All differential reinforcement procedures combine the use of what two principles of behavior?

5.05.5.2.1 Differential Reinforcement of Incompatible Behavior

Differential reinforcement of incompatible behavior (DRI) is a weakening procedure that is based on reinforcement. The behavior that is reinforced is not the target behavior, as this would clearly strengthen it, but rather is a behavior that is both desirable and topographically incompatible with the target behavior. The rationale is that the individual cannot simultaneously engage in two incompatible behaviors so as one is strengthened, the other must be weakened.

A DRI procedure could be used to intervene with a child who is continually out of his seat in the classroom. A desirable behavior that is topographically incompatible with roaming around the classroom is that of being seated. The teacher therefore could reduce the out-of-seat behavior by systematically reinforcing the child for being seated. Another example of a behavior that could be treated with DRI is that of a young child who, while watching television, continually “twiddles” with her hair to the extent that bald patches are appearing. Such hair “twiddling” could be reduced by reinforcing the child for cuddling a teddy bear.

The big advantage of DRI is that the focus is on doing something positive, and not simply the removal of an unwanted behavior. It is for this reason that DRI is sometimes referred to as a constructional approach to weakening behavior. One potential procedural difficulty with DRI is that it is not always easy to find an appropriate behavior that is topographically incompatible with the target behavior. Another potential difficulty is that the individual must be able to engage in the incompatible behavior for all of the time they would normally engage in the target behavior. This is usually referred to as complying with the 100% rule. Reinforcing piano playing could be used to treat the preschool child who hits other children, but the child could not be expected to play the piano for the total time in attendance at preschool.

(iii) Social consequences

Assessment may reveal that the aggressive behavior of a dementia patient is contingently associated with social consequences (e.g., attention by caregivers). Differential reinforcement of incompatible behavior (DRI) may be a reasonable option to consider in this situation. To illustrate, Rosberger and MacLean (1983) described the implementation of this procedure in the treatment of a 79 year-old stroke patient residing in a nursing home. The patient was aphasic and partially paralyzed. Problem behaviors included physically aggressive behavior directed at staff, other residents, and visitors, throwing objects, smearing feces, and disrobing in front of visitors. Mechanical restraint and medication (e.g. Haldol and Mellaril) resulted in unacceptable side effects. Following assessment of the behaviors identified as the most disruptive, a DRI intervention was implemented wherein staff were instructed to greet the resident and praise her cooperation whenever she was observed to be behaving appropriately. In addition, problem behaviors resulted in a brief time-out from social attention. The intervention proved effective in that the resident's aggressiveness was reduced from 16 episodes per week to zero by the third week. Treatment effects were maintained at a 5 month follow-up.

A case study involving a 72 year-old man diagnosed with AD also demonstrates the effective contingent application of social consequences for reducing agitated behavior (Bakke et al., 1994). The patient had been prescribed buspirone for the agitation and had been expelled from two day treatment centers previously because of the agitation. On the basis of interview and observational data it was hypothesized that this man's frequent efforts to leave a work program and statements about going home were maintained by consistent attention from staff, while on-task, in-seat behavior resulted in social attention only 15% of the time. The intervention involved increasing the rate of tangible and social reinforcement contingent on working while incorporating efforts to reduce fatigue, discomfort, and boredom through restructuring of the work setting and schedule. Results indicated that this multicomponent program produced a significant decrease in agitated behavior while maintaining work productivity.

Understanding the contingencies of reinforcement associated with the behavior of dementia patients presents unique challenges for researchers and caregivers in that the progression in neuropathology limits the adequacy of traditional standards of treatment generalization and maintenance (Fisher & Carstensen, 1990). As is the case for physiologically based interventions, the effects of environmentally based interventions for dementia patients require continuous monitoring of changes in the reciprocal influence between behavior and the environment associated with deterioration in brain function. Stimulus control procedures, designed to prevent the occurrence of problem behaviors may be the most reasonable approaches to pursue until information regarding the efficacy of other environmentally based interventions is available (Fisher & Carstensen, 1990; Hussian, 1988).

