. Adelinis.
[Tampa, Fla.] :
University of South Florida,
Thesis (M.A.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 48 pages.
ABSTRACT: Previous research has shown that providing students with the opportunity to choose the type of academic assignment could reduce a variety of problem behavior. However, procedural limitations of previous research prevent definitive conclusions regarding the mechanism by which choice interventions effect behavioral change. Furthermore, because research related to choice interventions has been limited primarily to children with developmental and emotional disabilities, the generality of such interventions is unclear. Therefore, the current study set out to extend the efforts of previous researchers by attempting to further isolate the mechanism by which choice procedures produce improved behavioral performance and attempted to further assess the generality of choice procedures by examining its effects on the behavior (e.g., maladaptive behavior, on-task behavior, academic performance) of a population (i.e., typically developing adolescent youth) not frequently targeted.
Adviser: Jennifer Austin, Ph.D.
x Applied Behavior Analysis
t USF Electronic Theses and Dissertations.
The Role of Choice Versus Preference: An Analysis of Why Choice Interventions Work by John D. Adelinis A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Arts Applied Behavior Analysis Graduate School University of South Florida Major Professor: Jennifer Austin, Ph.D. Kelli McCormack-Brown, Ph.D. Maria dePerczel-Goodwin, Ph.D. Date of approval: July 15, 2005 Keywords: school setting, typically develop ing, choice, establishing operation, multi-element design Copyright 2005, John Adelinis
Table of Contents List of Figures ii Abstract iii Chapter 1 Introduction 1 Chapter 2 Method 12 Chapter 3 Results 19 Chapter 4 Discussion 32 References 40 Appendices Appendix A: Interval Data Sheet 42 Appendix B: Social Validity Q uestionnaire 43 i
List of Figures Figure 1. Bar Graph for Sa m and Tony 20 Figure 2. Sams Graphs (S et One) 22 Figure 3. Sams Graphs (S et One) 24 Figure 4. Tonys Graphs (S et One) 26 Figure 5. Tonys Graphs (Set Two) 28 ii
The Role of Choice Versus Preference: An Analysis of Why Choice Interventions Work John Adelinis ABSTRACT Previous research has shown that providing students with the opportunity to choose the type of academic assignment could reduce a variety of problem behavior. However, procedural lim itations of previous research prevent definitive conclusions regarding the mec hanism by which choice interventions effect behavioral change. Furthermore, because research related to choice interventions has been limited primarily to children with developmental and emotional disabilities, the gener ality of such interventions is unclear. Therefore, the current study set out to extend the efforts of previous researchers by attempting to further isolate the mechanism by which choice procedures produce improved behavioral performance and attemp ted to further assess the generality of choice procedures by examining its effects on the behavior (e.g., maladaptive behavior, on-task behavior, academic performanc e) of a population (i.e., typically developing adolescent youth) not frequently targeted. iii
1 CHAPTER 1 INTRODUCTION Free will is a philosophical doctrine sugges ting that an individual is free to make a choice and that such a choice may be impervious to the pressure of external influences. It is a concept that pervades many facets of our lives. For example, children are frequently reminded th at they can grow to become what ever they want or that in order to succeed they must make good choices, presumably to avoid making bad ones. The concept of choice is so deeply engrained in our society that when one per ceives that an opportunity to make a choice has been restricted, or altogether removed, he may display a variety of measures of countercontrol (Skinner, 1971). Such measures may involve avoidance or escape from those condition s in which behavior is perceived to be under control. In more extreme cases, countercont rol may be exerted through force, aggression, or attack. Interestingl y, reactions involving countercontrol are less likely to occur when there is an opportuni ty to make a choice, even if that choice is merely an illusion. Given the so cietal importance of ones right to make a choice, it is only natural that it woul d become a topic of in vestigation within the psychological literature. The topic of choice has become mo re than just a peripheral concept deserving of attention only prior to init iating a study (e.g., informed consent)
2 and/or within the provision of clinical se rvices (e.g., person-centered planning; Kincaid, 1996). It has bec ome, in and of itself, a topic researchers consider worthy of exploration, as exemplified by a review study conducted by Kern, Vorndran, Hilt, Ringdhal, Adelman, and Dunl ap (1998). Kern et al. identified a large number of research articles within the behavior analytic lit erature that were related to the topic of choice. The authors noted that each of the studies fit into one of three general categories, including 1) examining the use of choice as a means of measuring preference, 2) exploring stra tegies for increasing choice responding, and 3) using choice as an independent variable (i.e., intervention). Although research related to choice is abundant in the literature, according to the authors, between the years of 1975 and 1996, only a small number of studies (i.e., 14) were conducted wherein choice as an intervention was the topic of investigation. Since 1996, only three additional publications on this subject matter were reported in the literature (K illu, Clare, & Im, 1999; Powell & Nelson, 1997; Romaniuk, Miltenberger, Conyers, Jenner, Jurgens, & Ringenberg, 2002) Given the relative paucity of research st udies evaluating the use of choice as an antecedent control invention, this line of research will serve as the topic for the following discussion. Parsons, Reid, Reynolds, and Bumgarner (1990), using an alternating treatments design, showed t hat conditions involving either therapist-selected high-preference activities (i.e., nochoice high preference) or participant selected activities (i.e., choice) were equally effective in increasing the on-task behavior of four adults with mental reta rdation, relative to a condition where participants
3 were provided therapist-selec ted low-preference activiti es (i.e., no-choice low preference). These results showed that providing a choice or simply providing access to previously identified high pr eference tasks could produce an increase in work productivity. Not only do the results of the Parsons et al. (1990) study convincingly illustrate the clinical value associated wit h the use of choice procedures within a vocational setting, but the authors use of a no choice high-preference condition allowed for a more fine grain examinat ion of the mechanism by which choice procedures affect behavior. That is, it is unclear wether the effects associated with the use of choice procedures are related directly to an individuals opportunity to select a stimulus (i.e., choice ), or alternatively, if the provision of choice simply results in access to more hi ghly preferred stimuli (i.e., preference). Regarding the latter supposition, providing one with choice among stimuli likely results in an increased probability that one will be provided access to relatively high-preference stimuli. Access to such high preference stimu li may serve as an establishing operation (Michael, 1993) that diminishes the evocative properties of the work context, in turn, decreasi ng the occurrence of escape maintained problem behavior (assuming that problem behavior occu rring during work related contexts is escape maintained). Ther efore, by including a no choice highpreference-task condition and showing t hat such a condition could produce outcomes similar to those obtained during a choice condition, results of the Parsons et al. study suggest that the more salient variable inherent in choice procedures is preference rather than the provision of choice.
