Journal of Mental Health Research in Intellectual Disabilities , 2:169–187, 2009 Copyright © Taylor & Francis Group, LLC ISSN: 1931-5864 print / 1931-5872 online DOI: 10.1080/19315860902725875 169 UMID 1931-5864 1931-5872 Journal of Mental Health Research in Intellectual Disabilities, Vol. 2, No. 3, May 2009: pp. 1–31 Journal of Mental Health Research in Intellectual Disabilities Aggression and Tantrums in Children with Autism: A Review of Behavioral Treatments and Maintaining Variables Aggression and Tantrums in Children with Autism J. Matson JOHNNY MATSON Department of Psychology Louisiana State University Aggression and tantrums are common co-occurring problems with autism. Fortunately, positive developments in the treatment of these challenging and stigmatizing behaviors have been made recently with psychologically-based interventions. Evidence-based methods employ behavior modification, which is also often described as applied behavior analysis and has been at the fore- front of these developments. Conceptually, researchers in the field note that many of the factors maintaining these challenging behaviors are communication related. A treatment that produces functionally equivalent reinforcement yet is socially appropriate has been proposed as a major thrust of intervention. In this article research-based treatments are reviewed for aggression and tantrums, the particular behaviors that have been effectively treated are dis- cussed, and strengths and weaknesses of the intervention strategies are critiqued. A rationale for why these methods may prove to be an alternative to pharmacological interventions is presented. KEYWORDS autism, aggression, property destructions, applied behavior analysis, children, behavior modification Autism spectrum disorders (ASD) are a set of five neurodevelopmental con- ditions with many common features (Matson, 2007a; Matson & Boisjoli, 2007). By far the most frequently studied of the ASD is autism (Matson, Nebel-Schwalm, & Matson, 2007). Core features of the disorder include Address correspondence to Johnny Matson, Department of Psychology, Louisiana State University, 234 Audubon Hall, Baton Rouge, LA 70803. E-mail: johnmatson@aol.com 170 J. Matson deficits in communication and social skills as well as stereotyped and ritual- istic behaviors (MacDonald et al., 2007). In addition to these basic features, a cascade effect is seen for many other important skill sets such as general adaptive functioning (Matson, 2007b). Further complicating the picture are co-occurring problems such as comorbid psychopathology (La Malfa et al., 2007) and challenging behaviors such as property destruction, self-injury, and aggression (Machalicek, O’Reilly, Beretvas, Sigafoos, & Lancioni, 2007; Matson & Minshawi, 2007; Plant & Sanders, 2007). Aggression and related behaviors such as tantrums and property destruction are common in children with ASD (Baghdadli, Pascal, Grisi, & Aussilloux, 2003; Fox, Keller, Grede, & Bartosz, 2007). Similarly, subtypes of these problems such as bullying have been found to occur in almost half of the children with ASD (Montes & Halterman, 2007). Furthermore, early detection of ASD is particularly critical because these aforementioned prob- lem behaviors are often linked to other core deficits in social skills and communication (Dominick, Davis, Lainhart, Tager-Flusberg, & Folstein, 2007; Hoch & Symons, 2007; Ray & Schlottmann, 2007). Interventions for these particular problem behaviors for very young children are often bun- dled into more comprehensive treatment programs. When these treatment approaches are used, researchers have proposed that the best results occur before the child reaches 5 years of age (G. Green, 1996; MacDonald et al., 2007). Unfortunately, in practice, differential diagnosis of ASD can be com- plicated. Difficulties with differential diagnosis may explain why the average age of diagnosis is 5 years, with parents consulting 4.5 professionals on average over a 3-year period to arrive at the result (Siklos & Kerns, 2007). Thus, it should not come as a surprise that these externalizing problems tend to persist and become chronic (Murphy et al., 2005). Furthermore, aggression, property destruction, and behaviors such as tantrums that accompany them are highly stigmatizing and are marked impediments to learning (Machalicek et al., 2007). Because of the serious nature and long-term course of aggression and property destruction, the most powerful interventions in the researchers and clinicians list of evidence-based interventions are typically employed. More often than not these interventions have in the past been complex behavioral programs that often incorporate some components of a restrictive nature or pharmacotherapy (Matson & Minshawi, 2006). Pharmacotherapy has been researched the most extensively for aggres- sion and property destruction in children and adults with ASD (Martin, Koenig, Anderson, & Scahill, 2003). This approach has been used