Assessing preferences for positive and negative reinforcement during treatment of destructive behavior with functional communication training

Research in Developmental Disabilities 26 (2005) 153–168

Assessing preferences for positive and negative reinforcement during treatment of destructive behavior with functional communication training Wayne W. Fisher*, John D. Adelinis, Valerie M. Volkert, Kris M. Keeney, Pamela L. Neidert, Alyson Hovanetz Marcus Behavior Center, Marcus and Kennedy Krieger Institutes, Johns Hopkins University Schools of Medicine, 1920 Briarcliff Rd., Atlanta, GA 30329, USA Received 18 August 2003; received in revised form 10 December 2003; accepted 23 January 2004

Abstract Results of prior studies (e.g. [J. Appl. Behav. Anal. 32 (1999) 285]) showing that participants chose alternative behavior (compliance) over escape-reinforced destructive behavior when this latter response produced escape and the former response produced positive reinforcement may have been due to (a) the value of the positive reinforcer overriding the value of the negative reinforcer or (b) the presence of the positive reinforcer altering the value of the negative reinforcer (i.e., lessening the aversiveness of the demands). In this investigation we evaluated the relative contributions of these alternative mechanisms with two girls with autism. We compared the relative effects of positive and negative reinforcement using equivalent communication responses under both a restricted-choice condition (in which participants could choose positive or negative reinforcement, but not both) and an unrestricted-choice condition (in which participants could choose one or both reinforcers). Both participants often chose positive over negative reinforcement in the restricted-choice condition. However, in the unrestricted-choice condition (in which participants could choose one or both reinforcers), one participant consistently chose both reinforcers by the end of the analysis whereas the other primarily chose only positive reinforcement. Results suggested that for one participant the value of the positive reinforcer overrode the value of the negative reinforcer, whereas for the other participant, the presence of the positive reinforcer in the demand context lessened the aversiveness of the demands. # 2004 Elsevier Ltd. All rights reserved. Keywords: Autism; Choice; Functional analysis; Positive reinforcement; Negative reinforcement

* Corresponding author. Tel.: þ1 404 419 4454. E-mail address: [email protected] (W.W. Fisher).

0891-4222/$ – see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ridd.2004.01.007

154

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

The approach to the treatment of aberrant behavior that is increasingly becoming the standard of care begins with a functional analysis to identify the reinforcer(s) for the response, which then leads to treatments that generally manipulate that reinforcer in ways that decrease the problem behavior. For example, a common approach to the treatment of aberrant behavior maintained by negative reinforcement is to eliminate the escape contingency that was shown to reinforce the response during the functional analysis, a treatment referred to as escape extinction (EE; Iwata, Pace, Kalsher, Cowdery, & Cataldo, 1990; Kelley, Piazza, Fisher, & Oberdorff, 2003). Although EE is generally effective as treatment for aberrant behavior maintained by escape, it is sometimes associated with negative side effects (e.g., bursts of the target behavior, induction of aggression or negative vocalizations; Goh & Iwata, 1994; Lerman & Iwata, 1996; Lovaas, Freitag, Gold, & Kassorla, 1965; Piazza, Patel, Gulotta, Sevin, & Layer, 2003). A variety of procedures have been combined with EE in order to lessen the negative side effects, including differential positive reinforcement of alternative behavior (DPRA; Patel, Piazza, Martinez, Volkert, & Santana, 2002), differential negative reinforcement of alternative behavior (DNRA; Marcus & Vollmer, 1995), noncontingent reinforcement (Fisher, DeLeon, RodriguezCatter, & Keeney, 2004; Vollmer, Marcus, & Ringdahl, 1995), or stimulus (or demand) fading (Pace, Ivancic, & Jefferson, 1994), but bursts may still occur (see Lerman & Iwata for a discussion). Another promising approach to the treatment of aberrant behavior maintained by negative reinforcement involves the identification of potent positive reinforcers that are then placed in direct competition with the escape contingency for problem behavior through DPRA (without extinction; Adelinis, Piazza, & Goh, 2001; DeLeon, Neidert, Anders, & Rodriguez-Cater, 2001; Lalli, Casey, & Kates, 1995; Lalli et al., 1999; Piazza et al., 1997). If the individual consistently chooses the positive reinforcer associated with the alternative response over the negative reinforcer associated with problem behavior, then EE becomes unnecessary. But even if the individual chooses the positive reinforcer over the negative reinforcer some of the time (and the alternative response increases), then EE will generally be implemented less often. Using this approach, Piazza et al. showed that the destructive behavior displayed by three participants was reinforced by escape (as well as positive reinforcement). For one participant, providing the same escape contingency for alternative behavior (compliance) and destructive behavior increased the alternative response and decreased destructive behavior to near-zero levels. For the other two participants, destructive behavior maintained at high rates when both responses produced escape. However, destructive behavior decreased with these latter two participants when the alternative response produced a break with access to positive reinforcement whereas destructive behavior only produced escape. A limitation of the Piazza et al. (1997) study was that it was unclear in these latter two cases whether it was necessary to provide escape for the alternative response or whether reinforcing this response only with positive reinforcement would have produced similar reductions in destructive behavior. Lalli et al. (1999) addressed this limitation with five participants who displayed escape-reinforced problem behavior. These investigators reinforced the alternative response (compliance) with positive reinforcement (food) while the escape contingency remained intact. That is, when participants displayed destructive behavior, they received a 30 s break, whereas when they displayed the alternative response,