4 Unfortunately, a limitation of the Parsons et al. study was the absence of a choice low-preference task condition. A condition in which choice was provided among low-preference activities may hav e functioned as a better control for preference. Furthermore, inclusion of su ch a condition would have allowed for a more stringent test of the treatment integrity of choi ce procedures by evaluating the extent to which such procedures c ould supersede the effects of exposing individuals to low-preference, someti mes aversive, activities (e.g., the hypothesized establishing operation). This is of importance since, for some persons identifying high pr eference vocational and/or academic tasks may be difficult, if not impossible. Another study that showed the effects of choice interventions in vocational settings was conducted by Seybert, Dunlap, and Ferro (1996). The authors evaluated the effects of a choice inte rvention within a vocational setting by providing participants between the ages of 13 and 22 years who had been diagnosed with moderate to se vere mental retardati on with a choice between several vocational and domestic tasks. Using a reversal design, the authors demonstrated that adaptive and maladaptiv e behavior occurred at higher and lower levels, respectively, in the choice condition relative to the no-choice condition. Findings of the authors study provide an additiona l demonstration of the utility of choice procedures for use in vocational settings as a means of improving performance related to ta sk completion and maladaptive behavior Research conducted by Bambara, Ager, and Koger (1994) further evaluated the effects of choice procedur es on behavior within a vocational setting
5 by conducting several manipulations across two studies. During study 1, the authors exposed three adult participants, who were diagnosed with moderate to profound mental retardation, to choice, no choice high-preference, and no choice low-preference conditions using a multi-el ement design. Results from study 1 showed that the choice and the no choice high-preference task conditions were equally effective in increasing on-task behavior relative to a condition where participants were assigned low-preference vocational tasks. In study 2, the authors exposed participants to choice and nochoice conditions while attempting to hold preference constant by using ta sks of similar prefer ence (e.g., moderately preferred) across choice and no-choice conditions. The results of study 2 showed little difference in the level of on-task behavior across the two experimental conditions. The authors c oncluded that the combined results of study 1 and study 2 suggest that the effe cts stemming from t he use of choice procedures may be a function of prefer ence rather than choice. Although the authors extended the work of Parsons et al (1990) by further examining the role of preference in choice procedures, mu ch like the Parsons et al. study, the authors did not include a condition wher ein participants would be provided a choice among low preference activities Again, such a condition would have provided a better test for the effects associated with choice procedures. Also, the absence of a baseline established prior to experimental manipulations prevents conclusions to be drawn regar ding the efficacy of the described procedures relative to naturally occurri ng conditions expected within a vocational setting.
6 In addition to vocational settings, researchers have assessed the use of choice procedures within academic contexts For example, Powell and Nelson (1997) evaluated the effects of providing a 7-year-old student diagnosed with Attention-Deficit/Hyperactivity Disorder with a choice of academic assignment on an aggregate measure of undesirable behavior. The authors, using a withinsubject experimental design (i.e., within series design) compared the effects of choice and no-choice experimental condi tions on several topographies of the students problem behavior. Results showed that providing the student an opportunity to select academic assignments (i.e., choice) produced a decrease in problem behavior. However, the authors noted that c onclusions regarding the effects of the choice intervention on ac ademic performance (i.e., achievement) could not be established due to the absenc e of such a measure. Despite this shortcoming, the authors contributed to the choice-making literature in at least two ways. First, they used choice as an intervention in a general education classroom, and second, they demonstr ated the ease with which choice procedures could be implemented by using staff members who were permanently assigned to the class. In an earlier study, Dunlap et al. ( 1994) explored the utility of choice procedures within an academic setting as a means of decreasing problem behavior and increasing student engagement. In an initial study, the authors showed that providing two emotionally handicapped 11-year-old students with an opportunity to choose their work assignments produced lower levels of problem behavior and higher levels of task engagement relative to conditions where
7 teachers selected assignments (i.e., no choi ce). In a followup study, the authors sought to further discriminate the me chanism by which behavior change occurs when individuals are provided with a choi ce by yoking the work assignment selections (i.e., the choice of book for an adult to read to him) made by a 5 yearold student during a choice condition to a subsequent no-choice condition. The authors found that the choice conditions produced gr eater clinical outcomes (increased attending and decreased off-ta sk behavior) than the no choice yoked conditions, even when the type of activi ty and the sequence in which they were delivered remained constant across both conditions. Therefore, the authors concluded that behavior resulting from the use of choice procedures was related to the act of choosing, rather than preference. By yoking the assignments selected during the choice condition to the subsequent no choice condition, the authors seemingly arranged an adequate co ntrol for preference. However, supplemental data collec ted during the second st udy showed considerable variability in the assignments selected by the participant across choice conditions. That is, the participant dem onstrated a shift in preference across two temporally distal choice conditions; ther efore, it is reasonabl e to suggest that the participants preference for assignments si milarly may have shifted in the time between the choice condition and the followi ng no-choice yoked condition. Such a preference shift may have mitigated the efficacy of the no-choice yoked condition as a control for preference. Ther efore, conclusions related to the results of study two should be regarded as tentative.