even with very young children (4–5 years of age). An often-cited rationale for pharmaco- therapy is that the drug is being administered to treat co-occurring disorders with ASD, such as hyperactivity, which leads to aggression and acting-out behavior (Aman, 2004). However, it has been cautioned that behavioral methods are highly effective with this group, and medications might be Aggression and Tantrums in Children with Autism 171 better utilized with adolescents and adults whose behaviors are more entrenched and resistive to behavioral interventions (Matson & Dempsey, 2008). Despite this rationale, there has not been a data-based review to date regarding behaviorally based treatment specific to aggression and tantrums for children who evince ASD. Given the expanding nature of interventions, particularly those that incorporate replacement behaviors based on a com- munications model, such a review would appear warranted. The goal of this article is to provide such a review. AGGRESSION AND TANTRUMS IN AUTISM Aggressive and tantrum behavior such as hitting, kicking, biting, punching, scratching, and throwing furniture are common among children with ASD (Hellings et al., 2005). Furthermore, although temperament is certainly a factor, setting events also play an important role in the expression of these behaviors as the form and intensity of aggression varies across time, setting, and changes in stimuli. For example, high-volume noise, changes in routine demands, and transitions from one task or environment to another are often precipitants of aggression. These problem behaviors are the most common reason for preschoolers to be referred for mental health services (Luby & Morgan, 1997; Renk, 2007). Furthermore, these behaviors tend to persist over time unless effectively treated (Keenan, Shaw, Walsh, Dellaquadri, & Giovanelli, 1997; Moffitt, 1990; Murphy et al., 2005). Children displaying problem behaviors also have a lower nonverbal IQ, poor expressive lan- guage, and marked deficits in social skills (Dominick et al., 2007). In addition, these behaviors often restrict the child’s activities and are a major impedi- ment to learning. For these and other reasons, treatment of aggression and other related behaviors such as property destruction must be one of if not the top priority for treatment for children with ASD. ETIOLOGY OF AGGRESSION The research on the etiology of aggression for children with ASD is largely nonexistent. Given the high rates of the problem in these children makes this issue of particular concern. Researchers and clinicians at present are largely left to extrapolate from very substantial literatures with the general childhood population. In this latter instance issues of temperament, self-esteem, parenting, other family factors (e.g., issues of alcohol abuse, antisocial behavior, etc.), and peer group characteristics are routinely cited (Barnow, Lucht, & Freyberger, 2004). The minimal research on etiology of aggression has focused on mentalis- tic abilities associated with abnormal regional cerebral activity (Anckarsater, 172 J. Matson 2006) and on functional assessment to establish reinforcers that maintain aggression (Dawson, Matson, & Cherry, 1998). The neurocognitive explanations are at this point largely correlational and inferential in nature. For example, violent offenders are said to have “autistic traits” (Soderstrom, Nilsson, Sjodin, Carlstedt, & Forsman, 2005; Soderstrom, Sjodin, Carlstedt, & Forsman, 2004), and the lack of empathy sometimes seen as characteristic of ASD may be related to aggression as well (Soderstrom, 2003; Soderstrom et al., 2005). Additionally, Anckarsater (2006) argues that ASD is an example of a social interaction disorder, and social deficiencies are related to higher rates and more egregious acts of aggression. Physiologically, this is explained as social, cognitive, and per- ception deficits linked to abnormalities in the amygdala, medial temporal lobes, hippocampus, and striatum (Aylward et al., 1999; Ryu et al., 1999). Similarly, neurochemical correlates such as serotonin activity have been associated with ASD and aggression (Anckarsater, 2006). However, these relationships, although showing some promise, are at this point highly speculative. Furthermore, even assuming that additional research can refine and prove such relationships, there is no clear link to any possible interven- tion to improve or inhibit these challenging behaviors in children with ASD. Additionally, frequently co-occurring disorders such as intellectual disability (ID; Matson, Dempsey, LoVullo, & Wilkins, 2007), attention-deficit/hyperactivity disorder (ADHD; Montes & Halterman, 2007), and language impairment (Dominick et al., 2007) may complicate this issue further. Although temperament and its related physiological and neurochemical correlates most likely play a role in aggression for those experiencing ASD, the only empirically established causes are learning based. The primary