Raters’ knowledge of treatment

Rasnake et al. (1993) evaluated the association of knowledge of behavior principles to treatment acceptability ratings. Participants included 57 directive care staff members employed at an intermediate care facility. A case description was presented to the participants with manipulations of severity levels of self-injurious behavior. Treatment vignettes using six different behavioral interventions (differential of reinforcement of other behavior, differential reinforcement of incompatible behavior, stimulus control, overcorrection, physical restraint, and contingent shock) were rated by participants. The IRP was used to rate treatment acceptability. Participants also completed a 25-item multiple forced-choice instrument as a means of measuring prior knowledge of behavioral principles. The data were analyzed using Pearson correlation coefficients. The researchers found an increased knowledge of behavioral principles did not result in increased treatment acceptability ratings. In addition, the stimulus control intervention was the only intervention to receive an overall mean acceptability ratings at the “acceptable” level.

Fairbanks and Stinnett (1997) evaluated treatment acceptability of different behavior interventions associated with different diagnostic labels as rated by members of different professional groups. A vignette was presented to 31 teachers, 33 school psychologists, and 33 school social workers from three school districts in southwestern Illinois. A copy of the IRP-15 was completed by the participants to rate the acceptability of the proposed treatment described within the vignette. Variables manipulated included the diagnostic label of the child described within the case vignette. Diagnostic labels manipulated were learning disabled (LD), behavior disordered (BD), and Attention Deficit Disorder (ADD). The case vignette described a third-grade boy who displayed behaviors such as excessive talking, out-of-seat, and overly active. A three-way analysis of variance was used to evaluate the data. Findings indicated no significant differences based on the diagnostic label of the child described in the case vignette. Among the members of the different professional groups, teachers rated the negative intervention (time-out from reinforcement with praise) more acceptable than school psychologists or school social workers. Also, school psychologists rated the negative intervention more acceptable than school social workers.

RECOMMENDATIONS FOR INTERVENTION SERVICES

1.

An intensive behavioral intervention plan should be developed between the psychologist and Mrs. Duffy to address Don's social skills deficits and competing problem behaviors.

Specific intervention procedures for Don's escape-motivated behaviors (tantrums, arguments, and talking back) might be effectively addressed by reconsidering and revising the task demands for classroom academic assignments. For example, Mrs. Duffy indicated that Don dislikes tasks requiring handwriting and math calculation. As such, task requirements could be modified by having Don provide answers either verbally in a tape recorder or in a short-answer format. Math calculations could be reduced and supplemented with math word problem tasks. In addition, Mrs. Duffy might consider giving Don choices among activities and interspersing easier tasks with more difficult tasks.

Specific intervention procedures for Don's social attention-motivated behavior of disturbing ongoing activities might include differential reinforcement of incompatible behaviors (DRI) for staying in his seat and completing his assigned work, noncontingent attention of behavior delivered on a time-based schedule (e.g., every 5 minutes), and presentation of preferred activities in the classroom.

For the social skills acquisition deficits of controlling temper and cooperating with peers, it is recommended that Don be placed in a social skills training group designed to remediate acquisition deficits by using modeling, coaching, and behavioral rehearsal strategies. Specific procedures for teaching these skills can be found in the Social Skills Intervention Guide (Elliott & Gresham, 1992), which teaches each social skill found on the SSRS.

For the performance deficits of following directions and making transitions, it is recommended that Mrs. Duffy use precorrection and incidental teaching strategies. For example, Mrs. Duffy could explain exactly what is required and what is going to happen prior to a given activity or transition. Also, when Don does not follow directions or appropriately make transitions, Mrs. Duffy could use these situations to prompt and reinforce correct performances of these skills.

To address Don's social skill acquisition and performance deficits and his competing problem behaviors, it is recommended that Mrs. MacMillan attend a parent training group that teaches fundamental knowledge and practical application of behavior change procedures.

Once the above intervention procedures are agreed upon, specific procedures for monitoring the integrity of the intervention plan and for evaluating the outcomes of these interventions should be developed by the psychologist and Mrs. Duffy.

If these intervention procedures are ineffective in bringing Don's behavior into acceptable levels, Don then should be considered for possible placement into a special education setting to receive more intensive intervention services.

III. Empiral Studies

Although overcorrection procedures have been utilized most commonly with persons with mental retardation in institutional settings, these procedures also have been employed in the treatment of autism, emotional disturbances, and behavior disorders in a variety of settings. Although many of the treatment studies have focused on children, significant numbers of studies have been conducted with adults as well. In their 1982 review of 97 overcorrection studies, Foxx and Bechtel classified the maladaptive behaviors treated with overcorrection techniques into categories of aggressive-disruptive behaviors, self-stimulatory behaviors, self-injurious behaviors, toileting behaviors, inappropriate oral behaviors, and educational-social development behaviors. Historically, the vast majority of applications of overcorrection have occurred with aggressive-disruptive, self-stimulatory, and toileting behaviors.