8 Killu, Clare, and Im (1999) used pref erence assessments to identify the relative preference of 20 familiar spel ling assignments prior to exposing three participants, who were diagnosed as emot ionally impaired, mentally impaired, and/or learning disabled, to a series of choice and no-choice experimental conditions. Using an ABCDEF design (choice of preferred tasks, choice of nonpreferred tasks, no choice of preferred tasks, no choice of non-preferred tasks, no-choice of preferred tasks [yoked control], and no-choice of non-preferred tasks [yoked control]), the authors c onducted systematic manipulations along dimensions of preference (i.e ., preferred v. non preferred) and choice (i.e., choice v. no choice). The authors found that participant on-task behavior increased during sessions where access to high preference activities was provided, independent of whether or not they we re provided an opportunity to select assignments (i.e., choice). These result s suggest that the vari able of importance for their participants was not the provisi on of choice, but access to preferred activities for which the provision of choice allows. The authors extended the research related to choice interventi ons and contributed to the behavior analytic literature by providing yet another demonstration of the clinical ut ility of choice procedures as an antecedent invention in academic settings and by more effectively examining the role of preference in outcomes resulting from such procedures. More recently, a study conducted by Romaniuk, Miltenberger, Conyers, Jenner, Jurgens, and Ringenberg (2002) inve stigated the extent to which the efficacy of choice interventions is re lated to behavioral function (e.g., escape
9 maintained v. attention maintained). T he authors conducted analyses to identify the function of problem behavior for se ven elementary school-aged participants with various psychological disorder s, including ADHD, mood disorder (unspecified), and developmental delays pr ior to exposing each participant to choice and no-choice conditions. Using a series of reversal designs, the authors concluded that problem behavior was much more sensitive to choice interventions when behavior was found to be maintained by negative reinforcement in the form of escape rather than positive reinforcement in the form of attention. Also, the authors provided evidence of the integrity of choice procedures by evaluating their utility in the absence of relevant extinction components (i.e., escape remained availa ble for instances of problem behavior) To date, the reliable outcomes produced by the use of choice procedures lend support to the argum ent that choice should not only be considered an ethical standard for which providers strive prior to the provision of clinical services, but also as an appropriate antecedent manipulation toward the treatment of a variety of escape mainta ined target behavior across distinct settings. However, the fairly narrow scope with which choice procedures have been tested indicate a need for additional re search. For example, of the number of research articles within the behavior analytic literature related to the use of choice as an antecedent intervention, only a small percentage (i.e., 18%) have been conducted with participants who have not been diagnosed with developmental disability and/or severe mental impairment. Of those studies conducted with participants wit hout developmental disa bilities, none have been
10 conducted with adolescents. T herefore, there remains a need for additional tests of generality. Also, limitations of previous rese arch preclude definitive conclusions regarding the mechanism by which choice procedures operate. For example, the absence of a choice low-preference condition in the Bambara et al. (1994) study prevents conclusions regarding the role of choice in outcomes produced by choice procedures. Also, the Dunlap et al (1994) investigation showed that the participant from study two displayed shifts in preference across time thereby brining into question the adequa cy of the authors no-choice yoked condition as a control for preference. Similarly, al though Parsons et al. (1990) showed that a no choice low-preference condition was as effective as a choice condition in producing desired outcomes, a comparison of such conditions does not provide insight to the variables res ponsible for the efficacy of the choice condition (i.e., choice v. preference). Final ly, although Killu et al. (1999) attempted to control for preference shifts across time so that the role of preference c ould be more clearly identified, a pre-st udy preference assessment may not be the most effective approach. That is, ones preference can va ry from minute to minute; therefore, more frequently conducted (i.e., pre-se ssion rather than pr e-study) preference assessments should be conducted to better capture and control for shifts in preference. The results of Killu et al. are impressive; however, similar results could be bolstered with the use of a more stringent experimental design (e.g., one allowing for reversal).
11 Although research has demonstrated the clinical utility of choice procedures in a variety of sett ings when used as an antecedent based intervention, there remains a need for additi onal research that will further explore the generality of outcomes produced by choi ce procedures and to better isolate the variables (i.e., choice v. preference) responsible for such outcomes. Therefore, the present res earch will use procedures si milar to those used in previously conducted research to furt her assess the generality of choice procedures by examining its effects on the behavior (e.g., maladaptive behavior, on-task behavior, academic performance) of a population (i.e., typically developing adolescent youth) not frequently targeted. Furthermore, the current study will extend the efforts of previous researchers by attempting to further isolate the mechanism by which choi ce procedures produce improved behavioral performance by first, using pre-sessi on preference assessments as a means of controlling for preference shifts ac ross time and second, by exposing participants to no choice high-preference, no choice low-preference, choice highpreference, and choice low-preference conditions.
12 CHAPTER TWO METHOD Participants and Setting Participants in the current study in cluded two typically developing (i.e., nondevelopmentally delayed) participants. Both Sam (age 13) and Tony (age 12) were in 7 th grade at a public charter school. Their teacher also participated in the study. Sam and Tony were selected fr om a small sample of students (four) that were eligible for inclusion based on the recommendation of the University of South Floridas Institutional Board Review Boar d (IRB) (i.e., the IRB recommended selecting participants am ong those students who lived with biological parents as opposed to foster parents). However, Sam and Tonys inclusion was supported by direct observa tion of each participant and interview of school staff that indicated each student displayed several topographies of problem behavior (e.g., cursing, aggre ssion, etc.) and/or demonstrated a sub par academic performance (e.g., below average grades). All sessions were conducted in each participants respecti ve classroom. For both Sam and Tony, sessions were conducted during their Social Studies class. All procedures were approved by the universitys Institutional Review Board. Parental consent and student assent also were obtained for each participant and their respective teachers prior to the start of data collection.