meth- odology to establish maintaining variables has been functional assessment (Matson & Minshawi, 2007). In an early study utilizing an adult population with profound ID and ASD, Dawson et al. (1998) established that aggression was maintained primarily for attention. Although no significant group differ- ences were found with regard to behavioral function (i.e., autism, Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS), and no ASD), the authors did report trends in the data. When examined separately, the autism group was less likely to engage in behaviors to gain attention when compared with the PDD-NOS group or no ASD group. Other factors that resulted in aggressive behavior with the participants in this study were access to tangible items such as food as well as nonsocial factors. The non- social factors consisted of withdrawing from social situations and the moti- vation associated with a lack of interest in the environment. Participants with PDD-NOS and autism were more likely to engage in behaviors for non- social reasons than participants diagnosed with ID and no ASD. These factors (i.e., social withdrawal) are consistent with symptoms of ASD (Rutter, 1978). Furthermore, communication deficits characteristic of ASD are related to higher rates of challenging behavior in this population (Chung, Jenner, Chamberlain, Aggression and Tantrums in Children with Autism 173 & Corbett, 1995; Schroeder, Schroeder, Smith, & Dalldorf, 1978; Sigafoos, 2000). Thus, the presence of ASD, and in many cases coexisting ID, may be setting events for aggression and tantrums. However, causes of specific problem behaviors appear to be environmental in origin and tend to respond to learning-based interventions (e.g., applied behavior analysis, behavior therapy). The literature here is massive. Hanley, Iwata, and McCord (2003) reported 277 studies using functional assessment. Most stud- ies are with ID populations, but given the high incidence of ID in ASD, extrapolating procedures to this latter group is not much of a stretch. The extrapolation of functional assessment methods to persons with ASD is bolstered by some other important aspects of the disorder. Commu- nication and social skills are two of the three core domains of ASD and have been linked to functional assessment from the outset (Campbell & Lutzker, 2005; Durand & Carr, 1992). Methods of identifying maintaining factors for the challenging behaviors of aggression and tantrums can be important not only in decreasing the interfering behaviors but also address- ing shortcomings in core ASD skill areas. Aggression and tantrums present a special set of problems in the applica- tion of functional assessment of children with ASD. The biggest potential issue involves the use of experimental functional analysis (EFA), particularly with older, bigger, stronger individuals. To adequately assess the challenging behav- ior with an EFA, conditions must be established that will produce the problem behavior. For obvious reasons regarding injury and continuation of the problem behavior once the assessment conditions have been terminated, an EFA should be the last-resort assessment method with such behaviors. Furthermore, EFA may not be advisable in some instances where the maintaining variables can be adequately identified with other methods. For aggression and tantrums that are serious in nature, we advise the use of checklists and antecedent-behavior- consequence, real-time data recording. The Questions About Behavior Function (QABF; Applegate, Matson, & Cherry, 1999; Paclawskyj, Matson, Rush, Smalls, & Vollmer, 2000) and the Functional Assessment for Multiple Causality (FACT; Matson et al., 2003) are brief Likert-type rating scales with established psycho- metrics that fit this description. Furthermore, other methodologies such as scatterplots can prove to be important adjunctive methods by helping narrow timeframes and better identify the conditions that precipitate challenging behaviors (Touchette, MacDonald, & Langer, 1985). PUBLISHED RESEARCH LITERATURE Interventions are broken down into two broad categories for the purpose of this review. These approaches include more traditional applied behavior analysis methods and functional assessment. A review of the data support- ing each of these general methods follows. 