Journal of Developmental and Physical Disabilities, Vol. 14, No.

3, September 2002 ( C 2002)

Functional Analysis of Self-Injury With and Without Protective Equipment1

Duy D. Le2 and Richard G. Smith2,3

Outcomes of functional analyses of self-injury during which participants wore protective equipment (PE) were compared to those during which PE was not worn. Results of analyses without PE suggested that self-injurious behavior (SIB) was maintained by negative reinforcement for 2 participants and nonsocial stimulation for a 3rd participant. However, SIB was eliminated either immediately or eventually for all 3 participants when PE was worn during their functional analyses. Thus, outcomes of functional analyses with PE did not match those without PE, and the functional properties of SIB could not be identied from functional analyses with PE. Implications for the use of PE during functional analyses are discussed.
KEY WORDS: functional analysis; protective equipment; self-injury.

Functional analysis procedures developed by Iwata et al. (1982/1994a) have been repeatedly demonstrated to be effective at identifying controlling variables for aberrant behavior (e.g., Hanley et al., 1997; Iwata et al., 1982/1994a; 1994b,c; Kahng et al., 1997; Smith et al., 1993; Vollmer et al., 1993). However, a potential drawback of this procedure is that individuals with self-injurious behavior (SIB) are exposed to risk because SIB is evoked during the assessment. Many parents, guardians, and advocates express concerns about participants safety during functional analyses and request that precautions be taken to reduce the risk of injury. There are several ways in which functional
1 This

research was completed in partial fulllment of the requirements for the Master of Science degree for Duy Le, who is now at CARITAS Peace Center. 2 Department of Behavior Analysis, University of North Texas, Denton, Texas. 3 To whom correspondence should be addressed at Department of Behavior Analysis, University of North Texas, P.O. Box 310919, Denton, Texas 76203; e-mail: rsmith@scs.unt.edu. 277
1056-263X/02/0900-0277/0
C

2002 Plenum Publishing Corporation

278

Le and Smith

analysis procedures can be modied to reduce the risk of injury. First, session termination criteria (e.g., maximum number of self-injurious responses, occurrence of tissue damage, etc.) can be established by an attending physician prior to the assessment. Second, the assessment can be shortened by using single-session conditions (Derby et al., 1992; Northup et al., 1991) or analyzing within-session patterns of responding (Vollmer et al., 1993). Assessment of appropriate behavior has also been used to make inferences about the variables that might maintain dangerous problem behavior (Derby et al., 1992). Although specically contraindicated by Iwata et al. (1982/1994a), protective equipment (PE) can be used during functional analysis procedures to reduce the risk of injury. Several studies have reported the use of PE to treat self-injury (e.g., Dorsey et al., 1982; Iwata et al., 1994b; Mace and Knight, 1986; Mazaleski et al., 1994; Silverman et al., 1984). For example, Dorsey et al. found that noncontingent application of PE (foam-padded gloves and football helmet) suppressed SIB for all 3 participants in their study. However, the controlling variables for the participants SIB were not known because no pretreatment functional analysis was conducted. Mazaleski et al. conducted a functional analysis of the self-injurious hand mouthing and tongue pulling of 2 participants and found that their behaviors were maintained by nonsocial variables. Subsequently, oven mittens were placed over the participants hands either contingent on SIB or continuously throughout sessions. Both procedures produced signicant reductions in SIB. Mace and Knight examined the effects of PE on the rate of pica. They found that the lowest rates of pica were observed when their participant did not wear a helmet, moderate rates of pica were observed when the participant wore the helmet without a face shield, and the highest rates of pica were observed when the participant wore a helmet with a face shield. Iwata et al. (1994b) reported that providing helmets to 2 of 3 participants in their study did not affect rates of responding as long as the consequences identied to maintain SIB (escape or attention) were intact. Given inconsistent results of previous studies on the effects of PE on the occurrence of SIB, it is difcult to predict what effects PE would have on the outcomes of functional analysis procedures. If the outcomes of functional analyses are not inuenced by PE, it would be preferable to provide participants with such protection. If, on the other hand, PE can signicantly alter results of functional analyses, empirical support for proscriptions against the use of PE would be provided. However, no available data have directly evaluated the effects of PE on outcomes of functional analyses. The purpose of this study was to examine the effects of PE on functional analysis outcomes by comparing results of functional analyses (Iwata et al., 1982/1994a) during which PE was worn to those during which PE was not worn.