13 Dependent Variables and Measurement During all sessions (except preference assessments), data were collected on four student behaviors (on-task, talking to other students, inappropriate verbalizations, and aggression) using a parti al interval recording and one student behavior (in-seat) using whole interval. On-task was defined as orienting toward the assignment, not directing attention to unrelated task for mo re than 5 s, and displaying behavior required for assignment completion (e.g., co mpleting a cross word puzzle). Talking to other students was defined as any verbal behavior between participant and other students exc ept for questions or statements relating to work assignment. Inappropriate verbalizations were defined as any verbal behavior consisting of cursi ng, verbal threats, or insults. Aggression was defined as any physical contact between par ticipant and others involving hitting, kicking, pulling hair, pushing, or pinching. In-seat was defined as contact between buttocks and chair seat. Data also were collected on teacher interactions, which were defined as any verbal/physical response by the teacher directed toward the participant. Each 20-min observation session was par titioned into 10-s intervals so that data collectors could indicate w hether or not responding occurred during each respective interval. A devise that emitted a tone to signal the end of one interval and the beginning of a subsequent interval was used to cue observers to record the presence or absence of target behavior within each interval. Permanent product data also were collected during each session. Assignment scores were defined as the total number of correct responses
14 divided by total number possible respons es multiplied by 100%. Assignments were collected and scored by the teacher at the end of each session. Data during all sessions were collected by the researcher and a teachers aid. The teachers aid wa s trained prior to the ons et of the study by the researcher. She was provided operational definitions for all dependent variables and was required to practice data colle ction until she reached a level of competency (90% agreement fo r 3 consecutive sessions). Competency was met in six 10-minute sessions. Interobserver Agreement During 25% and 31% of sessions for Sam and Tony, respectively, a second observer independently collected data on all relevant dependent variables. Interobserver agreement checks were spaced across the study so that measures were obtained across all conditions. During all assessments, interobserver agreement was calculated by di viding the number of intervals with agreement divided by the number of interv als with agreements plus intervals with disagreements and multiplying by 100%. For Sam IOA coefficients were as follows: on-task (M= 92%, range 89%-97%); in-seat (M= 95%, range 93%-98%); talking to other students (M= 98%, range 98%99%); inappropriate verbalizations (M= 97%, range 93%-100% ); aggression (M= 100%); and teacher interactions (M=98%, range 98%-99%). Fo r Tony, IOA coefficients were as follows: on-task (M= 83%, range 76%-89%); in-seat (M= 94%, range 92%-99%); talking to other students (M= 98% range 97%100%); inappropriate
15 verbalizations (M= 99%, range 98%-100% ); aggression (M= 100%); and teacher interactions (M=98%, range 95%-100%) Preference Assessments Prior to each experimental session, a paired-choice preference assessment as described by Fisher, Piazza, Bowman, Hagopian, Owens, and Slevin (1992) was conducted with each student. Just prior to the start of each session (i.e., no longer than 10 min. befor e the start of an assignment), the teacher provided the researcher with five variations of a Social Studies assignment (i.e., Crossword Puzzle, True or False, Fill in the Blank, Short Answer, and Multiple Choice.). Each type of assignment (e.g., True or False) was printed on a separate card. Two cards were then presented to the student and the student was asked to select t he one card with type of assignment he would most prefer. This procedure wa s repeated until each of the cards had been presented with every other ca rd once. During the pres entation of pairs, the researcher recorded the number of times each assignment was chosen. The assignment chosen most frequently wa s defined as the high preference assignment, whereas as the one chosen least frequently was as the low preference assignment. Although within each preferenc e assessment there were multiple types of assignments, the subject matter of each type of assignment was identical. That is, each type of assignment was pulled from the same chapter of a Social Studies curriculum. Therefore, the only difference between assignments was the format of the work. Experimental Condit ions and Procedures
16 No choice (baseline). During the no choice condition, participants were provided an assignment (i.e., Crossword Puzzle, True or False, Fill in the Blank, Short Answer, or Multiple Choice) that wa s selected by the participants instructor with no instruction from this investigator That is, the no choice condition was included as a free operant context that was intended to capture pre-intervention conditions and their effects on participant maladaptive and adaptive behavior. No choice / low preference assignments The no-choice / low preference condition was similar to the no choice c ondition in that participants were not provided an opportunity to choose their assignment. However, the assignments selected for the participant were the ones identified as the least preferred during the pre-session preference assessment. No choice / high preference assignments This condition was similar to the no choice / low preference condition in t hat each participant was not allowed to choose their assignment. However, rat her than selecting the least preferred assignment, the researcher selected the a ssignment identified as most preferred during the pre-session preference assessment. Choice / high preference assignments During this condition, each participant was provided a choice am ong the two most highly preferred assignments identified dur ing the pre-session preference assessment. Choice / low preference assignments This condition was similar to the choice / high preference condition, except that the participants were provided a choice among the two least preferred assignments identified during the presession preference assessment.
17 Assignments used across baseline and experimental sessions were consistent. That is, the same types of assignments were used for all conditions. They differed only with regard to student preference and provision of choice The relative effects of the experim ental conditions were assessed using an alternating treatments des ign (with an initial baseli ne phase). The order of treatment conditions was arbitrarily selected. That is, four pieces of paper, each with the name of one condition, were placed in a cup and one was blindly selected from the cup prior to each sessi on. Selected pieces of paper were not replaced until all pieces of paper had been se lected. However, during the initial sessions of the multielement phase, two sessions in the same condition were conducted daily. For example, if the c hoice / high preference condition was selected on a given day, two consecutiv e choice / high preference sessions would be conducted. For Sam, this me thod of conducting sessions was in place for two days (four sessions); for Tony, this method wa s in place for four days (eight sessions). Because there was a need to accelerate the manner in which sessions were conducted and a need to increase the number of exposures to each condition, three sessions were conducted daily and the condition for each session was arbitrarily selected without r eplacement using the same selection procedure described above. Social Validity The school principal and teachers were interviewed to solicit information to identify the array of assignments required fo r use in the study to ensure the use of socially valid procedures Furthermore, all consumers (i.e., teachers and
18 student participants) were asked to comp lete a questionnaire (Appendix B) prior to the start the study a nd once again at the end of the study to provide a measure of satisfaction with the procedures and outcomes.