174 J. Matson Perhaps the most comprehensive review to date pertaining to ASD and challenging behaviors was published by Machalicek and colleagues (2007). This paper departs from our review in that a broad range of target behaviors was covered: aggression, tantrums, self-injurious behavior, stereotypies, classroom disruptions, pica, and noncompliance. Also, the studies were lim- ited to classroom environments. Nonetheless, a number of conclusions were instructive with respect to our review. First, considerable overlap in inter- ventions, specifically functional assessment, self-management, and differen- tial reinforcement were noted. However, given that our review also covered hospital and home settings, more intrusive interventions have been described in these environments. Applied behavior analysis methods that are generic are covered first. Although functional assessment could be included in this section, we describe it separately because of the dispropor- tionate amount of research these procedures have received relative to other behavioral methods. A selective review of package procedures is followed by a description of contingency contracting. We included studies we could find using the search engine Scopus and hand searches of the Journal of Applied Behavior Analysis, Research in Autism Spectrum Disorders, and Journal of Autism and Developmental Disorders. Keywords used for the search included autism, aggression, tantrums, and treatment. It should be noted that a large number of studies included in this review involved behaviors in addition to aggression and/or tantrums. The rationale for including studies that address behaviors in addition to aggression and tan- trums was due to the high likelihood of individuals engaging in multiple topographies of challenging behavior at a given time (Lowe et al., 2007), thereby, increasing the generalizability of this review. We use the term “selective” in the event that an article or articles were missed. Data from these studies are presented in Table 1. Package Procedures Foxx and Garito (2007) describe a package of behavioral treatments for the severe aggression, self-injury, tantrums, induced vomiting, and inappropriate toileting of a 12-year-old boy with autism. Treatment occurred across the home, community settings, and a self-contained classroom of a public school. Intervention procedures included a high density of social reinforce- ment for appropriate behaviors, tokens, choice making, and punishment procedures including contingent exercise and overcorrection when chal- lenging behaviors occurred. A functional assessment was not done, but the authors report reducing these problems to near zero levels, which were maintained at a 2-year follow-up. Paisey, Fox, Curran, Hooper, and Whitney (1991) confirmed their treat- ment of an 11-year-old with autism for severe aggression and tantrums to the home. Parental attempts to implement differential reinforcement of other 175 TABLE 1 Behavioral Treatments of Aggression and Tantrums of Children with Autism Authors and year Persons treated Target behaviors Intervention Setting ABA Hagopian, Bruzek, Bowman, & Jennett (2007) 2 children with autism, 6 and 12 years old Pinching, scratching, kicking, hitting, throwing objects, screaming Functional analysis, noncontingent reinforcement, task demands, interruption Social reinforcement for appropriate behaviors, tokens, choice making, contingent exercise, overcorrection University hospital Foxx & Garito (2007) Boy with autism, 12 years old Severe aggression, tantrums, self-injury, induced vomiting, inappropriate toileting Home, community setting, classroom Thompson, Fisher, Piazza, & Kuhn (1998) Boy with pervasive developmental disorder and severe ID, 7 years old Hitting, kicking, pinching, scratching, grinding chin against others Functional analysis, communication training, blocking, physical guidance, redirection University hospital Functional assessment Wetzel, Baker, Roney, & Martin (1966) Boy with autism, 6 years old Tantrums Functional assessment and communication training Outpatient clinic, home Kern, Carberry, & Haidara (1997) Girl with autism and severe ID, 15 years old Aggression and self-injury Mand training, extinction, and increasing delays in reinforcement Residential facility Campbell & Lutzker (2005) Child with autism, 8 years old Severe tantrums and property destruction Functional communication training and activity planning Home Sigafoos & Meikle (1996) 2 boys with autism and moderate to severe ID, 8 years old Aggression, self-injury, property destruction, disruption, stereotypies Functional analysis and functional communication training School Vollmer, Borrero, Lalli, & Daniel (1999) 2 boys with autism, both 9 years old Severe aggression Reinforcing appropriate mands Inpatient hospital Braithwaite & Richdale (2000) Boy with autism Aggression and self-injury Communication training and extinction School Hagopian, Wilson, & Wilder (2001) Boy with autism, 6 years old Aggression, disruptive behavior, self-injury Functional analysis, noncontingent reinforcement Inpatient hospital Mueller, Wilczynski, Moore, Fusilier, & Trahant (2001) Boy with autism, 8 years old Aggression Presentation of alternative high- preference items School Note: ABA = Applied Behavior Analysis; ID = Intellectual Disability. 