The initial application of overcorrection procedures occurred as a method of toilet training individuals with mental retardation in institutional settings. An extensive set of procedures known as dry-bed training was used following bowel and bladder accidents. The procedures usually consisted of mopping the floor, cleaning wet and soiled items, redressing oneself in clean clothing and replacing bed linens, and repeatedly walking to the toilet and performing a series of responses (pulling pants down, sitting, etc.) involved in appropriate toileting. These procedures later were modified for application to normal children with greater emphasis placed on the positive practice component (i.e., repeatedly walking to the toilet and rehearsing appropriate toileting behaviors). Similar procedures have been applied to children diagnosed with enuresis or encopresis. Despite some variation across studies and populations treated, the duration of overcorrection with toileting behaviors usually has been 30 min or greater, often ranging up to 45 min.Because overcorrection procedures usually have been combined with other effective procedures such as Mowrer and Mowrer's bell-and-pad training and reinforcement for appropriate voiding, it is not possible to determine the relative contribution of overcorrection to the successful outcomes reported in such multicomponent treatment programs. However, reductions of greater than 80% in wetting and/or soiling usually have been reported, with near 100% reductions often being achieved within 1 to 3 months and maintained at 2- to 18-month follow-up.

With aggressive-disruptive behaviors, overcorrection has been employed rather extensively in treating relatively mild problems, such as out-of-seat behavior or talking out, as well as in treating more extreme acts, such as hitting, biting, and assaultive sexual behavior. Typical overcorrection procedures for aggressive-disruptive behaviors include picking up thrown or ripped items, apologizing to the victim, and/or assisting in medical care of the victim. In a few studies, overcorrection has involved requiring the aggressor to lie down, to pat and stroke the victim, or to engage in a series of arm movements. The duration of these overcorrection procedures has ranged from less than 1 min to 2 hr, with the most frequent durations being 5 to 10 min.In relatively few of these studies has overcorrection been used as the only treatment procedure. Instead, many studies have combined overcorrection with procedures such as verbal warnings and positive reinforcement during periods when the client was not engaged in overcorrection acts. Using DRI (differential reinforcement of incompatible behaviors) or DRO (differential reinforcement of other behaviors) procedures, positive reinforcement has been made contingent either on responses that are incompatible with the target behaviors or on the absence of aggression or disruption for specified intervals. In investigations of overcorrection treatments, reductions of greater than 85% have been observed within 2 weeks to 2 months for most aggressive-disruptive behaviors, with a large number of researchers reporting reductions of near 100%. Maintenance of behavior change has been reported in most studies, with follow-up periods ranging from 5 weeks to 1 year.

In a large number of studies, overcorrection has been used to treat a variety of self-stimulatory behaviors including hand flapping and posturing, stereotyped vocalizations, rocking, hair pulling, and mouthing objects. Behaviors most frequently treated have been hand flapping, rocking, and mouthing. The most commonly used overcorrection procedures for these responses consist of required movement of the body parts involved in the self-stimulatory behaviors. Other common procedures have included enforced toy play and required toothbrushing. The duration of the overcorrection procedures for self-stimulatory behavior has ranged from 30 sec to 20 min, with a typical duration of 2 min.Relatively few treatments for self-stimulatory behaviors have employed overcorrection alone, as most combine overcorrection with other procedures. Additional treatment procedures have included verbal warnings, prevention of self-stimulatory behavior by physical restraint or other means, and/or positive reinforcement (i.e., DRI or DRO procedures). Nearly all investigators reported reductions in target behaviors of greater than 80%, with near 100% reductions observed in the majority of studies. However, follow-up data have been reported in very few studies, with maintenance of behavior reductions rarely reported for longer than 1 to 3 months.

Self-injurious behaviors, such as face slapping, head banging, hand biting, and eye poking and gouging also have been the focus of a relatively small number of overcorrection studies. The most frequently treated self-injurious behaviors have been head banging and biting. Overcorrection procedures for these behaviors usually have consisted of required movement of the body part involved in the self-injury, sometimes combined with required toothbrushing for self-biting, hair combing for head banging, required bed rest, and applying medication or cream to the affected area. The duration of these overcorrection procedures typically has ranged from 5 to 10 min.Overcorrection has been utilized as the only treatment procedure in most studies but has been combined with positive reinforcement of alternate behaviors in a few instances. Reductions in self-injurious behaviors of 95 to 100% have been reported in less than 1 week of treatment for most cases. The majority of studies have reported follow-up data, with maintenance of treatment effects being demonstrated for 4 to 33 months posttreatment.