19 CHAPTER THREE RESULTS Figure 1 shows results of the pr eference assessments for Sam and Tony. Each type of assignment presented during the preference assessment is represented by a single color. The Xaxis represents a ranking system ranging from 1 (most preferred) to 5 (least preferred). The Y-axis represents the percentage of preference assessment sessions wherein a type of assignment (e.g.. True or False) ranked a particular rank ing. For example, for Sam Fill in the Blank, True or False, and Puzzle were all selected at least once as the most highly preferred type of assignment during the course of the study. Data reveal that both participants demonstrated shifts in preference across the course of the study. For example, Sam showed a shift in preference relating to the fill in the blank assignment that ranged in prefer ence from most preferred to least preferred. More specifically, three a ssignment types were identified as most preferred with four different types of assignments having been identified as second most preferred during at least one preference assessment. Similar shifts occurred with the lower ranked assignment types as well. Tony also demonstrated shifts in preference, though his preferences were less varied than Sams. For example, puzzle assignm ents varied in preference from most preferred to third most preferred. Ho wever, such shifts were not noticeable amongst the least preferred assignments (fill in the blank and short answer
ranked as fourth and fifth, respecti vely, during 100% of the preference assessment sessions.) Although shifts for either participant were not considerable, these results suggest t hat had assignment selection (participant and/or teacher based selection) been bas ed solely on the results of a single preference assessment, participants ma y not have been provided access to the highest or lowest preferred assignments. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 12345Ranking (1= most preferred / 5 = least preferred)Percentage of Sessions Selected Multiple Choice Fill in Blank Short Answer True or False Puzzle 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 12345Sam Tony Figure 1. Bar Graph for Sam and Tony Figure 2 presents Sams data for ontask behavior, in-seat behavior, and talking-to-others. On-task behavior occurre d at fairly low but stable levels across all conditions. However, there was s light increase in on-task behavior observed during the choice/high preference conditi ons. During baseline, Sam was on task 20
21 an average of 7% of the intervals observed (range, 2% to 10%). During the choice/high preference condition, t he mean percentage of intervals on task increased to 19.7% (range, 1.6% to 28%), with all data points except one falling above the baseline range. For all other conditions, percentage of intervals remained within the baseline range (choice/l ow preference, M = 5.3%, range, 0% to 13%; no choice/high preference, M= 3. 6%, range, 0% to 11%; no choice/low preference was 0% across all sessions). Results related to in seat behavior were considerably more variable than on-task behavior with no clear response differentiation observed across phases or among experimental conditions. The m ean percentage of intervals with in-seat behavior for baseline was 54.3% (range, 38% to 75%), though data were trending downward before the start of experimental manipulations. The choice/high preference condition yielded a mean score of 61.5% and data were highly variable (range, 28% to 96%). C onsiderable variability also was observed in the other conditions, though means tended to be lower (i.e., choice/low preference, M = 48.6%, range, 5.8% to 99%; no choice/hi gh preference, M = 45.3%, range, 32% to 66%; no choice/low preference, M= 49.3, range, 5% to 99%). During baseline, mean percentage of intervals with the occurrence of talking to others was 37.3% (range, 28% to 45%). Talking to others occurred at a consistently lower level across all experi mental conditions relative to baseline with the lowest level of behavior observ ed during the no choice/high preference condition. Levels were stable across all ex perimental condition s (i.e., choice/high
preference, M=7.5%, range, 2.5% to 12%; choice/low pr eference, M=7.1%, range, 2.5% to 10%; no choi ce/high preference, M=18. 3%, range, 11% to 30%; no choice/low preference, M= 15.0% range, 4% to 26%). 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w/ On-Tas k 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w/ In-Seat 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 Session NumberPercent Interval w/ Talking to Other s =ChoiceHigh Pref ChoiceLow Pref No ChoiceHi gh P r ef No ChoiceLow Pref Sam Change in Condition Sequence Change in Condition Sequence Change in Condition Sequence Figure 2. Sam Graph (Set One) Figure 3 represents Sams data fo r inappropriate verbalizations, aggression, and academic scores. During baseline, Sams level of inappropriate verbalizations occurred at low and stable levels (M=15.6%; range, 10% to 20%). This behavior occurred at a lower level dur ing experimental conditions relative to 22
23 baseline (i.e., choice/high preference, M=1.8%, range, 0% to 5%; choice/low preference, M=0.8%, range, 0% to 2.5% ; no choice/high preference, M=3.9%, range, 1.6% to 5%; no choice/low pref erence, M=3.2%, range, 0% to 8%). Aggression occurred at zero levels across all conditions. During baseline, Sams assignment scores were relatively low and somewhat variable (M = 6.6%; range, 0% to 35%). During experimental conditions, assignment scores were relative ly high but variable in the choice/high preference condition (M = 25.5%; range, 0% to 52%), but the range of scores during the choice/high preference condition was comparable to that of baseline with the exception of the second sessi on where assignment score was higher than any baseline scores. Interestingly, Sam obtained assignment scores of zero in all other experimental conditions.
24 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / Inapp. Verb. 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / A g g 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 Session NumberScore on A cadem ic Tas ChoiceHigh Pref ChoiceLow Pref No ChoiceHigh Pref No ChoiceLow Pref Change in Condition Sequence Change in Condition Sequence Sam k Change in Condition Sequence Figure 3. Sam Graph (Set Two) Figure 4 represents Tonys data for on-task behavior, in-seat behavior, and talking-to-others. During baselin e, a downward trend in on-task behavior was observed with a mean level of 51% (range, 32% to 72%) Introduction of each experimental condition resulted in a considerable increase in on-task behavior. More specifically, on-task behavior occurred at high and stable levels during the choice/high preference (M = 90%; range, 81% to 98%) and no choice/high preference (M = 93%; range, 85% to 100%). Levels of on-task behavior was slightly more variable during the choice/ low preference condition
25 (M = 63%; range, 2.5% to 87%) with the first three sessions resulting in a level higher than that observed during baseline; however, on-task behavior dropped to near zero during the last session of the choice/low preference condition. Finally, following an initial session during the no choice/low preference where the level of on-task behavior was higher than baseline, a considerable drop in level was observed during the final three sessi ons (M = 30%; range, 5% to 95%) In seat behavior occurred at high and stable levels during all experimental conditions relative to base line (M = 78%; range, 68% to 73%). The experimental phase produced slightly higher levels of in-seat behavior across conditions (i.e., choice/high preference condition, M=98%, range, 94% to 100%; choice/low preference, M = 96%, range, 92% to 99%; no choice/high preference, M= 98%, range, 96% to 100%; no choice/low preference was 96%, range 92% to 100%). Talking to others occurred at a c onsistently lower level across all experimental conditions relative to bas eline with the lowest level of behavior observed during the no choice/high preferenc e condition. During baseline, mean percentage of intervals with the occurrence of talking to others was 16% (range, 2% to 30%). Following an upward trend in baseline, experimental conditions were introduced and a decrease in talking to other students was observed across all conditions. The greatest reduction in talking to others was observed during the no choice/high preference condition wi th a mean of 0%. The mean percent of intervals with talking to others during the choice/high preference condition was 2.8% (range, 0% to 6%). The choice /low preference condition produced a mean
of 7.8% (range, 0% to 20%). Finally, although the no choice/low preference condition also resulted in a reduction in talking to others with a mean of 10.4% (range, 1.6% to 18), a slight upward trend was observed. 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / O n-Tas k 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / In-Sea t 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 Session NumberPercent Interval w / Talking to O ther s Tony = ChoiceHigh Pref ChoiceLow Pref No ChoiceLow Pref No ChoiceHigh Pref Change in Condition Sequence Change in Condition Sequence Change in Condition Sequence Figure 4. Tony Graph (Set One) 26 Figure 5 represents Tonys data fo r inappropriate verbalizations, aggression, and academic scores. Du ring baseline, an upward trend in inappropriate verbalizations was observed with a mean percent of 12.3% (range, 0% to 22%). The introduction of the ex perimental phase resulted in one of the more dramatic changes in behavior with inap propriate verbalizations occurring at
27 near zero levels across all conditions (choi ce/high preference, M=0%; choice/low preference, M=0%; no choice/h igh preference, M=0%; no choice/low preference, M=0.5%, range, 0% to 2%). As was the case with Sam, aggression occurred at zero levels across all conditions. Finally, a considerable downwar d trend in assignment scores was observed during baseline with a mean of 55% (range, 30% to 100%). As was the case with on-task behavior, the onset of the experimental phase produced clinically significant changes in assignm ent scores across all conditions. The choice/high preference produced the greates t improvement in assignment scores with a mean of 91.3%, (range, 80% to 100% ), followed by the no choice/high preference condition with a mean of 91% (range, 86% to 100%). Assignment scores were also high during the first thr ee sessions of the choice/low preference condition before dropping to near zero dur ing the final session (M= 59%, range, 0% to 82%). Finally, during the no choi ce/low preference c ondition, an initial session where assignment score was 75% was followed by a considerable drop in assignment score to near zero during the final three sessions (M= 18.7% range, 0% to 75%).