176 J. Matson behavior in conjunction with publicly posting good behavior rules, physical management training, and redirection to relaxation did not prove to be effective. Control was established only when the child was reinforced for compliance with task demands in conjunction with extinction and imple- mented in the home by professional staff. Thus, at least in this instance, consequences for maladaptive behavior considered unpleasant by the child and implemented by professionals was required. Thompson, Fisher, Piazza, and Kuhn (1998) described the treatment of aggression in the forms of hitting, kicking, pinching, scratching, and firmly pressing his chin and grinding against others, often resulting in physical injury to the caregiver. The boy (Ernie), who was 7 years of age, had severe ID and was diagnosed with pervasive developmental disorder. An EFA was used to establish attention as a maintaining variable for his aggression. Communication training consisted of a picture card of Ernie hugging a ther- apist that he could hand to a care provider to get attention. Aggression resulted in Ernie being ignored, and his communication card was not made available. However, although some aspects of aggression were effectively treated using function-based interventions, more robust contingencies were required for chin grinding. Ernie was provided with a rigid piece of plastic tubing as an alternative to a person’s arm for chin grinding, and blocking, physical guidance, and redirection consisting of the therapist putting his hands on Ernie’s shoulders and guiding him away from attempts to chin grind on others. This combination of methods was effective in reducing chin grinding. The multifactorial nature of behavioral antecedents is demonstrated in this study. Furthermore, this study demonstrated the need for mild punishers in the forms of extinction, manual guidance, and redirection to effectively reduce problem behaviors. Hagopian, Bruzek, Bowman, and Jennett (2007) describe the treatment of tantrums, Self-injurious behavior (SIB), and aggression in two children (ages 6 and 12 years) who evinced ID and autism. Among the aggressive acts displayed in the inpatient hospital setting where they were treated were pinching, scratching, kicking, hitting, and tantruming in the form of throw- ing objects and screaming. A traditional EFA, as described by Iwata and col- leagues (1994), was employed with both of these children. The children were placed in each of the following conditions: demand, social attention, tangible, and play. However, low rates of behavior were found even in con- ditions hypothesized to be maintaining the behaviors. The authors termi- nated the more traditional EFA and subsequently developed conditions where the child was instructed to “do” or “don’t do” an action. The authors found that when the children were interrupted during preferred activities by these instructions, the target behaviors were displayed at higher rates. This intervention phase involved progressively longer periods of time where ongoing activities were interrupted by “do” requests. Obviously, the chil- dren could not successfully learn communication and other skills without Aggression and Tantrums in Children with Autism 177 active intervention. Thus, gradually exposing them to a compliance regimen regarding instructions when preferred activities were not available was an attempt to accomplish this goal, resulting in decreased problem behaviors. Although using the more traditional EFA results was inconclusive, the authors identified establishing operations that were likely to result in either high or low levels of problem behavior. Complex behaviors that are more severe and that have persisted for years may be particularly hard to pinpoint with respect to relevant maintaining variables. As this study points out, using more traditional methods of EFA may not result in information leading to effective treatment. Thus, the examiner needs to adapt the intervention through consideration of other variables that influence the occurrence of behavior. Functional Assessment We have chosen the term “functional assessment” because it provides an over- arching definition that incorporates EFA and scaling methods (Herzinger & Campbell, 2007). Typically, the functional assessment is followed by treatment strategies that are skill building in nature and that are designed to replace the function of the challenging behavior or behaviors. It is important to stress however, that challenging behaviors may have multiple functions. When this latter issue is true the intervention picture is complicated further. Example studies are noted in Table 1 and provide exemplars of incorporat- ing functional assessment into treatment studies for children with ASD and aggression and/or temper tantrums. One of the first studies to employ an assessment of behavior function that led to intervention was by Wetzel, Baker, Roney, and Martin (1966). The addition of extinction or punishment procedures to consequate the challenging behaviors have also frequently been added. In addition to addressing the determined function of a behavior and tailoring the inter- vention to the function, the authors also claim this to be the first study applying operant techniques with this population through