A handful of investigations have addressed maladaptive oral behaviors in individuals with mental retardation. This category of behaviors includes drooling, vomiting, rumination (the repeated rechewing and swallowing of regurgitated food), pica (the ingestion of nonnutritive substances such as paper or cigarette butts), and coprophagia (the ingestion of fecal material) . Overcorrection procedures have consisted of picking up trash, required practice in correct vomiting, cleaning of vomited matter from various surfaces including walls and floors, and required handwashing, toothbrushing, and mouth wiping. Durations for such procedures have varied considerably, often involving periods of less than 2 min for rumination, drooling, and pica as contrasted with 20 min to 2 hr for coprophagia and vomiting. Brief durations of overcorrection have been used most often combined with other procedures such as DRO and positive reinforcement for appropriate behaviors. Except for drooling, near 100% reduction in these maladaptive oral responses has been reported at posttreatment. The majority of studies conducted follow-up assessments and reported maintenance of these reductions at 3 to 12 months posttreatment.

A limited number of studies have addressed various responses identified by Foxx and Bechtel as educational-social development behaviors. Maladaptive responses in this broad category include errors on academic-related tasks (e.g., oral reading, spelling, writing proficiency, manual signing) and failure to comply with directions/demands to stay on-task, attend class, share with other children, make eye contact, vocalize, eat appropriately, and perform tasks with adequate speed. Overcorrection procedures for these behaviors typically have consisted of requiring clients to repeatedly correct academic errors, complete written academic tasks, comply with verbal instructions, and engage in required movements of specific body parts (e.g., hand movements with eating utensils or puzzle pieces, head movements in the direction of the therapist). Modeling and reinforcement procedures often have been included as treatment components in these studies. Combinations of these procedures usually have resulted in significant decrements in maladaptive responses, as well as significant improvements in compliance with instructions and accurate responding. However, relatively few studies examining the use of overcorrection procedures with social-academic behaviors have addressed issues of maintenance of behavior change.

As with other punishment procedures, the literature on overcorrection is replete with numerous reports (usually anecdotal in nature) of positive and negative side effects. The majority of studies that have provided data-based observations of side effects have examined stereotyped behaviors of a self-stimulatory or self-injurious nature. Associated with overcorrection have been reported increases in prosocial behaviors such as compliance, cooperation, and appropriate toy play, as well as increases in negative responses such as aggression, emotional outbursts, and nontargeted self-stimulatory behaviors.

Application of Reinforcement Principles in Neuropsychological Rehabilitation

Operant principles are perhaps the most widely used among the behavior therapy strategies [18]. Operant procedures are based on the manipulation of consequences, either by providing rewards contingent on appropriate behavior (positive programming) or the withdrawal of rewards or reinforcing environments contingent on inappropriate behaviors (response cost or time out). Reinforcement refers to any stimulus or event whose presence increases the probability of occurrence of a behavior. Reinforcements have been widely used to modify undesirable behaviors such as aggression. Despite the fact that patients with TBI may have cognitive deficits, learning is not hampered when concrete reinforcements are used.

The withdrawal of rewards leads to a decrease in the probability of the occurrence of an undesired behavior. Some commonly used techniques include response cost or withdrawal of rewards or tokens earned, and time out, which involves the removal of the person from a pleasant/neutral environment to a nonreinforcing environment. Response cost is a cognitive behavioral technique based on the operant principle of contingent punishment, or negative punishment, and has been used to enhance behavioral control. Alderman and Burgess [26] report the use of response cost along with cognitive overlearning in a patient with herpes encephalitis, who demonstrated repetitive speech and other problems of poor social skills, as well as incontinence.

Schedules of reinforcement determine the manner in which reinforcements will be delivered. They allow greater discrimination of cues, such as the use of differential reinforcement. These involve the manipulation of antecedent cues and thereby reduction or elimination of behaviors that occur exclusively in response to cues. The success of operant procedures is based on the careful and detailed analysis of the behaviors as well as the settings in which they occur and factors that maintain them. The criterion for success with reference to target behaviors must also be specified (such as frequency of an appropriate behavior and magnitude).