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

155

food was delivered and the presentation of demands continued. In all five cases, destructive behavior decreased and the alternative response increased when the alternative response produced positive reinforcement and destructive behavior produced escape relative to when both responses produced the same negative reinforcer (i.e., escape for 30 s). Lalli et al. (1999) suggested that their findings might be explained in terms of the principles of choice responding wherein the alternative response and destructive behavior were concurrent operants and the participants’ motivation for food was often greater than their motivation for escape, and thus they frequently emitted the response (compliance) associated with food. With this explanation, the value of escape was not altered; rather, its value was momentarily overridden by the value of food (i.e., the demands were tolerated, and compliance was emitted during most trials in order to get the food). An alternative explanation was also briefly mentioned by Lalli et al. (see the last paragraph of Section 3), one in which the presence of positive reinforcement acted as a motivating operation (MO) that momentarily altered (i.e., abolished) the value of escape and thus decreased its effectiveness as negative reinforcement for destructive behavior (Laraway, Snycerski, Michael, & Poling, 2003).1 Lalli et al. presented these two explanations as alternatives (reductions in problem behavior due to positive reinforcement overriding versus altering the value of escape), but its possible that one or both are operative with different children. It may be important to distinguish between these two alternative explanations for the effects of competing positive reinforcement on escape-reinforced problem behavior, because if problem behavior decreases solely because the positive reinforcer overrides the negative reinforcer, then one might expect destructive behavior to increase over time. That is, the participants may eventually learn that by displaying both the alternative response and then destructive behavior, it is possible to obtain both reinforcers, which would result in an increase in problem behavior. By contrast, if the presence of the positive reinforcer altered rather than overrode the value of the negative reinforcer, then destructive behavior should remain relatively low. In the current investigation we attempted to separate the effects of these two mechanisms (the positive reinforcer overriding v altering the value of the negative reinforcer) with two girls with autism. We compared the relative effects of positive and negative reinforcement using equivalent and novel communication responses under two choice arrangements. We used equivalent and novel communication behaviors as the operants to eliminate the potential effects of response bias (cf. DeLeon, Fisher, Herman, & Crosland, 2000; DeLeon et al., 2001) and reinforcement histories (Lattal & Neef, 1996). In addition, we evaluated participant choices for positive and negative reinforcement in both a restricted-choice condition (in which participants could choose positive or negative reinforcement, but not both) and an unrestricted-choice condition (in which the participants could choose one or 1 Laraway et al. (2003) have proposed the concept ‘‘motivating operation’’ (in place of establishing operation) as a general term to describe antecedent events that (a) alter the value of reinforcers and punishers and (b) alter the probability of behaviors that historically have been influenced by those consequences. Antecedents that increase the effectiveness of a consequence (a reinforcer or a punisher) are called establishing operations, whereas those that decrease the effectiveness of a consequence are called abolishing operations. We have tried to use the terms recommended by Laraway et al. in this manuscript.

156

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

both reinforcers). If participants chose positive reinforcement over negative reinforcement in the restricted-choice arrangement but chose both reinforcers in the unrestricted-choice arrangement, it would support the hypothesis that the value of the positive reinforcer overrode the value of the negative reinforcer. However, if the participants chose positive reinforcement over negative reinforcement in both the restricted- and unrestricted-choice arrangements, it would support the hypothesis that the positive reinforcer altered the value of the escape contingency.

1. Method 1.1. Participant, materials, and setting Betty, age 13, and Kim, age 14, were admitted to a day-treatment program for the assessment and treatment of severe destructive behavior. Betty was diagnosed with moderate to severe mental retardation and autism. Kim’s diagnoses were chromosomal abnormality and autism. At the time of their respective admissions, they both displayed a few gestures but did not speak (e.g., pointing to mand for a preferred item), and they both followed a few simple commands (e.g., ‘‘Stand up’’). Sessions were conducted either in a treatment room or the corner of a classroom (last phase with Kim) equipped with chairs, a table, and other stimuli (e.g., toys, work materials) needed for the condition in effect. During treatment sessions, two communication cards (approximately 10 cm  10 cm) were available. For both participants, one card was red and had the word ‘‘break’’ printed on it (the SR card). For Betty, the other card had a picture of her playing with another individual (the SRþ card). For Kim, the SRþ card had a picture of food on it. 1.2. Response measurement and inter-observer agreement Observers collected data on destructive behavior and communication responses on laptop computers. Betty’s destructive behavior included aggression (hitting, kicking, pulling, pushing, pinching, or biting others), property destruction (banging, kicking, ripping, or throwing objects), and self-injurious behavior (SIB; self-hitting, body slamming, or headbanging). Kim’s destructive behavior included aggression (slapping, scratching, hitting, kicking, biting, or headbutting others), property destruction (throwing, banging, or forcefully swiping items off of tables and counters), and SIB (headbanging, hitting head with objects). Functional communication for positive reinforcement (Com/Srþ) was defined as handing the SRþ card to the therapist and functional communication for negative reinforcement (Com/Sr) was defined as handing the SR card to the therapist. A second observer independently collected data on 35% of functional analysis sessions and 33.3% of treatment-analysis sessions. Exact agreement coefficients were calculated by comparing observer agreement on the exact number of occurrences or nonoccurrences of a response during each 10 s interval of a session; an agreement was scored if both observers recorded exactly the same number of responses in an interval. Agreement coefficients were computed by dividing the number of intervals with agreements by the total number of intervals in a session, and then multiplying the quotient by 100%. Average agreement