outpatient con- tact only (Wetzel et al., 1966). The participant was a 6-year-old boy (David) with autism. He engaged in daily tantrums consisting of throwing items at others, pulling hair, pinching, face slapping (his own), and screaming. The experimenters and the child’s mother agreed on increasing approach behaviors and decreasing maladaptive behaviors as the targets for treatment. During baseline sessions (with limited items/toys in the room), David became aggressive toward the experimenters. At that time it was hypothesized that his maladaptive behaviors were attention seeking in nature. The experimenters explored this hypothesis in subsequent ses- sions, removing all attention during the maladaptive behavior. Attention was reinstated when appropriate behavior was evinced. The authors stated that the removal of attention served as a potent punisher for the 178 J. Matson maladaptive behavior. Additionally, the boy was provided with reinforce- ment in the form of verbal praise and interactive play when he appropri- ately approached the experimenters. Decreases in maladaptive behavior and increases in approach behavior were reported. In addition to the aforementioned treatment, the experimenters used attention as a rein- forcer to increase David’s verbal repertoire. Preferred interactions (e.g., tickling, bouncing) were made contingent on appropriate verbalizations. Furthermore, the intervention techniques were taught to his mother and implemented at home, school, and in the community. During a time when behavior therapy was just surfacing, and the more traditional psychotherapy was often used in treatment, these experimenters were successful in iden- tifying the variables maintaining a behavior, employing reinforcers and pun- ishers, and teaching the child more appropriate ways to communicate his desires, all using behavioral technology. More recently Kern, Carberry, and Haidara (1997) used mand training and extinction following an EFA for a 15-year-old girl with autism displaying aggression and self-injury. Similarly, Vollmer and colleagues (1999) trained mands to obtain reinforcement in lieu of aggression for two 9-year-old boys with autism. The fact that this procedure was in effect replicated in a related study suggests that this treatment/assessment package be given seri- ous consideration in the treatment of aggression and tantrums of children with autism. Braithwaite and Richdale (2000) also conducted a single-participant case study of a boy with autism in a school setting for aggression and self- injury. Escape from difficult tasks and access to preferred objects were the identified maintaining variables. Treatment involved teaching the child an alternative request and placing the challenging behaviors on extinction. The authors concluded that aggression and self-injury could be replaced with functionally equivalent communication, which was a centerpiece of all three of the treatments just described. Campbell and Lutzker (2005) used a single-case research design and EFA to establish the maintaining factors for severe tantrums and property destruction of an 8-year-old child with autism. Functional communication training and activity planning were used in the home setting as replacement behaviors for the maladaptive responses. This procedure departs from the studies teaching mands in that those skills could be functional or not, relative to the challenging behavior. In the Campbell and Lutzker study, communi- cation training was more “target behavior specific.” Along these same lines, Sigafoos and Meikle (1996) treated the aggression, self-injury, and disruptions of two boys with autism. Using an EFA and a multiple baseline design to assess treatment effects, they focused intervention on function- ally equivalent alternative methods of communication. Self-injury and disrup- tions were directly treated by teaching mands to obtain attention and to request preferred objects. Treating these target behaviors also resulted in a decrease in rates of aggression. This latter finding suggests the interconnection Aggression and Tantrums in Children with Autism 179 of the target behaviors and the presence of similar maintaining variables across all topographies of challenging behaviors. This phenomenon is not uncommon. Inconclusive results following a functional assessment are a problem some experimenters and clinicians may face. A possible reason for not being able to accurately identify the behavioral function may be subcatego- ries of function within larger behavioral function categories (Taylor & Carr, 1992). For example, a person engaging in a behavior to escape a task or to escape a social interaction are both escape but would require treatments that differ contextually (Hagopian, Wilson, & Wilder, 2001). Hagopian et al. (2001) treated a 6-year-old boy who was an inpatient for severe challenging behav- iors, which included aggression (i.e., hitting, kicking, scratching, head butting, and throwing objects