Serno [18] provides a detailed discussion of the main behavior modifications and their applications, as well as limitations to neuropsychological rehabilitation. Some of the main techniques highlighted in this review are the use of positive reinforcement, prompts, cues, and fading in the management of aphasia, and the use of response hierarchies, shaping, and differential reinforcement in the management of stereotyped verbal responses. Programmed therapy is another contribution of the operant school of learning to rehabilitation programs. In a programmed progression, response hierarchies are created, and cues and prompts to be delivered are also determined.

Accelerative techniques are those techniques that increase the frequency or intensity of target behaviors, whereas decelerative behaviors are those that help decrease the frequency or intensity of target behaviors. Accelerative behavioral techniques include positive programming, shaping, and chaining; decelerative techniques include differential reinforcement of incompatible behavior, other behaviors (DRO), and low-rate behaviors, overcorrection, stimulus change, stimulus satiation or massed practice, time out, and response cost. Complex strategies include contingency contracting, stimulus control, and token economy.

The most researched area in the behavioral management of TBI is the management of aggression and other behavioral disturbances that occur as a consequence of TBI, such as self-injurious behaviors. Aggressive behavior is one of the most disturbing consequences and is a serious hindrance to the progress of rehabilitation as it hampers the use of other rehabilitation strategies and interferes with the social functioning of the patient. Frequently used behavioral strategies in the management of aggression and behavioral disturbances include the use of positive reinforcement and modeling [27], differential reinforcement and token economy [28–30], overcorrection and restitution [31,32], and time out and response cost [26]. Some studies have also used a combination of strategies such as self-monitoring of anxiety, cognitive restructuring, and instrumental mastery skills [33]. It must be noted that the most effective behavior therapy programs are a combination of both decelerative and accelerative techniques, rather than the isolated use of strategies. Punishment and other aversive techniques are used based on the principle of the least coercive methods and are rarely used in the absence of positive programming strategies. Thus, while one target behavior is on a deceleration program, such as response cost or time out, a desirable target behavior identified in the ABC analysis is placed on a positive reinforcement schedule. This ensures that while maladaptive behaviors are eliminated or decreased, the patient has an opportunity to develop an adaptive repertoire as well.

Aggression has also been managed by changing antecedents of aggression, decreasing stimulation, increasing predictability by scheduling, signaling an impending event, and approaching a patient from a side that is not affected by visual neglect. Management of inappropriate behaviors, including nonverbal activities such as throat clearing, spitting, hoarding, and verbal behaviors such as screaming, shouting, and complaining, have been effectively managed with straightforward reinforcement methods such as DRO and other methods such as the contingent withdrawal of rewards through time out and response cost.

Alderman and Burgess [26] report the use of response cost along with cognitive overlearning in a patient with herpes encephalitis who demonstrated repetitive speech. Burgess and Alderman [33] describe the use of response cost and training in self-monitoring in patients with prefrontal lobe lesions who had problems such as verbal abuse, poor social skills, and urinary incontinence.

One of the major difficulties with the use of operant procedures in the management of behavioral disturbances resulting from TBI is the maintenance of gains in the natural environment. Often the cues and contingencies present in the patient’s natural environment do not support changes. One way in which these difficulties can be overcome is through the involvement of significant others in treatment programs. In cultures where the patient continues to stay in the home with families and is not under hospital care alone, this is a promising method.

Similarly, token economy systems have been used to modify several behaviors within a complex program [34]. Alderman and Knight [35] report the use of differential reinforcement in dealing with physically aggressive behavior in a patient following TBI. In a single case study with multiple baseline assessment, Hegel and Ferguson [36] report using differential reinforcement in a young man with TBI and aggressive behaviors as sequelae. In differential reinforcement, the undesirable behavior is placed on an extinction schedule, while the desired behavior is positively reinforced.

Becker and Vakil [37] describe the steps in a behavioral program adopted to treat two types of frontal lobe syndromes called disinhibition and adynamia. The steps outlined include establishing a therapeutic alliance, diagnostic evaluation, identification of target problems, implementation of the behavioral strategies including contingency contracting, reinforcement, relaxation training, social skills, and rehearsal and role plays. The final steps in this program include the generalization of skills to naturalistic settings. According to the authors, these patients with frontal lobe injury lacked sensitivity to feedback; the behavioral strategy of manipulating the environment provides this feedback. The exaggerated reinforcements that are provided help in building adaptive mechanisms.