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

157

coefficients for Betty were: aggression, 95.7%; property destruction, 97.7%; SIB, 99.02%; and communication, 98.6%. Average agreement coefficients for Kim were: aggression, 98.9%; property destruction, 97.8%; SIB, 97.6%; Com/SRþ, 98%; and Com/SR, 99.6%. 1.3. Procedures 1.3.1. Functional analysis The functional analyses for both participants were based on the procedures described by Iwata, Pace, Cowdery, and Miltenberger (1994) and Iwata, Wong, Riordan, Dorsey, and Lau (1982). All sessions lasted 10 min and were conducted in a quasi-random order in accordance with a multielement design. In addition, with Betty, we conducted (a) a series of ignore conditions and (b) a pair-wise comparison between the demand and toy-play conditions. The series of ignore sessions was run to rule out the possibility that Betty’s destructive behavior was maintained by automatic reinforcement. The pair-wise analysis between demand and toy play was conducted because Betty was observed to have high rates of destructive behavior in the classroom during certain school tasks and those tasks were integrated into this pair-wise analysis. During the demand condition, the participant and the therapist were seated in chairs at a table and academic tasks from the participants’ individual education plan were presented to the participant using a graduated prompting sequence (successive spoken, modeled, and physical prompts). The participant was given 5 s to respond after each successive prompt and the sequence continued until the participant: (a) displayed a destructive behavior (which resulted in a 30 s break from the task), (b) correctly completed the task without physical prompts (which resulted in brief praise and presentation of the next task) or (c) was guided by the therapist to complete the task (which was followed by presentation of the next task). In the attention condition, the participant had access to moderately preferred stimuli (identified in a prior preference assessment as described by Fisher et al., 1992). A therapist sat in a chair in the room reading a magazine and ignored the participant unless a destructive behavior occurred. Contingent on the occurrence of destructive behavior, the therapist delivered brief attention, which consisted of statements of disapproval and concern (e.g., ‘‘Don’t do that. You’ll hurt yourself.’’) and physical interaction (e.g., touching the participant on the shoulder). In the ignore condition, the therapist was present in the room with the participant but provided no interaction. No alternative stimuli were available in the room. In the tangible condition, the participant was given access to a preferred toy (identified in a prior preference assessment as described by Fisher et al., 1992) for 2 min prior to the onset of the session. When the session began, the therapist removed the toy and returned it to the participant for approximately 20 s contingent on destructive behavior. In the toy-play condition, the therapist was present in the room with the participant and high preference items (identified via the preference assessment) were continuously available. The therapist provided attention (e.g., saying ‘‘You’re playing nicely.’’) approximately once every 30 s after a 5 s interval in which no destructive behavior occurred. Otherwise, destructive behavior was ignored.

158

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

1.3.2. Communication training Prior to the treatment analysis, the participants were trained to emit each of the communication responses using sequential vocal, modeled, and physical prompts. Each communication response was trained separately in different sessions, and each training session consisted of a block of 10 discrete training trials. Training sessions continued for each response until that response was emitted independently during at least 9 of 10 trials for two successive sessions. Training trials for Com/SRþ began with the therapist and child sitting in chairs on the same side of a table with the Com/SRþ card placed on the table about 0.3 m in front of the child. The therapist waited 5 s at the start of each trial for the child to emit the Com/SRþ response independently. If the child did not, the therapist provided a vocal prompt, ‘‘Hand me the card.’’ If the child still did not emit the communication response, the therapist modeled the appropriate response while saying, ‘‘Hand me the card, like this. You do it.’’ If the child still did not emit the communication response, the therapist physically guided the child to complete the response while repeating the first vocal prompt, and then immediately began the next trial. If the child emitted the Com/SRþ response independently or after the vocal or modeled prompt, the therapist delivered the positive reinforcer, which was 30 s of social attention for Betty (e.g., praise, pats on the shoulder or one bite of food for Kim). These stimuli were selected as positive reinforcers for the participants because they had previously been effective at increasing at increasing compliance with demands. Training trials to teach the Com/SR response also began with the child and therapist seated on the same side of the table with the communication card between them. Each trial began with the therapist presenting a vocal prompt that presented a choice to the child to either complete a task (as in demand condition of the functional analysis) or to emit the Com/SR response (e.g., ‘‘Fold the towel or hand me the card if you want a break’’). If neither response was emitted within 5 s of the vocal prompt, the therapist presented a modeled prompt to emit the Com/SR response while saying, ‘‘Hand me the card like this if you want a break.’’ If neither response was emitted within 5 s of the modeled prompt, the therapist physically guided the child to hand over the Com/SR card and then immediately presented the next task. Completing the task at any point in the prompting sequence resulted in brief praise followed immediately by presentation of the next task. Handing the Com/SR to the therapist either following the vocal or modeled prompt resulted in immediate removal of all prompts and task materials for 30 s. 1.3.3. Treatment analysis The treatment analysis began after each participant demonstrated mastery-level performance for both communication responses (i.e., independent performance on 90% of the trials for two consecutive sessions). This analysis consisted of three conditions, baseline, restricted choice and unrestricted choice. In addition, when the restricted- and unrestricted-choice conditions failed to reduce destructive behavior for Betty, a 30 s basket-hold time-out (which was implemented contingent on destructive behavior) was added to these two conditions. The basket-hold time-out was selected as a consequence for destructive behavior because it was previously effective at reducing dangerous acts. The baseline and treatment conditions were compared using a