at others), disruptive behavior (i.e., screaming, throwing objects, tearing objects, swiping objects off tables), and self-injury (i.e., head hitting with hand and head banging). An EFA was conducted with the following conditions: social attention, demand, play, and alone. For the ini- tial analysis the tangible condition was not employed due to parent report that they did not provide the boy with items contingent on the maladaptive behavior; therefore, his behavior was not maintained by access to tangibles. The results of the analysis were inconclusive. Using the information pro- vided from the initial analysis, the experimenters decided to conduct an additional modified analysis. In addition to the escape from demand condi- tion and social attention conditions, a tangible condition and escape from social attention condition were employed. The results of the modified EFA indicated that the boy’s maladaptive behavior was maintained by access to tangibles and escape from social attention. A treatment plan was imple- mented using functional communication training to teach the boy to appro- priately request to play with a toy alone. Additionally, noncontingent reinforcement was provided on a fixed interval schedule consisting of access to other toys and books. This study highlighted the importance of using multiple methods of assessment. During the interview it was deter- mined that the boy’s behavior was not maintained by access to items, yet according to the EFA access to items was a function of his maladaptive behavior. Mueller, Wilczynski, Moore, Fusilier, and Trahant (2001) treated an 8-year- old boy with autism who was enrolled in an extended summer school program. An EFA revealed that restriction of tangible items resulted in aggression. Thus, the recommended treatment was the presentation of alter- native high-preference items so that he would not become aggressive when often highly preferred items were restricted. However, the authors did not actually test their intervention, and we worry that such an approach might actually end up reinforcing and thus strengthening the aggression. At the very least, a treatment of this sort would need to be implemented very precisely and carefully. 180 J. Matson Contingency Contracting Mruzek, Cohen, and Smith (2007) describe contingency contracting to pro- mote rule following for two boys, one with autism (age 10) and one with ADHD and possible Asperger’s syndrome (age 9). Target behaviors included antisocial comments, tantrums, and physical aggression. Contracts, which provided rewards for proper conduct, were initiated and revised as improvement was noted. A self-monitoring aspect of the program was also included. Although such methods may need to be limited to older children and those with normal intelligence or mild ID, such methods are cost effec- tive and involve choice making by the child. A great deal more research with these methods seems to be in order. CONCLUSION Functional assessment is a methodology receiving the greatest amount of attention among behavioral treatments in the literature on aggression and tantrum behaviors of children with ASD today. However, it must be cau- tioned that functional assessment is not a treatment in and of itself. Rather, this theoretical model provides a conceptual pathway to the selection of parsimonious treatments that dovetail into communication targets. These communication targets are then trained with conventional applied behav- ioral analysis methodologies. The aim is to change the focus from behavior suppression to skill acquisition and maintenance through teaching adaptive behaviors that replace the function of the maladaptive responses. Although conventional wisdom is that this approach can decrease or eliminate the use of aversive events for the child with ASD, no direct comparisons have been made between these two behavioral methodologies. It is cautioned that claims of effectiveness should be framed within the confines of the available data. Unfortunately, this does not appear to be the case at present for many clinicians and researchers. These optimistic scenarios may yet prove to be correct and hypotheses await. On a positive note, some data is emerging that suggest functional assessment may be variable as a means of “enhanc- ing” overall treatment effectiveness. The functional assessment literature has been developed for persons with ID, a group that overlaps considerably with ASD. Thus, results from the functional assessment literature should prove instructive for ASD researchers and may provide in many cases a way of increasing the power of behavioral technologies. In our view, a disturbing trend in the literature on ASD is the use of antipsychotic drugs to treat even very young children (Matson & Dempsey, 2008). Some studies have included children as young as 4 years of age. Recently, investigators have found that Risperidone and Haldol were no more effective than placebo in treating aggression in