1 Introduction

Behavior problems are more likely to occur in adults with intellectual disabilities (ID) compared to the general population (Hudson & Chan, 2002; Moss et al., 2000). Lowe et al. (2007) reported that approximately 16–35% of adults with ID exhibit problem behaviors.

A variety of applied behavior analysis (ABA) techniques have been used to manage behavior problems in adults in this population. As far as the use of aversive interventions is concerned, standards for practice in ABA severely restrict the use of aversive events such as physical punishment; these are only used when the problem behavior has not responded to gentler procedures, and are a threat to the safety of the individual or others (Association for Behavior Analysis International, 1993). The use of aversive strategies in the management of problem behaviors has been explored in several studies. For example, McAdam, Sherman, Sheldon, and Napolitano (2004) in a review of behavioral interventions of pica in persons with ID, found overcorrection and physical restraint to be the most frequently used. Matson and Keyes (1988) noted that aversive strategies were the most widely-used intervention for management of aggression and self-injurious behaviors (SIB). Reviews of behavioral treatments for chronic hair pulling (Friman, Finney, & Christophersen, 1984; Long & Miltenberger, 1998) also indicate widespread use of punishment contingencies to manage such behaviors in individuals with ID.

The use of aversive interventions generated widespread controversy in the last decades as the least-restrictive-alternative (LRA) guideline of providing treatment gained growing importance (Kazdin, 2001). This doctrine proposes that the procedure which is least intrusive, yet likely to be effective, should be used first. A similar perspective has been extended by Van Houten, Axelrod, Bailey, Favell, et al. (1988) in the Task Force on the Right to Effective Treatment appointed by the Association for Behavior Analysis. Consistent with the philosophy of least restrictive yet effective treatment, these authors maintained that it is unacceptable to expose an individual to a non-restrictive intervention if assessment results indicate that other procedures would be more effective. They further mentioned that “a slow-acting but non-restrictive procedure could be considered highly restrictive if prolonged treatment increases risk” (p. 113).

In any case, the LRA guideline received particular attention in the context of using aversive events, such as time-out, physical restraint, and electric shock, for individuals with ID who may not be able to consent to such treatment (Repp & Singh, 1990). Consequently, professional organizations, parent groups, and state legislatures have been increasingly proactive in establishing resolutions and laws regarding the use of intrusive interventions for behavior management in this population (Beare, Severson, & Brandt, 2004). The American Association of Intellectual and Developmental Disabilities (AAIDD, formerly AAMR), the National Association of Developmental Disabilities Council, and the Association for Retarded Citizens (ARC) have all passed policy statements supporting the development of non-aversive techniques that are both effective in reducing maladaptive behaviors and are respectful of the dignity and welfare of the individual (Bird, Hepburn, Rhodes, & Moniz, 1991).

Differential reinforcement (DR) is a non-aversive, reinforcement-based behavioral intervention found to be effective in suppressing maladaptive behaviors in individuals with ID. DR does not involve extended interruption of ongoing activities (e.g., time out), contingent removal of positive reinforcers (e.g., response cost), or presentation of aversive stimuli (e.g., punishment). These characteristics of DR make it among the least intrusive of all behavioral interventions and probably account for its widespread popularity (Cowdery, Iwata, & Pace, 1990).

2.6 Treatment efficacy calculations (PRB and PZD)

The efficacy of the treatments provided within individual studies are presented in Tables 1 and 2. Table 3 provides an overview of treatments, target behaviors and treatment efficacy calculations. Results are presented in the following order: overview of study characteristics; analysis of treatment efficacy; summary of research strength ratings and Z scores indicating EBP status. Table 3 shows that Vocal RIRD, involving verbal interruption and redirection to an alternative vocal response, was the most frequently implemented intervention. However, although an average reduction of 77.66% in stereotypy was identified from baseline phase to intervention across studies, a PZD analysis indicated that it was an ineffective intervention for suppressing stereotypy. Vocal RIRD was combined with a variety of other interventions such as verbal operant training, the provision of matched stimulation, differential reinforcement of incompatible behavior, and medication. In all cases, the target behavior was reduced by more than 50% from baseline levels. However, PZD statistics indicated that each of these interventions was either ineffective, or showed questionable effectiveness, in suppressing target behaviors.