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

159

multielement design for Kim and in a combination reversal and multielement design for Betty. All sessions lasted 10 min. The baseline condition was identical to the demand condition of the functional analysis. The restricted-choice condition was identical to baseline except that (a) destructive behavior was placed on extinction (i.e., it no longer produced escape), (b) the SRþ and SR communication cards were placed on the table in easy reach of the participant, (c) handing the SR card produced a break from demands for 30 s, and (d) handing the SRþ card resulted in 30 s of continuous attention for Betty (e.g., praise, pats on the back, while the prompts to complete the task continued) or food for Kim (one bite every 10 s, while the prompts to complete the task continued). If the participant handed one of the cards to the therapist, both cards were removed during the 30 s reinforcement interval. The unrestricted-choice condition was identical to the restricted-choice condition except that when the participant handed one card to the therapist, that card was removed during its subsequent reinforcement interval while the other card remained on the table. That is, the reinforcement schedules for Com/SRþ and Com/SR were completely independent in the unrestricted-choice condition (e.g., the Com/SRþ card was replaced after its 30 s reinforcement interval even if a reinforcement interval for Com/SR remained in effect). Thus, the participant could choose positive reinforcement or negative reinforcement (but not both at the same time) in the restricted-choice condition, whereas in the unrestrictedchoice condition the participant could choose either or both reinforcers at the same time, at different times, or at partially overlapping times. For Betty, a basket-hold time-out was added to the restricted- and unrestricted-choice conditions. This consequence lasted 30 s and was delivered contingent on destructive behavior on a FR 1 schedule. For Kim, a final phase was conducted in her classroom in which the card for the negative reinforcement was available but the card for positive reinforcement was not. This phase was included to show that her motivation to escape demands returned when positive reinforcer was unavailable.

2. Results 2.1. Functional analysis The top panel of Fig. 1 shows the results of the functional analysis conducted with Kim. As can be seen, Kim displayed the highest rates of destructive behavior in the demand condition (M ¼ 2.3) followed by social attention (M ¼ 1.1), tangible (M ¼ 0.9), alone (M ¼ 0.4), and toy play (M ¼ 0.3). These results suggest that Kim’s destructive behavior was, at least in part, reinforced by escape from demands. The middle panel of Fig. 1 shows the initial functional analysis conducted with Betty in the first phase and the series of ignore sessions in the second phase. During the first phase, Betty showed high rates of destructive behavior in the attention (M ¼ 3.3) and ignore condition (M ¼ 3.4) and much lower rates in the toy play (M ¼ 0.2), tangible (M ¼ 0.4) and demand (M ¼ 0.5) conditions. In the second phase, destructive behavior decreased to near-zero levels for the last three sessions. The bottom panel of Fig. 1 shows that Betty

160

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

Fig. 1. Destructive responses per minute during the functional analyses for Kim (top panel) and Betty (middle and bottom panels).

displayed higher rates of destructive behavior in the demand condition (M ¼2.7) relative to the toy play condition (M ¼ 0.8) when certain classroom tasks were incorporated into the demand sessions. Taken together, these results suggest that Betty’s destructive behavior was reinforced by attention and escape from certain tasks.

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

161

Fig. 2. Destructive responses per minute during the treatment analyses for Kim (top panel) and Betty (bottom panel).