0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / Inapp. Verb. 0 10 20 30 40 50 60 70 80 90 1000 5 10 15 20Percent Interval w / A g g 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 Session NumberScore on A cadem ic Tas k ChoiceHigh Pref ChoiceLow Pref No ChoiceLow Pref No ChoiceHigh Pref Tony Change in Condition Sequence Change in Condition Sequence Change in Condition Sequence Figure 5. Tony Graph (Set Two) In addition to student behavior, data we re also collected on teacher behavior. More specifically, data were collected on the percentage of intervals with teacher-student interaction such t hat any potential relation between such interactions and change in targeted student behavior could be monitored and adequately interpreted. Figure 6 displays these data. For Sam, the percent of teacher interaction was low across bas eline with a mean of 4.3% (range, 0% to10%). The level of teacher inte raction dropped during the choice/ high 28
29 preference condition to a mean of 1.75% (range 0% to 3%). A similar level of teacher interaction was observed during the choice/ low preference condition (M=1.7%, range 0% 5%). Conversely, the level of teacher interaction during the no choice/ high preference condition was co mparable to that of baseline with a mean of 4.3% (range 1% to 9%). Finally, the lowest level of teacher interaction was observed during the no choice/ low preference condition with a mean of 0.3% (range 0% 1%). Most import antly there appeared to be no correlation between teacher interaction and change in Sams behavior. For Tony, teacher interaction was hi ghest during baseline with a mean of 7.3% (range 3%-12%). T eacher interaction occurred at lower levels during the choice/ high preference condition (M= 5.3% range 2% to14%). The lowest level of teacher interaction was observed duri ng the choice/ low preference condition with a mean of 2.5% (r ange 0% to 5%). Teacher intera ction occurred at a slightly higher level during the no choice/ high pr eference condition with a mean of 3.2% (range 0% to 9%). Finally, the no c hoice/ low preference condition was correlated with the highest level of teac her interaction rela tive to the other experimental conditions with a mean of 5.7% (range 0% to 12%). As was the case with Sam, there appeared to be no re lation between interactions between Tony and the teacher and changes in behavior observed during experimental conditions. Finally, participants (teacher and two st udents) were administered a social validity questionnaire prior to the initiation of the study and again at completion of the study. However, Sam refused complete bot h. Results related to results of the
social validity questionnaire can be found in Table 1. In general, both the teacher and student found the use of choice procedures to be an acceptable strategy toward the treatment of classroom disr uptive behavior. Furthermore, responses related to questions remained constant across the pre and post questionnaires. Table 1 Social Validity Questionnaire Student (Tony) Teacher Pre Study/Post Study Pre Study/Post Study Need for Services (Pre-Study Only) Do you feel that you (or the student) are in need improved behavior during academic contexts? Yes Yes In which academic subject do you (or the student) need improvement? Soc. Stud. Soc. Stud. and Math Which behavior, specifically, would you like to see improved? On-Task, In Seat On-Task and and Aggression In Seat Acceptability of Procedures (Pre-Study and Post-Study) Would you like to be provided (or provide) choices with respect to academic assignments? Yes/Yes Yes/Yes Do you think being provided (or providing) a choice of academic assignment would result in a positive change in your (or the students) behavior? Yes/Yes Yes/Yes Teachers Only: Do you think it would be feasible to provide students with a choice of academic assignment? Yes/Yes 30
Teachers Only: Do you think you would be willing to provide students with a choice of academic assignment? Yes/Yes Acceptability of Outcomes (Post-Study Only) Did you notice a change in your (or the students) behavior at any point during the study? A little Yes Which of the following conditions do you believe resulted in the greatest change in your (or the students) behavior. Choi ce/High Preference All Which behavior, specifically, do you think was improved? On-Task and Agg. On-Task 31
32 CHAPTER FOUR DISCUSSION The previous work of researcher s clearly has shown that providing individuals with choice can be an effect ive behavior change strategy for various topographies of behavior across a variety of settings (e.g., Bam bara et al, 1994; Dunlap et al, 1994; Powell & Nelson, 1997; Seybert et al, 1996). Although the efforts of these researcher s and others have clearly demonstrated the viability of choice based interventions, there re mained continued debate related to the mechanism by which such procedures resu lt in change in behavior. Specifically, it is unclear whether exposure to choice procedures simply placed participants in contact with high-preference activities; ther eby mitigating the av ersive qualities of the instruction (e.g., academic, vocational, etc.); or if alternatively, the clinical outcomes often observed with the use of choi ce procedures is a function of the act of choice making. Although choice related research has resulted in a considerable contribution both to the clinical and resear ch literature, concerns relating to prior methodology limit one from making defin itive conclusions regarding the mechanism by which choice engenders behavioral change. Furthermore, participants in previously conducted re search have been diagnosed with either developmental and/or emotional disabilities, thus limiting the generality of the
33 findings. Given these procedural limitat ions and the narrow scope with which the procedures have been tested, additional resear ch related to choice was needed. Therefore, the current inve stigation set out to test the generality of choice procedures by evaluating their efficacy with typically developing students. Further, although the efforts of previous researchers have shed some light on the role of choice versus preference, limi tations of their research only allow for tentative interpretations. The current inve stigation sought to further examine the role of choice and preference by exposi ng two students to baseline, choice highpreference, choice low-preference, no c hoice low-preference, and no choice high preference conditions using a modified mu ltielement experimental design (with a baseline phase). Additionally, preferenc e assessments were conducted prior to the start of each session to better control for shifts in preference across time, a confound encountered during previous research. Surprisingly, results of the current in vestigation did not reveal consistent patterns of behavior change related to choice or preference variables for either of the participants. However, some effects ar e worth noting. Sams results (Figures 2 and 3) showed changes in select topographies of behavior. For example, ontask behavior occurred at a slightly higher level in the choice/high preference condition relative to baseline and the other experimental conditions. It is also interesting to note that Sams assi gnment scores were higher during the choice/high preference condition relative to other experimental conditions, but only one data point fell above the baseline range. The aforementioned results suggest that for Sam, the provision of choice amongst high preference activities