persons with developmental Aggression and Tantrums in Children with Autism 181 disabilities (Tyrer et al., 2008), and the discontinuation of such practices has been suggested (Matson & Wilkins, 2008). Data such as these presented here provide an even more compelling argument for behaviorally based interven- tions with emphasis on functional assessment, communication, reinforcement, and where necessary, extinction, blocking, and time-out as alternatives. Didden, Korzilius, van Oorsouw, and Sturmey (2006) conducted a meta-analytic study based on 80 articles using behavioral treatments for challenging behaviors of people with mild ID. They found that in studies where a functional assessment was used, larger effect sizes were observed. More specifically, in studies that employed EFA, a significantly larger effect size was noted than when behavioral methods were used following descrip- tive functional assessments. Another study, by Machalicek and colleagues (2007), looked at school interventions for children with challenging behav- iors and ASD. The authors found that half of the studies included in the review did not employ some type of functional assessment prior to the intervention. Furthermore, there was no difference with regard to interven- tion effectiveness if a functional assessment had or had not been conducted. However, the authors did note that some of the strategies implemented without conducting a functional assessment might have actually addressed the function of the challenging behavior. Furthermore, using daily sched- ules for transitions or video modeling for social skills training may also assist with decreasing challenging behaviors without the completion of a functional assessment. These interventions, for example, address the core features of ASD such as social skills deficits and insistence on sameness, which may contribute to challenging behaviors (Machalicek et al., 2007). Although an abundance of studies exist that report the effective identifica- tion of maintaining variables for challenging behavior, what are urgently needed are comparative studies that look at efficacy of treatments with and without functional assessment. Only one large group-controlled study of this type has been done, and it was with adults with ID (Matson, Bamburg, Cherry, & Paclawskyj, 1999). The complexity of the various disorders, types of functional assessment, variations in follow-up interventions such as skill- building methods (e.g., communication and social skills programs), and aversive consequences (e.g., extinction and time-out) require a great deal of additional study. The research presented in our review is representative of the available literature. As a group of studies they demonstrate that behavioral proce- dures are effective, although we were a bit surprised that so few studies had been published with this particular population focusing on aggression and tantrums. However, and we think unfortunately, there is considerable “white noise” in the ASD field, with bogus interventions frequently being described as foolproof alternatives to applied behavior analysis. V. A. Green and colleagues (2006) provide evidence of our concerns. They note that the average number of treatments parents undertake for children with ID was 182 J. Matson high, and the number was even higher for persons with severe autism. Fur- thermore, frequency with which treatments were used had little bearing on the level of data available to support the treatment. Over half of the children were on medication for their developmental disability. Given the emphasis on antip- sychotics to treat aggression and tantrums, we suspect that children with these problems have even higher rates of drug use than other children with ASD despite data suggesting that antipsychotics, the most highly prescribed drug class for these particular challenging behaviors in the developmentally disabled population, may not be effective for these purposes (Matson & Wilkins, 2008). The bulk of the treatment studies reviewed did use some sort of func- tional communication training. However, we point out that “functional” might not be all that functional in some instances. Although brief treatments and little or no follow-up can demonstrate some functional control, programmatic factors may also weigh heavily on long-term prognosis. For example, escape is a common function of challenging behaviors. Signaling for breaks can be trained to replace the maladaptive behavior that serves an escape function. However, if the child just dislikes the task and is constantly signaling break, such communication training will serve little relevance beyond letting care- givers know they do not want to do the task. Thus, trying to create more enjoyable tasks may be the most critical aspect of intervention. Similarly, if a child is communicating attention at a rate higher than it can be reasonably provided, this communication response may also be impractical to some extent. Although