Table 3. Applications and efficacy of treatment, or treatment packages, with response redirection or response interruption and redirection. PRB (percentage reduction from baseline), PZD (percentage zero data), NCR (noncontingent reinforcement), DRO (differential reinforcement of other behaviors), RIRD (response interruption and redirection), SSRI (selective serotonin reuptake inhibitor), NCS (noncontingent competing stimuli), DRA (differential reinforcement of other behavior).

Motor RIRD, involving verbal interruption and redirection to an alternative motor response, has been studied less commonly. In the current review, Motor RIRD alone was demonstrated to produce reductions of greater than 80% in vocal and motor stereotypy but was shown to be an ineffective method of suppressing such behaviors. Motor RIRD used in combination with differential reinforcement of alternative behavior and the provision of competing stimuli lead to a 96.8% reduction in pica. According to Wendt's (2009) guidelines for the interpretation of PZD scores this was a fairly effective method of suppressing the behavior (Hagopian et al., 2011). A variety of combined interventions incorporating response redirection were evaluated as treatments for motor and vocal stereotypy, self-injurious behavior, pica, and breath holding (see Table 3). Target behavior was redirected in a variety of ways including: redirection to a motor activity; redirection to preferred stimuli; response blocking and redirection to preferred stimuli; physical redirection; physical redirection in combination with noncontingent reinforcement and differential reinforcement of other behaviors, RIRD to academic tasks, and response interruption. All implementations led to at least a 40% reduction in target behaviors although some treatment packages were markedly superior to others (range of PRB: 43.48–100%). However, behavior suppression (PZD) was only reported for pica demonstrating response blocking and redirection to preferred stimuli as a highly effective treatment for this behavior.

As can be seen in Tables 1 and 2, the efficacy of both redirection and RIRD interventions were variable. However, the use of functional assessment or analysis did not appear to account for differences in treatment efficacy. Function-based interventions resulted in an average PRB of 77.10 and an average PZD score of 19.9, indicative of ineffective behavioral suppression. Correspondingly, non function-based interventions resulted in an average PRB of 74.84 and an average PZD score of 1 also indicative of ineffective behavioral suppression.

A comparison of the effectiveness of treatments incorporating a redirection component and treatments incorporating RIRD was conducted. Treatments incorporating a redirection component led to a mean reduction of 72.53% (range 43.48–100) from baseline levels of the target behaviors, and received a mean PZD score of 16.35 which indicated that such treatments are typically ineffective at suppressing challenging behaviors. Treatment incorporating a RIRD component was slightly more effective and led to a mean reduction of 77.25% (range 43.64–99.48) from baseline levels of the target behaviors, but led to a mean PZD score of 9.78 which indicated that it too was ineffective at suppressing target behaviors.

Increases in alternative appropriate behavior, including appropriate vocalizations, on-task behavior, leisure item engagement, and independent discards of inedible items, were more frequently reported (70%) in studies evaluating treatments with an RIRD component than in studies (25%) evaluating treatments with a redirection component.

Few data were presented on the maintenance or generalization of treatment effects. Two studies reported some maintenance of treatment effects (Ahearn et al., 2007; McEntee et al., 1996). One study reported that the treatment effects did not generalize to novel settings or instructors (Cassella et al., 2011). Two studies compared the efficacy of interventions in a clinical setting and in a natural setting with both finding that intervention in the natural environment was more effective in reducing challenging behavior (Hagopian et al., 2011; Turner et al., 1996).

1 Introduction

Problem behaviors such as aggressive, stereotyped, and self-injurious behavior are highly prevalent among persons with autism (e.g., Matson & LoVullo, 2008; Murphy, Healy, & Leader, 2009). The problem behaviors bring along major risks for the individual with autism and his/her family with regard to their physical, emotional, and social well-being, and can accordingly reduce their quality of life (e.g., Walsh, Mulder, & Tudor, 2013). In order to reduce problem behavior in persons with autism, several (cognitive-)behavioral interventions are used, such as differential reinforcement of other behavior (DRO), differential reinforcement of incompatible behavior (DRI), differential reinforcement of alternative behavior (DRA), antecedent control, antecedent exercise, noncontingent reinforcement, social stories, picture exchange communication system (PECS) interventions, and mindfulness-based interventions.

Many studies published in the domain of behavioral intervention research for reducing problem behavior among persons with autism are single-subject studies. In order to synthesize the results of these studies and to study which variables are moderating the effectiveness of the behavioral interventions, meta-level research is needed. Accordingly, Campbell (2003) conducted a quantitative synthesis of single-subject studies published between 1966 and 1998 on the efficacy of behavioral interventions for reducing problem behavior in persons with autism. In the meantime many more studies were published in this domain (cf. Matson & LoVullo, 2009), and an update of the meta-analysis of Campbell (2003) was warranted. The present article provides a double update of this meta-analysis: one from a methodological perspective and one from a temporal perspective.

First, we applied a methodological update. Campbell (2003) calculated three single-subject nonparametric statistics for estimating the effects of the behavioral treatments: the percentage of nonoverlapping data (PND; Scruggs, Mastropieri, & Casto, 1987), the percentage of zero data (PZD; Scotti, Evans, Meyer, & Walker, 1991), and the mean baseline reduction (MBLR; Kahng, Iwata, & Lewin, 2002). However, in the meantime new single-subject nonparametric statistics have been developed that avoid some of the drawbacks of the earlier developed statistics (e.g., see Heyvaert, Wendt, Van den Noortgate, & Onghena, in press; Parker & Brossart, 2003, and Parker, Vannest, & Davis, 2011, for overviews). Therefore we included the percentage of data points exceeding the median of baseline phase (PEM; Ma, 2006) and the percentage of all nonoverlapping data (PAND; Parker, Hagan-Burke, & Vannest, 2007) in our update. In contrast with some other recently developed nonparametric statistics, PEM and PAND have comparable ease of use to PND, PZD, and MBLR (cf. Parker et al., 2011): All five effect sizes can easily be calculated by practitioners.

PND is the most frequently used effect size index across single-subject syntheses in the field of disability research (cf. Maggin, O’Keeffe, & Johnson, 2011). In order to meet PND's main drawback (i.e., the deficient performance in the presence of data outliers in the baseline phase) Ma (2006) developed PEM as an alternative effect size for summarizing results of single-subject studies: Whereas PND takes into account the one most extreme value of the baseline phase, and refers to the percentage of data points in the treatment phase that exceeds this most extreme value, PEM takes into account the median value of the baseline phase. PEM leans very close to PND in its calculation and interpretation. PAND was also developed to meet the drawbacks of PND, but conceptually deviates a bit more from it. The main advantages of PAND over PND are: (1) PAND uses all data from the baseline and intervention phases, avoiding the criticism leveled at PND for overemphasis on one unreliable data point; and (2) PAND can be translated to Pearson's Phi and Phi2, and because Phi and Phi2 have known sampling distributions, p values are available, statistical power can be estimated, and confidence intervals can be included to indicate effect size reliability (Parker et al., 2007). Accordingly, for the present study we calculated PND, PZD, MBLR, PEM, and PAND for estimating the effects of behavioral interventions for reducing problem behavior in persons with autism. We compared these five nonparametric statistics with one another and examined to what extent they agreed in the analysis of the same data set. Answers to such questions are needed for scientist-practitioners to confidently use nonparametric statistics in the analysis of single-subject data (Parker & Brossart, 2003).

Second, for the temporal update we included single-subject studies published between 1999 and 2012 in our meta-analysis. Analogous to what Campbell (2003) did for the studies published between 1966 and 1998, we summarized single-subject studies published between 1999 and 2012: We studied the overall efficacy of behavioral interventions in reducing problem behavior in individuals with autism, examined whether some behavioral interventions were more effective than others, and investigated which variables, if any, moderated the overall efficacy of the behavioral interventions. Furthermore, we compared the two sets of studies and examined whether there were differences in the use of behavioral interventions and their effectiveness in reducing problem behavior in individuals with autism.

Accordingly, the research questions addressed in the current review were: (1) What is the overall efficacy of behavioral interventions in reducing problem behavior in individuals with autism; (2) Are some behavioral interventions more effective than others in reducing certain types of problem behavior in individuals with autism; (3) Do participant, treatment, or experimental variables influence the overall efficacy of behavioral interventions; (4) Are there any differences between the three older effect sizes (i.e., PND, PZD, and MBLR; Campbell, 2003) and the two more recently developed effect sizes (i.e., PEM and PAND) regarding treatment efficacy and moderating variables; and (5) Are there any differences between the single-subject studies published between 1966 and 1998 (Campbell, 2003) and the studies published between 1999 and 2012 regarding the use of behavioral interventions and their effectiveness in reducing problem behavior in individuals with autism?