34 produced some differential responding, albeit clinically insignificant. That is, choice or access to high preference assignments in isolation would not have been sufficient as evidenced by the absenc e of behavior change in the other test conditions. Perhaps the most compelling finding for Sam is that the percentage of intervals in which he engaged in talking to other students was slightly lower in all experimental conditions relative to base line. However, it is unclear how such an outcome was produced given the absenc e of similar patterns in other topographies of behavior. For Tony, several improvements in target responses were observed (Figures 4 and 5). First, an increase in on-task behavior was observed relative to baseline in all conditions except no choice/low pref erence. In fact, on-task behavior during the no choice/low preferenc e condition occurred at a level lower than that observed during the initial basel ine phase. This outcome suggests that for Tony, on task behavior was equally affected by choice and preference. That is, the provision of choice could ov erride the evocative properties of nonpreferred tasks, but choice is not necessa ry if the assignment selected by the teacher is highly preferred However, although the results related to on-task behavior constitute a clinically significant outcome, its worth noting that the downward trend observed in on-task behavior during baseline may have resulted in levels comparable to those observ ed during the no choice/low preference condition if additional baseline sessions were conducted. A second interesting finding in Tonys data was that all experimental conditions produced higher levels of in-seat behavior relative to baseline sessions. This outcome is somewhat
35 surprising since the baseline condition was comparable, presumably, to the no choice/low preference experimental condition. This pattern may have been an artifact of experimental se ssions more closely approximating a discrete trial with more salient start and stop points. That is, Tony may have been more motivated to work since it was made clear at t he onset of the session that the session would terminate in 20 min, which afforded more predictability. Third, Tonys level of inappropriate verbalizations occurred at near zero levels during all experimental conditions relative to baseline sessions. Here again, this outcome may have been related to a perception of increased structure or formality which may have affected behavior. Finally, an interesting trend was observed with Tonys academic performance. More spec ifically, both choice conditions and the no choice/high preference conditions produced an increase in academic performance relative to baseline and no-choice/low preference sessions. Furthermore, it should be noted that during those sessions where academic score was at, or near, zero, Tony attempted little work. Therefore, low scores were the result of not doing work rat her than doing the work incorrectly (as evidenced by Tonys level of On-Task behavio r). This distinction is of importance because if low academic scores were related to the latter, one could argue that any differentiation observed amongst text conditions related to academic score could be an artifact of task complexity. These results suggest that for Tony simply providing access to high pr eference assignments can improve on-task behavior and academic performance; however, the provision of choice also appeared to have an effect on these behaviors as observed during the
36 choice/low preference condition. Therefor e, for Tony choice and preference may be equally effective. Neither participant showed significant shifts in preference across the course of the study. However, minor sh ifts (e.g., alternating between two types of assignments) were obser ved and could have resulted in an experimental confound had preference asse ssments not been conducted prior to each session to control for such shifts More specifically, the abs ence of frequent preference assessments could have resulted in participant selection from among low preference assignments during those condi tions calling for the use of high preference assignments and visa versa. Although the procedures described in the current study address many of the methodological shortcomi ngs of previous research related to the topic of choice, the current investigation is not without its own limitations. First, even if clear differentiation amongst conditions had been observed, the number of participants in the study precludes any st rong conclusions regarding the effects of preference and/or choice nor does it allow for an adequate assessment of the external validity of choice interventions Second, the sequence in which sessions were conducted during the multi-element phase was inconsistent. That is, sessions within a given condition were in itially run consecutively, but later switched so that exposure to experiment al conditions was sequential. This inconsistency was a direct function of t he number of days remaining in the school year and the need to maximize the number of sessions conducted. Although this procedural shift likely did not result in a confound (e.g., sequence effects), it
37 brings into question the experimental rigor of the current study. Finally, for Tony, additional sessions should have been conduc ted to allow for behavioral patterns to stabilize. More specifically, the in stability observed during the choice/low preference session toward the end of the st udy, particularly in relation to on-task behavior and assignment score, warrants additi onal exposure to all experimental conditions to be capable of making more definitive conclusions. Limitations of the current study as ide, there remains many reasons for why additional research related to the efficacy and acting mechanism of choice procedures for typically devel oping students is warranted. A direct replication of the procedures described in the current investigation should be considered to further evaluate the generality of choice procedures in academic settings and to more clearly elucidate the operant me chanism(s), which may account for behavior change. Future researchers should certainly conduct pre-session preference assessments if interested in learning more regarding the role of choice and preference. Also if the following study were to be replicated, it is recommended that a single session be co nducted daily to minimize the probability of confounds (e.g., sequence effect s, multiple treatment interference) that could result from conducting sessions in rapid succession. Those interested in replicating the current study may also choose to include a more precise control condition. For example, the current st udy used a free operant baseline condition wherein the participants teacher assi gned one of the 5 assignment types used throughout the study. However, during 100% of baseline sessions for Sam and Tony, the teacher selected short answer assignments, which were found to be
38 amongst the least preferred type of assignm ent. An alternative and more precise control condition would involv e having the teacher randomly assign one of the 5 assignment types per session until eac h had been presented at least once. Finally, given the experimental component ( analysis of choice versus preference) of the current investigation, it may be of interest to future researchers to conduct such a study in a more controlled setting. Although applied research is important it is certainly not without its challenges and can sometimes hinder ones ability to draw definitive conclusions regarding t he relation between variables due to the lack of adequate control. Therefore, only af ter the role of choice and preference has been clearly identified should res earchers apply choice procedures with typically developing students in tradition al classroom settings so that the feasibility/generality of the choice interventions can be further assessed. In addition to altering the procedures of the current study as described above, there remains several areas of rese arch related to choice interventions that require further investigation. Fo r example, It may be of interest to researchers/clinicians for extend the wo rk of Romaniuk, Miltenberger, Conyers, Jenner, Jurgens, and Ringenberg (2002) by further assessing the relation between the efficacy of choice procedures and behavioral function. The current investigation did not identify the func tion of each participants targeted behavior prior to the start of the study. Therefore, behavioral function may account for the discrepancies in outcomes observed across participants (e.g. moderate-small outcomes for Tony; small outcomes for Sam).