Functional Analysis With Protective Equipment

279

METHOD Participants and Settings Three adult males participated in the study. At the time of the study, Tom, Mike, and Fred lived in state residential facilities for people with developmental disabilities. Tom was 37 years old and was diagnosed with profound mental retardation. Tom was referred to a day program for assessment and treatment of self-injury because of his chronic and severe face slapping. He was nonambulatory and required assistance to propel his wheelchair during transportation. Tom was nonverbal. Available records indicate that mitt restraints had occasionally been used with Tom as an emergency procedure for aberrant behavior. Mike was 40 years old and was diagnosed with profound mental retardation. Mike was referred to a day program for assessment and treatment of self-injury because his ngernail biting resulted in tissue damage (i.e., the ngertips were chronically raw and red). He was nonambulatory and required assistance to propel his wheelchair during transportation. His expressive language consisted of several short sentences (e.g., I want to eat, Im cold, etc.) and shaking and nodding his head. He could respond to a few simple instructions. Available records did not indicate that rubber gloves had been used with Mike as an intervention for SIB. Fred was 35 years old and was diagnosed with profound mental retardation. Fred was referred for assessment and treatment of head banging, which resulted in lacerations that required sutures on several occasions. He walked with an extremely unsteady gait and often used a wheelchair for transportation. Fred could propel his wheelchair with his feet but rarely did so. Fred was nonverbal. Available records did not indicate that helmets had been used with Fred as a behavioral intervention for SIB. Fred, however, routinely wore his helmet, which was originally recommended by his physical therapist because of his unsteady gait. Sessions for Tom and Mike were conducted at a day program for assessment and treatment of self-injury located on the premises of their residence. Chairs, a table, a couch, and appropriate materials for each condition were present in the room. Freds sessions were conducted in a bedroom at his residence, where two beds, a table, two dressers, and appropriate materials for each condition were present. Apparatus Mittens (oven mittens), nonlatex gloves (medical gloves), and a helmet were selected as PE for Tom, Mike, and Fred, respectively, because they provided protection from, but did not prevent occurrences of, the

280

Le and Smith

topographies of SIB emitted by each participant. Freds helmet did not cover his ears, and had a clear face shield (with openings for his mouth and eyes), which was positioned approximately 2 in. away from his face. Three straps prevented Fred from removing the helmet. Target Behaviors, Observation Procedures, and Interobserver Agreement Tom engaged in face slapping, dened as audible contact of the palm (or that part of the mittens covering the palm during Assessment with PE) with any part of the face. Mike engaged in ngernail biting, dened as contact of the teeth with ngernail/ngernails (or that part of the gloves covering the ngernails during Assessment with PE). Fred engaged in head banging, dened as audible contact of the back of the head (or helmet during Assessment with PE) with the top edge of the back of his wheelchair or stationary chair. Each occurrence of the target behaviors was scored by trained observers using handheld computers (Apple Newton MessagePad Models 100 and 130) containing data-collection software (Behavior Observer System software). Data were calculated as rate (responses per min [rpm]) of SIB by dividing the recorded number of SIBs by the number of min in a session. A second observer simultaneously but independently scored data during 20.5, 33.3, and 33.3% of sessions for Tom, Mike, and Fred, respectively. Interobserver agreement scores were calculated by dividing the session length into 10-s intervals, dividing the smaller number by the larger number of responses scored within each interval, averaging the results across the session, and multiplying by 100. Mean interobserver agreement scores for SIB were 99.0% (range = 93.8100%) for Tom, 96.2% (range = 85.7100%) for Mike, and 99.8% (range = 95.19100%) for Fred. Experimental Design Functional analysis conditions (Alone, Attention, Play, and Demand for all 3 participants, and a fth condition, Escape from Wheelchair, for Fred) were conducted using a multielement format. All 3 participants were initially exposed to the multielement assessment without PE (hereafter referred to as No PE) until differentiation was observed in the data. For Mike, SIB occurred in the Alone condition, and the data were undifferentiated (i.e., SIB occurred across several other conditions) during the initial multielement assessment; therefore, an Extended Alone condition was conducted to rule out the possibility that SIB occurring in the Alone condition was a

Functional Analysis With Protective Equipment

281

function of lack of discrimination among conditions (a potential problem of the multielement design; Vollmer et al., 1995). Following the No PE assessment, PE was placed on the participants and procedures identical to the No PE assessment (i.e., same contingencies, therapists, number of sessions per condition, and sequence of conditions) were conducted (hereafter referred to as Assessment with PE). PE was then removed, and a second No PE assessment was conducted. Thus, the effects of PE were assessed using an A-B-A reversal design for Tom and Fred and an A1-B1-A2-B2-B1 reversal design for Mike. Procedures All sessions were 15 min in length except for the Extended Alone sessions with Mike (45 min). One to ve sessions were conducted 5 days per week. Procedures similar to those described by Iwata et al. (1982/1994a) were used to test whether SIB was sensitive to certain consequences as maintaining variables. Below is a brief description of the experimental conditions. Alone The participant was seated in a chair with no leisure materials, and SIB produced no programmed consequences. This condition tested whether SIB persisted in the absence of a programmed social contingency. If so, then it was inferred that nonsocial variables were (at least partially) responsible for the production and/or maintenance of SIB. Attention The participant was seated in a chair with leisure materials available. Self-injury resulted in attention from the therapist in the form of reprimands and/or concern statements. This condition tested whether SIB was sensitive to attention from the therapist as a reinforcing consequence. Play The participant was seated in a chair with leisure materials available. A therapist delivered neutral or positive statements on a xed-time (FT) 30-s schedule. No demands were presented to the participant, and SIB produced no programmed consequences. This condition served as the control against which the other conditions were compared.

282

Le and Smith

Demand The participant was seated in a chair without leisure materials. Approximately every 30 s, a therapist presented a task demand (e.g., folding a towel, wiping the table), using a three-prompt (verbal, modeling, and physical guidance) sequence. Compliance prior to physical guidance resulted in social praise from the therapist and a break until the next scheduled trial. If physical guidance was necessary to complete the task, no praise was provided, and the participant received a break until the next scheduled trial. Any SIB occurring during the three-prompt sequence resulted in the termination of the task until the next interval. Any SIB occurring at the end of a 30-s interval delayed the presentation of a demand for 5 s. This condition tested whether SIB was sensitive to escape from demands as a reinforcing consequence. Escape From Wheelchair (for Fred) Casual observation and anecdotal reports indicated that Fred engaged in head banging while sitting in the wheelchair but rarely did so when he was not in the wheelchair. Therefore, a fth condition was added to Freds assessment. Fred was seated in the wheelchair without leisure materials. A therapist approached him and delivered neutral statements and physical interaction on a FT 30-s schedule. Head banging resulted in removal from the wheelchair into a stationary chair for 1 min. The stationary chair was modied so that the response effort (i.e., the distance Freds head had to travel to bang his head on the top edge of the back of either chair) involved in head banging while sitting in either the wheelchair or stationary chair was equalized. Session time was suspended whereas Fred was out of the wheelchair. Sessions were terminated if Fred either sat in the wheelchair for a total of 15 min or escaped from the wheelchair 30 times. Sessions during which head banging occurred lasted up to 50 min in actual time. This condition tested whether head banging was sensitive to escape from the wheelchair as a reinforcing consequence.

RESULTS AND DISCUSSION Figure 1 shows the results of Toms assessments. During the rst No PE assessment, face slapping occurred almost exclusively in the Demand condition, suggesting that it was maintained by negative reinforcement (i.e., face

Functional Analysis With Protective Equipment

283

Fig. 1. Responses per minute of face slapping across Toms functional analysis conditions with and without PE.

slapping was evoked by the presentation of task demands and maintained by their termination). Face slapping did not occur in any condition when mittens were placed on Toms hands. When the mittens were removed, face slapping again occurred almost exclusively in the Demand condition. Results of the No PE assessment suggested that Toms face slapping was escape-maintained; however, no conclusion could be drawn from those obtained during the Assessment with PE. Thus, outcomes of assessment with mittens did not match those of the No PE assessment and did not provide independent evidence about the functional properties of Toms SIB. No hypothesis about variable(s) associated with SIB could be formed on the basis of results of the Assessment with PE alone. Figure 2 shows the results of Mikes assessments. Fingernail biting occurred across all four conditions of the No PE assessment, although it was generally higher during the Alone condition, suggesting that it was automatically maintained or multiply controlled. Fingernail biting persisted for seven 45-min sessions in the No PE Extended Alone condition (rates of ngernail biting were at least 1.90 rpm during six sessions, and the condition mean rate was 3.05 rpm), suggesting that it was maintained, at least in part, by nonsocial variable(s). Mikes Assessment with PE was also undifferentiated,

284

Le and Smith

Fig. 2. Responses per minute of ngernail biting across Mikes functional analysis conditions with and without PE.

although ngernail biting occurred at lower rates as compared to those from the No PE assessment. Fingernail biting never exceeded 1.0 rpm during the Extended Alone with PE condition (the condition mean rate was 0.24 rpm), and decreased to zero or near-zero rates during the last four sessions of this condition. Finally, rates of ngernail biting comparable to those from the rst No PE Extended Alone were observed during the second No PE Extended Alone condition (the condition mean rate was 2.71 rpm). Although the results of the multielement No PE assessment and Assessment with PE were similar, neither of these assessments provided denitive information about the functional properties of ngernail biting. Results of the Extended Alone with PE condition did not match those from the No PE Extended Alone condition and did not provide independent evidence about variable(s) associated with Mikes ngernail biting. Results from the No PE Extended Alone condition showed that Mikes ngernail biting persisted in the absence of social contingencies, suggesting that it was maintained, at least in part, by sources that are not socially mediated; however, ngernail biting was eventually eliminated during the Extended Alone with PE condition. Thus, no conclusions about the functional properties of ngernail biting could be drawn from the Extended Alone with PE condition. In summary,

Functional Analysis With Protective Equipment

285

Fig. 3. Responses per minute of head banging across Freds functional analysis conditions with and without PE.

providing PE altered assessment outcomes for Mike, making interpretation of data difcult. Figure 3 shows the results of Freds assessment. Head banging occurred almost exclusively during the Escape from Wheelchair condition of the No PE assessment, suggesting that it was maintained by negative reinforcement in the form of escape from his wheelchair. Head banging ceased to occur in all conditions when he wore the helmet during Assessment with PE. During the replication of the No PE assessment, head banging again occurred exclusively in the Escape from Wheelchair condition. Outcomes obtained from No PE assessment were not replicated during the Assessment with PE. Results of the No PE assessment suggested that Freds head banging was escape-maintained, wheareas no conclusion could be drawn from results of the Assessment with PE. As with the 2 other participants, Freds results indicate that providing PE during a functional analysis can alter its outcomes and interfere with the interpretation of data. GENERAL DISCUSSION The present ndings suggest that providing PE during a functional analysis of SIB may affect its outcome. For all participants in this study, results obtained from assessments with PE did not match those without PE

286

Le and Smith

and did not provide independent evidence about the variables associated with SIB. Guardians, parents, and advocates often request that PE be used during functional analyses of SIB to reduce risk of injury. However, results of this study indicate that such precautions can confound the outcomes of assessment. Such negative outcomes can be useful if they result in the discontinuation of ineffective or potentially deleterious practices of a discipline. The present ndings provide an empirical basis for Iwata et al.s recommendation . . . that subjects be allowed to engage in self-injurious behavior while free from mechanical, physical or chemical restraint (Iwata et al., 1994a, p. 199). The suppressive effects of PE on the SIB of all 3 participants of this study provide additional support for previous ndings that PE can be used as part of a treatment package (Dorsey et al., 1982; Mazaleski et al., 1994; Silverman et al., 1984). However, long-term treatments should involve alteration of the contingency responsible for behavioral maintenance. Otherwise, as indicated by the present ndings, SIB is likely to reemerge when PE is removed. The present ndings, however, contradict previous studies that demonstrated that PE either did not have an effect on SIB maintained by social contingencies (Iwata et al., 1994b) or increased pica when controlling variables were not known (Mace and Knight, 1986). The reasons for inconsistent results between this and previous studies (Iwata 1994b; Mace and Knight, 1986) are not clear. One likely effect of PE is to attenuate the pain that may be produced by SIB. If so, then PE should not suppress (and might increase) SIB maintained by social contingencies as long as those contingencies remain intact. Future studies investigating the effects of PE on SIB maintained by social contingencies might provide information on whether and how PE can be used to assess social variables during functional analyses. It is possible that the suppressive effects of different types of PE involve different mechanisms (e.g., response cost, stimulus control, extinction, punishment) or that the mechanisms vary among individuals. For example, it is possible that Fred did not engage in head banging when the helmet was in place because more effort was required (i.e., response cost). If so, then padding the top edge of the wheelchair would not be expected to affect Freds SIB because no additional effort would be involved. A history with a particular type of PE may also contribute to its effect on current behavior. Mace and Knight (1986) found that lowest levels of pica were observed when a participant was not wearing a helmet. Furthermore, reprimands were more likely to follow pica if the helmet was not worn, suggesting that the helmet may have acquired a discriminative function by being correlated

Functional Analysis With Protective Equipment

287

with a thinner schedule of punishment. Similarly, it is possible that mittens had a discriminative function for Toms face slapping. For example, if Tom had previously worn mittens during escape extinction procedures, they may have acquired an S-delta function, suppressing SIB during the functional analysis with PE. If mittens were the only type of PE correlated with escape extinction, then other types of PE (e.g., face masks) might not have altered functional analysis outcomes. Time-out or punishment may also be responsible for the suppressive effects of PE on SIB. Mazaleski et al. (1994) found that contingent application of oven mittens decreased SIB by more than 80% as compared to that during baseline for both participants of that study. These researchers suggested that their ndings are more consistent with either punishment or time-out as the operating process rather than than sensory extinction. Similarly, PE may have generally reduced the SIB of the participants in this study because of punishment or a historical correlation with time-out procedures. Future studies will be necessary to determine how, when, and why PE interacts with behavioral contingencies during functional analyses. Such studies might investigate interactions among variables, such as functional properties of SIB, topographies of SIB, types of PE, and previous history with PE. Another possible direction for future research is the effects of PE during treatment procedures. For example, extinction procedures can produce increases in responding that may persist for several sessions (e.g., Goh and Iwata, 1994), thus exposing participants to prolonged risk of injury. PE may be useful to protect participants during extinction; however, PE could interfere with extinction if it eliminates SIB and, thus, contact with the extinction contingency. Both Tom and Fred tolerated events that had previously evoked escape behavior (i.e., task demands or being in the wheelchair) when PE was placed on them. For these 2 participants, PE had generalized suppressive effects on their SIB, even though contingencies identical to the No PE assessment remained intact. Participants in Iwata et al.s study, on the other hand, continued to engaged in SIB while wearing PE as long as contingencies for SIB were identical to baseline condition (Iwata et al., 1994b). Given these inconsistent results, future studies should further investigate how contingencies and histories affect performances with and without PE. A possible limitation of this study is that the No PE assessment always preceded the Assessment with PE. Thus, results obtained during the Assessment with PE might have been confounded because it always followed the No PE assessment. However, the purpose of this study was to compare results of functional analyses with PE to those without PE. The No PE assessment was always conducted rst to establish the conditions (e.g., number of

288

Le and Smith

sessions, additional analyses if applicable, etc.) necessary to draw condent conclusions about variables associated with SIB. Further, previous exposure to a No PE assessment should enhance differentiation during the Assessment with PE because of previous exposure to assessment contingencies. The elimination of SIB with all 3 participants suggests that no such order effect was operating. A second limitation of this study is that no participants SIB was maintained by socialpositive reinforcement (e.g., attention or tangible items, etc.). The SIB of the participants of this study was maintained either by socialnegative reinforcement (Tom and Fred) or by nonsocial mechanisms (Mike). Therefore, the effects of providing PE during a functional analysis of SIB maintained by socialpositive reinforcement were not evaluated. Results of the Iwata et al.s study suggest that PE may not have an effect on SIB maintained by socialpositive reinforcement during a functional analysis (Iwata et al., 1994b). However, Tom and Freds results, showing that socially maintained SIB was eliminated during functional analysis, were not consistent with those of Iwata et al. Additional investigation will be necessary to clarify the reasons for these differences. Finally, Tom and Freds experimental question could have been answered in a more expeditious fashion. During the second No PE assessment for both participants, irrelevant conditions (Alone, Attention, and Play for Tom, and Alone, Attention, Play, and Demand for Fred) could have been omitted because there was prior evidence to suggest that SIB would not occur during these conditions. However, because of the relatively short period of time required to conduct these conditions, and because whether previous exposure to the Assessment with PE would alter outcomes of a subsequent No PE assessment (including the irrelevant conditions) was not known at the time of this study, those conditions were presented in the nal No PE assessment. Future studies could produce more rapid results by excluding irrelevant conditions or by using a multielement design to compare assessments with and without PE. In summary, this study investigated effects of PE during functional analyses. In each case, PE altered the outcomes of functional analysis and limited conclusions about behavioral function. In addition, data showed that PE can eliminate SIB maintained by both social and nonsocial mechanisms, even when social reinforcement contingencies remain intact. Additional research on the effects of PE during functional analysis may clarify the conditions under which assessment outcomes are altered, as well as conditions under which PE attenuates, but does not eliminate, SIB. Such information could provide an empirical basis to guide the use of PE during the assessment and treatment of SIB.

Functional Analysis With Protective Equipment

289

ACKNOWLEDGMENTS Support for this investigation was provided by a grant from the Texas Higher Education Coordinating Board. The authors express their appreciation to Robert Churchill, Jay Cole, Wendy Jacobs, Lori Russo, Rita Ybarra, and the staff and administration of the Denton State School for their assistance during this investigation. Thanks also to Bridget Shore and Amy Crye for their helpful comments on previous versions of this paper.

REFERENCES
Baer, D. M., Wolf, M. M., and Risley, T. R. (1968). Some current dimensions of applied behavior analysis. J. Appl. Behav. Anal. 1: 9197. Barlow, D. H., and Hersen, M. (1984). Single Case Experimental Designs, 2nd edn., Pergamon Press, New York. Derby, M. K., Wacker, D. P., Sasso, G., Steege, M., Northup, J., Cigrand, K., and Asmus, J. (1992). Brief functional assessment techniques to evaluate aberrant behavior in an outpatient setting: A summary of 79 cases. J. Appl. Behav. Anal. 25: 713721. Dorsey, M. F., Iwata, B. A., Reid, D. H., and Davis, P. A. (1982). Protective equipment: Continuous and contingent application in the treatment of self-injurious behavior. J. Appl. Behav. Anal. 15: 217230. Hanley, G. P., Piazza, C. C., Fisher, W. W., Contrucci, S. A., and Maglieri, K. A. (1997). Evaluation of client preference for function-based treatment packages. J. Appl. Behav. Anal. 30: 459 473. Goh, H. L., and Iwata, B. A. (1994). Behavioral persistence and variability during extinction of self-injury maintained by escape. J. Appl. Behav. Anal. 27: 173174. Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., and Richman, G. S. (1994a). Toward a functional analysis of self-injury. J. Appl. Behav. Anal. 27: 197209. (Reprinted from Anal. Interv. Dev. Disab. 2: 320, 1982). Iwata, B. A., Pace, G. M., Cowdery, G. E., and Miltenberger, R. G. (1994b). What makes extinction work: An analysis of procedural form and function. J. Appl. Behav. Anal. 27: 131144. Iwata, B. A., Pace, G. M., Dorsey, M. F., Zarcone, J. R., Vollmer, T. R., Smith, R. G., Rodgers, T. A., Lerman, D. C., Shore, B. A., Mazaleski, J. L., Goh, H., Cowdery, G. E., Kalsher, M. J., McCosh, K. C., and Willis, K. D. (1994c). The functions of self-injurious behavior: An experimentalepidemiological analysis. J. Appl. Behav. Anal. 27: 215 240. Kahng, S. W., Iwata, B. A., DeLeon, I. G., and Worsdell, A. S. (1997). Evaluation of the control over reinforcement component in functional communication training. J. Appl. Behav. Anal. 30: 267276. Mace, F. C., and Knight, D. (1986). Functional analysis and treatment of severe pica. J. Appl. Behav. Anal. 19: 411416. Mazaleski, J. L., Iwata, B. A., Rodgers, T. A., Vollmer, T. R., and Zarcone, J. R. (1994). Protective equipment as treatment for stereotypic hand mouthing: Sensory extinction or punishment effects? J. Appl. Behav. Anal. 27: 345355. Northup, J., Wacker, D., Sasso, G., Steege, M., Cigrand, K., Cook, J., and DeRaad, A. (1991). A brief functional analysis of aggressive and alternative behavior in an out-clinic setting. J. Appl. Behav. Anal. 24: 509522. Sidman, M. (1960). Tactics of Scientic Research, Authors Cooperative, Boston.

290

Le and Smith

Silverman, K., Watanabe, K., Marshall, A. M., and Baer, D. M. (1984). Reducing self-injury and corresponding self-restraint through the strategic use of protective clothing. J. Appl. Behav. Anal. 17: 545552. Smith, R. G., Iwata, B. A., Vollmer, T. R., and Zarcone, J. R. (1993). Experimental analysis and treatment of multiply controlled self-injury. J. Appl. Behav. Anal. 26: 183 196. Vollmer, T. R., Iwata, B. A., Zarcone, J. R., Smith, R. G., and Mazaleski, J. L. (1993). Withinsession patterns of self-injury as indicators of behavioral function. Res. Dev. Disab. 14: 479492. Vollmer, T. R., Marcus, B. A., Ringdahl, J. E., and Roane, H. S. (1995). Progressing from brief assessments to extended experimental analyses in the evaluation of aberrant behavior. J. Appl. Behav. Anal. 28: 561576.