2.2. Treatment analysis Fig. 2 shows the rates of destructive behavior during baseline and the restricted- and unrestricted-choice conditions for Kim (top panel) and Betty (bottom panel). As can be seen Kim showed relatively high rates of destructive behavior in baseline (M ¼ 1.1) and lower rates in both the restricted- (M ¼ 0.2) and unrestricted-choice (M ¼ 0.3) conditions. For Betty, rates of destructive behavior were high during baseline (M ¼ 2.2) and maintained at similar levels when the restricted- (M ¼ 2.1) and unrestricted- (M ¼ 2.3) choice conditions were introduced without the basket-hold time-out. The addition of the basket-hold time-out reduced destructive behavior to a mean of 0.3 in the restrictedchoice condition and to 0.2 in the unrestricted choice condition. Fig. 3 shows the rates Com/SRþ and Com/SR in the restricted- and unrestrictedchoice conditions for Kim (top panel, first phase) and Betty (bottom panel). Kim displayed high rates of Com/SRþ and low rates of Com/SR in the restricted choice condition, indicating a clear preference for positive over negative reinforcement. Interestingly, in the unrestricted-choice condition (in which she could choose positive reinforcement, negative

162

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

Fig. 3. Communication responses per minute for positive and negative reinforcement in the restricted- and unrestricted-choice condition during the treatment analysis for Kim (top panel) and during the choice þ baskethold phases of the treatment analysis for Betty (bottom panel).

reinforcement, or both), Kim emitted high rates of Com/SRþ and moderate rates of Com/ SR initially, but the rates of Com/SR decreased to near-zero levels whereas rates of Com/SRþ remained high throughout the analysis, suggesting that the presence of the positive reinforcer decreased the value of the negative reinforcer for Kim and reduced motivation to escape demands. This interpretation is further supported by the fact that rates of Com/SR increased markedly in Kim’s classroom when only the negative reinforcer was available (i.e., the card for Com/SR was present but the card for Com/SRþ was not; top panel, second phase). Betty’s choices were more variable and interpretation of her data is somewhat more complicated. In the first phase of FCT þ baskethold (see the bottom panel of Fig. 3), Betty responded almost exclusively for the attention card in the restricted-choice condition (96% of her responses were for the attention card and she responded exclusively for the attention card in seven of the eight sessions). She also showed a preference for attention in the unrestricted-choice condition during the first FCT þ baskethold phase (75% of her

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

163

responses were for the attention card and she responded more for the attention card than for the escape card in all eight sessions). However, her patterns of responding changed in both the restricted- and unrestricted-choice conditions in the second phase of FCT þ baskethold. During the second FCT þ baskethold phase, Betty initially showed a clear preference for attention in the restricted-choice condition (87% of her responses were for attention during the first four sessions); then her responding in the restricted-choice condition became more variable in the middle of this phase (52% of her responses were for escape during the middle four restricted-choice sessions); and finally, at the end of the second FCT þ baskethold phase, she showed a preference for negative over positive reinforcement in the restricted-choice condition (70% of her responses were for escape). In the unrestrictedchoice condition during the second FCT þ baskethold phase, Betty showed a preference for attention over escape during the first two sessions (Sessions 37 and 39). Then, beginning in Session 42, Betty appeared to learn that she could access both attention and escape in the unrestricted-choice condition. Prior to this session, she rarely chose both reinforcers simultaneously. However, after this session, she regularly picked up one card then immediately picked up the other card in the unrestricted-choice condition. This pattern of picking up one card and then immediately picking up the other card in the unrestricted-choice condition persisted throughout the remainder of the second FCT þ baskethold phase. It is important to note that Betty began responding for both positive and negative reinforcement in the unrestricted-choice condition while she still showed a clear preference for positive reinforcement in the restricted-choice condition (early in the second FCT þ baskethold phase). In addition, she continued to respond for both positive and negative reinforcement in the unrestricted-choice condition at times when she showed a preference for negative reinforcement in the restricted-choice condition (late in that same phase). This suggests that for Betty, the availability of contingent attention in the demand context did not substantially lessen (or abolish) motivation for escape in the unrestricted-choice condition, even when the positive reinforcer was preferred over the negative reinforcer early on during the second FCT þ baskethold phase (the opposite effect than was shown for Kim).

3. Discussion In the current investigation, destructive behavior displayed by two girls with autism was treated using a package that included differential reinforcement in which they could choose positive or negative reinforcement (but not both) in one condition (restricted choice) and choose one or both forms of reinforcement in the other condition (unrestricted choice). For Kim, these differential reinforcement treatments reduced destructive behavior when combined with extinction, whereas for Betty, destructive behavior decreased significantly only when a punishment component (basket-hold time-out) was added. These results are consistent with prior research on the effects of FCT when combined with extinction and/or punishment (e.g., Hagopian, Fisher, Sullivan, Acquisto, & LeBlanc, 1998). Comparisons of the communication (or choice) responses across the restricted- and unrestricted-choice conditions produced interesting and quite different patterns for the two girls. Kim almost exclusively chose positive over negative reinforcement in

164

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

the restricted-choice condition, when she could choose one but not both reinforcers, indicating a clear preference for positive over negative reinforcement. Moreover, when she could choose both reinforcers simultaneously in the unrestricted-choice condition, Kim chose positive reinforcement at high rates throughout the phase and negative reinforcement at moderate levels initially, which decreased to zero over time, suggesting that the availability of positive reinforcement altered (lessened) her motivation to escape demands. That is, Kim could have simultaneously chose positive and negative reinforcement by emitting both Com/SRþ and Com/SR, but instead she frequently chose positive reinforcement and continued working, suggesting that when positive reinforcement was available, the work was not sufficiently aversive to maintain Com/SR. Betty showed clear preference for positive over negative reinforcement in the restrictedchoice condition during the first phase and the beginning of the second phase of FCT þ baskethold, but then her preferences became more variable and toward the end of the analysis and her preference shifted toward negative reinforcement. Betty also showed a preference for positive over negative reinforcement in the unrestricted-choice condition initially, but beginning in Session 42 and for the remainder of the treatment analysis, she consistently chose both positive and negative reinforcement. Our interpretation of Betty’s data is that she did not discriminate that she could obtain both positive and negative reinforcement simultaneously in the unrestricted-choice condition until the second phase of FCT þ baskethold (beginning in Session 42). After this point, she consistently chose both positive and negative reinforcement in the unrestrictedchoice condition, while her preference shifted in the restricted-choice condition. If our interpretation is correct (that she initially failed to discriminate the contingencies in the unrestricted-choice condition), then Betty’s motivation for escape was not substantially altered by the availability of attention in the demand context (unlike Kim). The differences in responding between Kim and Betty may have also been influenced by the fact that different positive reinforcers were used with the two participants (food for Kim; attention for Betty). Another interpretation of Betty’s data is that she accurately discriminated that both attention and escape were available in the unrestricted-choice condition throughout the analysis, but that initially, when she displayed a clear preference for attention, the availability of attention acted as an abolishing operation and lessened her motivation for escape. In addition, it is possible that as her preference shifted toward negative reinforcement, she began choosing both reinforcers in the unrestricted-choice condition because the availability of attention no longer altered Betty’s motivation for escape. This interpretation seems much less plausible to us, because Betty began consistently choosing both reinforcers in the unrestricted-choice condition (in Session 42) before her preference began to shift in the restricted-choice condition (in Session 46). Thus, during the initial portion of the second phase of FCT þ baskethold, Betty showed a clear preference for attention over escape in the restricted-choice condition while she chose both positive and negative reinforcement in the unrestricted-choice condition (indicating that the availability of attention did not alter her motivation for escape, at least during this period). The results of the current investigation extend prior research involving the treatment of problem behavior through DPRA (DeLeon et al., 2001; Lalli et al., 1995, 1999; Piazza et al., 1997). First, we evaluated the participants’ relative preferences for positive and

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

165

negative reinforcement using topographically equivalent responses (both picture-exchange communication responses). In most previous research involving competition between positive and negative reinforcement (DeLeon et al., 2001; Lalli et al., 1995, 1999; Piazza et al., 1997), one response (destructive behavior) was reinforced by negative reinforcement and a topographically distinct response (i.e., communication or compliance) was reinforced by positive reinforcement. Evaluating relative preferences for two different reinforcers (positive and negative) with two topographically distinct responses (e.g., compliance and destructive behavior) may not always produce comparable results. For example, one of the participants in the Piazza et al. (1997) study continued to show relatively high rates of destructive behavior and only moderate levels of compliance when the former response produced only a break whereas the latter response produced a break plus a tangible reinforcer. Similarly DeLeon et al. (2000) presented data on an individual who chose destructive behavior over communication, even when the former response produced reinforcement at a rate that was one-half the rate of reinforcement for the latter response. Such response biases may be due to differences in response effort or the reinforcement histories of the two different topographies. The finding that competing positive reinforcers sometimes produce their reductive effects on escape-reinforced problem behavior by acting as a MO and altering the value of escape may be important because most treatments for problem behavior using abolishing operations have involved the alteration of task requirements (at least temporarily). That is, treatments based on MO manipulations generally include decreases in the amount or difficulty of work the individual is required to complete (e.g., Horner, Day, Sprague, O’Brien, & Heathfield, 1991; Pace et al., 1994; Vollmer et al., 1995), although there have been notable, yet idiosyncratic exceptions (see the participant Charlie in McComas, Hoch, Paone, & El-Roy, 2000; O’Reilly, Lacey, & Lancioni, 2000). One potential limitation of altering task requirement is that the individual may acquire skills at a slower rate. The approach involving competition between positive and negative reinforcement, as exemplified by Lalli et al. (1999), may offer a more general method for decreasing escapereinforced problem behavior without lowering task requirements and potentially limiting the individual’s skill acquisition. Another potential contribution of the current investigation is that it provided a potentially useful model for determining whether the presence of one reinforcer affects a response associated with a different reinforcer (a) through direct schedule competition, in accordance with the principles of choice responding (e.g., Matching Law; see Fisher & Mazur, 1997 for a discussion); or (b) by acting as an MO and altering the effectiveness of the latter reinforcer (Laraway et al., 2003; Michael, 1982, 2000). That is, if schedule competition is the sole operative mechanism, then an individual should show a preference for one reinforcer when a restricted-choice arrangement is in effect but should choose both reinforcers when the unrestricted-choice choice arrangement is in effect. Betty’s responding appeared to fit this pattern in the second phase of FCT þ baskethold. Conversely, if the presence of one reinforcer acts as an abolishing operation that momentarily alters the effectiveness of the other reinforcer, then this latter reinforcer should have diminished effectiveness in the unrestricted-choice condition (Kim’s responding showed this pattern). In lay terms, the value of escape was not lessened for Betty in the restricted-choice condition; rather, its value was overshadowed by the value of the positive reinforcer

166

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

(attention). But this became clear only through the inclusion of the unrestricted-choice condition, where she could access both reinforcers simultaneously. By contrast, the value of escape was altered by the presence of the positive reinforcer for Kim, as she chose escape with decreasing frequency in the unrestricted-choice condition when she could have chosen the positive and negative reinforcers at the same time. Methods similar to those applied in this study might be used to help determine the operative mechanisms for other interventions as well. For example, one effective intervention for problem behavior referred to as noncontingent reinforcement (NCR) involves the delivery of either the functional or an alternative reinforcer on a fixed-time (FT) schedule. Some investigations have attributed the often immediate and large reductions in destructive behavior produced by NCR to reinforcer satiation, an MO (Vollmer, Iwata, Zarcone, Smith, & Mazaleski, 1993), whereas others have suggested that schedule competition and choice principles provide a better account (e.g., Fisher et al., 1999). We could have examined these two alternative explanations of NCR with the current (or similar) participants by substituting NCR (with food or attention) in place of the positive reinforcement schedules used in the restricted- and unrestricted-choice conditions (e.g., delivery of food on an FT schedule with removal of the escape card during the reinforcement interval in the restricted condition but not in the unrestricted condition). The results of the current investigation should be interpreted with appropriate caution and relative to several limitations. First, there were just two participants in the current investigation. Second, the results would have been strengthened had we included a condition in which only the escape card was present during the multielement comparisons between the restricted- and unrestricted-choice conditions. This was done for Kim after the comparison between the restricted-choice and unrestricted-choice conditions, but it was done in a separate setting (for clinical reasons). Future investigations should replicate these findings with more participants and include an additional control condition in which negative reinforcement, but not positive reinforcement is available through a communication response. Third, Betty displayed destructive behavior that was reinforced not only by escape but also by attention, which was used as the competing positive reinforcer in the demand context. Thus, it is possible that some of the destructive behavior she displayed in the demand context was reinforced by attention. We view this as an unlikely possibility because during demand sessions, the discriminative, establishing, and reinforcing stimuli for escape-maintained destructive behavior were present and the corresponding stimuli for attention-maintained destructive behavior were absent. In addition, Betty never received attention for destructive behavior in any of the demand sessions (functional analysis, baseline, or treatment). Furthermore, one form of attention (delivery of demands) was in direct competition with escape during the functional analysis (as is typically the case), and this is probably why we identified the escape function only when we introduced more difficult demands during the pair-wise functional analysis. If attention reinforced destructive behavior in the demand context, then this effect should have occurred both when easy and when difficult demands were delivered. Nevertheless, future investigators studying competition between positive and negative reinforcement may wish to include only participants whose destructive behavior is solely reinforced by escape to eliminate the potential limitation.

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

167

Acknowledgements This investigation was supported in part by Grant No. MCJ249149-02 from the Maternal and Child Health Service and Grant No. 5 R01 HD37837-03 from the National Institute of Child Health and Human Development within the US Department of Health and Human Services.

References Adelinis, J. D., Piazza, C. C., & Goh, H. (2001). Treatment of multiply controlled destructive behavior with food reinforcement. Journal of Applied Behavior Analysis, 34, 97–100. DeLeon, I. G., Fisher, W. W., Herman, K. M., & Crosland, K. C. (2000). Assessment of a response bias for aggression over functionally equivalent appropriate behavior. Journal of Applied Behavior Analysis, 33, 73–77. DeLeon, I. G., Neidert, P. L., Anders, B. M., & Rodriguez-Catter, V. (2001). Choices between positive and negative reinforcement during treatment for escape-maintained behavior. Journal of Applied Behavior Analysis, 34, 521–525. Fisher, W. W., DeLeon, I. G., Rodriguez-Catter, V., & Keeney, K. M. (2004). Enhancing the effects of extinction on attention-maintained behavior through noncontingent delivery of attention or stimuli identified via a competing stimulus assessment. Journal of Applied Behavior Analysis, 37, 239–242. Fisher, W. W., & Mazur, J. E. (1997). Basic and applied research on choice responding. Journal of Applied Behavior Analysis, 30, 387–410. 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. Fisher, W. W., Thompson, R. H., DeLeon, I. G., Piazza, C. C., Kuhn, D. E., Rodriguez-Catter, V. et al. (1999). Noncontingent reinforcement: Effects of satiation vs. choice responding. Research in Developmental Development, 20, 411–427. Goh, H. L., & Iwata, B. A. (1994). Behavioral persistence and variability during extinction of self-injury maintained by escaped. Journal of Applied Behavior Analysis, 27, 173–174. Hagopian, L. P., Fisher, W. W., Sullivan, M. T., Acquisto, J., & LeBlanc, L. A. (1998). Effectiveness of functional communication training with and without extinction and punishment: A summary of 21 inpatient cases. Journal of Applied Behavior Analysis, 31, 211–235. Horner, R. H., Day, H. M., Sprague, J. R., O’Brien, M., & Heathfield, L. T. (1991). Interspersed requests: A nonaversive procedure for reducing aggression and self-injury during instruction. Journal of Applied Behavior Analysis, 24, 265–278. Iwata, B. A., Pace, G. M., Cowdery, G. E., & Miltenberger, R. G. (1994). What makes extinction work: An analysis of procedural form and function. Journal of Applied Behavior Analysis, 27, 131–144. Iwata, B. A., Pace, G. M., Kalsher, M. J., Cowdery, G. E., & Cataldo, M. F. (1990). Experimental analysis and extinction of self-injurious escape behavior. Journal of Applied Behavior Analysis, 23, 11–27. Iwata, B. A., Wong, S. E., Riordan, M. M., Dorsey, M. F., & Lau, M. M. (1982). Assessment and training of clinical interviewing skills: Analogue analysis and field replication. Journal of Applied Behavior Analysis, 15, 191–203. Kelley, M. E., Piazza, C. C., Fisher, W. W., & Oberdorff, A. J. (2003). Acquisition of cup drinking using previously refused foods as positive and negative reinforcement. Journal of Applied Behavior Analysis, 36, 89–93. Lalli, J. S., Casey, S., & Kates, K. (1995). Reducing escape behavior and increasing task completion with functional communication training, extinction, and response chaining. Journal of Applied Behavior Analysis, 28, 261–268. Lalli, J. S., Vollmer, T. R., Progar, P. R., Wright, C., Borrero, J., Daniel, D. et al. (1999). Competition between positive and negative reinforcement in the treatment of escape behavior. Journal of Applied Behavior Analysis, 32, 285–296.

168

W.W. Fisher et al. / Research in Developmental Disabilities 26 (2005) 153–168

Laraway, S., Snycerski, S., Michael, J., & Poling, A. (2003). Motivating operations and terms to describe them: Some further refinements. Journal of Applied Behavior Analysis, 36, 407–414. Lattal, K. A., & Neef, N. A. (1996). Recent reinforcement-schedule research and applied behavior analysis. Journal of Applied Behavior Analysis, 29, 213–230. Lerman, D. C., & Iwata, B. A. (1996). A methodology for distinguishing between extinction and punishment effects associated with response blocking. Journal of Applied Behavior Analysis, 29, 231–234. Lovaas, O. I., Freitag, G., Gold, V. J., & Kassorla, I. C. (1965). Experimental studies in childhood schizophrenia: Analysis of self-destructive behavior. Journal of Experimental Child Psychology, 2, 67–84. Marcus, B. A., & Vollmer, T. R. (1995). Effects of differential negative reinforcement on disruption and compliance. Journal of Applied Behavior Analysis, 28, 229–230. McComas, J., Hoch, H., Paone, D., & El-Roy, D. (2000). Escape behavior during academic tasks: A preliminary analysis of idiosyncratic establishing operations. Journal of Applied Behavior Analysis, 33, 479–493. Michael, J. (2000). Implications and refinements of the establishing operation concept. Journal of Applied Behavior Analysis, 33, 401–410. O’Reilly, M. F., Lacey, C., & Lancioni, G. E. (2000). Assessment of the influence of background noise on escape-maintained problem behavior and pain behavior in a child with Williams syndrome. Journal of Applied Behavior Analysis, 33, 511–514. Pace, G. M., Ivancic, M. T., & Jefferson, G. (1994). Stimulus fading as treatment for obscenity in a brain-injured adult. Journal of Applied Behavior Analysis, 27, 301–305. Patel, M. R., Piazza, C. C., Martinez, C. J., Volkert, V. M., & Santana, C. M. (2002). An evaluation of two differential reinforcement procedures with escape extinction to treat food refusal. Journal of Applied Behavior Analysis, 35, 363–374. Piazza, C. C., Fisher, W. W., Hanley, G. P., Remick, M. L., Contrucci, S. A., & Aitken, T. L. (1997). The use of positive and negative reinforcement in the treatment of escape-maintained destructive behavior. Journal of Applied Behavior Analysis, 30, 279–298. Piazza, C. C., Patel, M. R., Gulotta, C. S., Sevin, B. S., & Layer, S. A. (2003). On the relative contribution of positive reinforcement and escape extinction in the treatment of food refusal. Journal of Applied Behavior Analysis, 36, 309–324. Vollmer, T. R., Bethany, A., & Ringdahl, J. E. (1995). Noncontingent escape as treatment for self-injurious behavior maintained by negative reinforcement. Journal of Applied Behavior Analysis, 28, 15–26. Vollmer, T. R., Iwata, B. A., Zarcone, J. R., Smith, R. G., & Mazaleski, J. L. (1993). The role of attention in the treatment of attention-maintained self-injurious behavior: Noncontingent reinforcement and differential reinforcement of other behavior. Journal of Applied Behavior Analysis, 26, 9–21.