39 An additional area worthy of research in the future is one wherein the efficacy of choice procedures is co mpared across a variety of parameters including participant diagnosis, adjunctive behavioral measures (e.g., indices of student attitudes toward school, percept ions of autonomy, etc.) and behavioral intensity, to name a few. Such parametri c research may provide clinicians with a guide to the effective and efficient management of behavior displayed by individuals with distinct needs.
40 REFERENCES Bambara, L. M., Ager, C., & Koger, F. (1994). The effects of choice and task preference on the work performance of adults with severe disabilities. Journal of Applied Behavior Analysis, 27 555-556 Dunlap, G., dePerczel, M., Clarke, S., Wilson, D., Wright, S., White, R., & Gomez, A. (1994). Choice making to promote adaptive behavior for students with emotional and behavioral challenges. Journal of Applied Behavior Analysis, 27, 505-518. Fisher, W., Piazza, C. C., Bowman, L. G., Hagopian, L. P ., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying reinforcers for persons with severe and profound disabilities. Journal of Applied Behavior Analysis, 25, 491-498. Kern, L., Vorndran, C.M., Hilt, A., Ringdhal, J.E., Adel man, B.E., & Dunlap, G. (1998). Choice as an intervention to improve behavior: A review of the literature. Journal of Behavioral Education, 8 151-169 Killu, K., Clare, C.M., & Im, A. (1999). Choice vs. Preference: The effects of choice and no choice of preferred and non-preferred spelling tasks on the academic behavior of students with disabilities. Journal of Behavioral Education, 9 239-253 Kincaid, D. (1996). Person-centered planning. In L.K. Koegel, R.L. Koegel, & G. Dunlap (Eds.), Positive behavioral support: Incl uding people with difficult
41 behavior in the community (pp. 439-465). Baltim ore, MD: Paul H. Brookes. Michael, J. (1993) Establishing Operation The Behavior Analyst 16, 191-206 Parsons, M. B., Reid, D. H., Reynolds, J., & Bumgarner, M. (1990). Effects of Chosen versus assigned jobs on the work performance of persons with severe handicaps. Journal of Applied Behavior Analysis, 23, 253-258 Powell, S., & Nelson, B. (1997). Effects of choosing academic assignments on a student with attention deficit hyperactivity disorder Journal of Applied Behavior Analysis, 30, 181-183. Romaniuk, C., Miltenberger, R., Conyer s, C., Jenner, N., Jurgens, M., & Ringenberg, C. (2002). The influence of activity choice on problem behaviors maintained by escape versus attention. Journal of Applied Behavior Analysis, 35 349-362. Seybert, Dunlap, and Ferro (1996). The ef fects of choice-making on the problem behaviors of high-school students wit h intellectual disabilities. Journal of Behavioral Education, 6 49-65 Skinner, B.F. (1971). Beyond Freedom and Dignity. New York: Knopf.
42 Appendix A Partial/Whole Interval Data Collection Participant Name _______ Session Number ______ IOA: Y or N Date Interval Number 1 2 3 4 5 6 7 8 9 10 On-task (W) In Seat (W) Maladaptive Behavior #1 (P) Maladaptive Behavior #2 (P) Teacher Interaction Session Number ______ IOA: Y or N Date Interval Number 1 2 3 4 5 6 7 8 9 10 On-task (W) In Seat (W) Maladaptive Behavior #1 (P) Maladaptive Behavior #2 (P) Teacher Interaction Session Number ______ IOA: Y or N Date Interval Number 1 2 3 4 5 6 7 8 9 10 On-task (W) In Seat (W) Maladaptive Behavior #1 (P) Maladaptive Behavior #2 (P) Teacher Interaction
43 Appendix B Social Validity Questionnaire Name: ____________ Circle one: Pre-Study or Post Study Need for Services (Pre-Study Only) 1) Do you feel that you (or the st udent) are in need improved behavior during academic contexts? 2) In which academic subject do you (or the student ) need improvement? 3) Which behavior, specifically, would you like to see improved? (Circle all that apply) a) On-Task b) In-Seat c) Academic Performance (i.e., better grades) d) Aggression e) Inappropriate Ve rbalizations f) List Others: ____________ Acceptability of Procedures (Pre-Study and Post-Study) 1) Would you like to be provided (or provide) choices with respect to academic assignments? 2) Do you think being provided (or providing) a choice of academic assignment would result in a positive change in your (or the students) behavior? 3) Teachers Only: Do you think it would be feasible to provide students with a choice of academic assignment? 4) Teachers Only: Do you think you would be willing to provide students with a choice of academic assignment?
44 Acceptability of Outcomes (Post-Study Only) 1) Did you notice a change in your (o r the students) behavior at any point during the study? 2) Which of the following conditions do you believe resulted in the greatest change in your (or the students) behav ior? Circle all that apply. ChoiceHigh Preference Choice-Low Preference No Choice High Preference No ChoiceLow preference 4) Which behavior, specifically, do you think was improved? (Circle all that apply) a) On-Task b) In-Seat c) Academic Perf. (i.e., better grades) d) Aggression e) Inappropriate Ver balizations f) List Others: