JOURNALOFAPPLIEDBEHAVIORANALYSIS 2006, 39, 381–384 NUMBER3 (FALL2006) A PARADOXICAL EFFECT OF PRESESSION ATTENTION ON STEREOTYPY: ANTECEDENT ATTENTION AS AN ESTABLISHING, NOT AN ABOLISHING, OPERATION CHRISTINA F. ROANTREE AND CRAIG H. KENNEDY VANDERBILTUNIVERSITY Previous studies have shown that presession attention for problem behavior can serve as an abolishingoperationwhenattentionfunctionsasapositivereinforcer.Inthecurrentstudy,we showthatthestereotypyofachildwithseveredisabilitieswasundifferentiatedduringstandard analoguefunctionalanalysisconditions.However,whennoncontingentpresessionattentionwas provided, stereotypy occurred for social attention as a positive reinforcer, suggesting that the antecedent manipulationfunctioned asanestablishing operation. DESCRIPTORS: stereotypy, functional analysis, positive reinforcement, establishing operation,motivating operation, presession attention _______________________________________________________________________________ Motivating operations (MOs) change the nopresessionattention.Suchanalysesshowthat probability of behavior under the control of presession attention can serve as an AO for reinforcement contingencies by altering the attention as a positive reinforcer. However, at valueofreinforcingstimuli(Laraway,Snycerski, least one well-known behavioral technique, Michael, & Poling, 2003). When an MO priming,suggeststhatpresessionattentioncould functionsasanestablishingoperation(EO),the function as an EO, rather than an AO, thus reinforcing effects of a stimulus are increased. increasing the probability of responding rather When an MO functions as an abolishing than decreasing it. Priming involves the non- operation (AO), the reinforcing effects of contingent presentation of a stimulus prior to a stimulus are decreased. Correspondingly, its use as a contingent reinforcer and can serve EOs serve to increase the probability of as an EO for the reinforcing effects of the responding, whereas AOs decrease response stimulus (Catania, 1998). probability. In the current analysis, we studied whether Several studies have analyzed the motivation presession attention could function as an EO ofproblembehaviorintermsofAOs(e.g.,Berg rather than an AO. Initially, the stereotypy of et al., 2000; McComas, Thompson, & John- a child with severe disabilities was analyzed son, 2003; O’Reilly, 1999). For example, using standard analogue functional analysis McComas et al. showed that presession atten- conditions (i.e., alone, attention, control, and tion reduced the probability of problem demand). We then provided noncontingent behavior maintained by positive reinforcement presession attention under the same analogue intheformofattentionrelativetosessionswith conditions to see if attention as a positive reinforcer increased in value via increases in This report is based on a thesis submitted by the first response probability. We then withdrew the author to Vanderbilt University in partial fulfillment of the requirements for the MEd degree. We thank Maria presession attention and observed whether Couppis, Nea Houchins-Jua´rez, Pablo Jua´rez, Michael behavioral patterns reversed to baseline levels. May, and Jill Parks for their comments on a previous versionofthe manuscript. Address correspondence to Craig H. Kennedy, De- METHOD partment of Special Education, Box 328 – Peabody, Vanderbilt University, Nashville, Tennessee 37203 (e- Participant, behavior, and setting. The child mail:[email protected]). doi:10.1901/jaba.2006.97-05 was a 10-year-old boy with severe mental 381 382 CHRISTINA F. ROANTREE and CRAIG H. KENNEDY retardation. He was nonverbal, but used several sion attention. Prior to the start of the no- gestures and vocalizations to communicate his presession-attention sessions, there was no wantsandneeds.Thestereotypicalresponsewas interaction with the participant for 20 min. In a full-body flex that involved quickly raising his noncontingent presession attention, the child hands above his head while simultaneously was provided with 20 min of attention before extending his legs. All sessions were conducted the analogue functional analysis sessions began. in a room (4 m by 4 m) with several tables and Noncontingent attention consisted of positive chairs. verbal comments delivered every 30 s. All Measurementandinterobserveragreement.Ses- presession manipulations occurred in the same sions were videotaped, and the stereotypical roomasthefunctionalanalysis.Eachpresession response was scored as frequency counts using manipulation was followed by a daily block of a paper-and-pencil system. Two observers in- four functional analysis conditions as described dependently scored 50% of sessions. Interob- previously. The order of the functional analysis server agreement was measured by dividing the conditions was randomly determined for the smaller value by the larger and multiplying by first phase and was repeated in each subsequent 100%. Mean interobserver agreement was 88% phase to assess possible sequence effects. (range, 75% to 100%). Designandprocedure.AcombinedA-B-Aand RESULTS AND DISCUSSION multielementdesignwasused(Kennedy,2005). An initial functional behavioral assessment The multielementaspect of the design was used using interviews and observations resulted in an to study stereotypy during an analogue func- inconclusive outcome, but some informants tional analysis involving alone, attention, con- noted that the child enjoyed attention. We trol, and demand conditions (see Iwata, Slifer, conducted an initial analogue functional anal- Dorsey, Bauman, & Richman, 1982/1994). ysis to test the hypothesis that attention was One of each type of condition was conducted a positive reinforcer (see Figure 1). However, once per day. Each condition was 5 min in during the initial analysis with no presession length, with a 5-min break between conditions. attention, an undifferentiated pattern of stereo- Allcontingenteventswereforstereotypy.Inthe typy was observed. To further test the hypoth- alone condition, the participant was in a room esis that attention was a positive reinforcer, we without any activities or social stimulation, and optedtouseprimingofthisbehavioralfunction no consequences were delivered. During the by providing noncontingent attention prior to attention condition, social interaction was functional analysis sessions. When noncontin- withheld until stereotypy was emitted, which gent presession attention was delivered during occasioned 5 s of attention. In the control the second phase, elevated levels of stereotypy condition, the child had access to attention for occurred during the attention condition, but the absence of stereotypy and noncontingent low levels of stereotypy were observed in other access to preferred activities. In the demand conditions.Areturntotheinitialno-presession- condition, instructional requests were made attention procedure again resulted in an un- every 10 s. If the task was completed, the child differentiated pattern of responding, suggesting was praised. If stereotypy occurred, the re- that the behavior was sensitive to attention as searcher removed the materials and did not a positive reinforcer, but that an EO was present them again until stereotypy was absent required. for 30 s. Our data suggest that presession attention The A-B-A aspect of the design compared no acted as an EO, increasing the value of social presession attention with noncontingent preses- attention as a positive reinforcer during the ATTENTION AND MOTIVATING OPERATIONS 383 Figure1. Frequencyofstereotypy(yaxis)andindividualanaloguefunctionalanalysissessions(xaxis).Conditions includedalone,attention,control,anddemand.DuringtheAphasesnopresessionattentionwasprovided,andduring PhaseBnoncontingent presession attention was provided. subsequent functional analysis. In previous the addition of noncontingent presession atten- studies of MO effects on problem behavior, tion. It is possible that presession attention presessionattentionhasbeenshowntofunction actedtoprimeattentionasapositivereinforcer, as an AO but not an EO. Therefore, one thereforeinducingamoredifferentiatedpattern implication of our findings is that individuals than seen during the no-presession-attention whoconductfunctionalassessments ofproblem phases. Given that the undifferentiated pattern behavior need to consider the possibility that was reestablished with the withdrawal of presessioneventscanserveaseitherAOsorEOs presession attention, the effects were probably for attention. It is possible that a curvilinear notduetochangesinthediscriminativecontrol relation exists between exposure to presession exerted by the functional analysis conditions, attention and its effect as either an EO or AO, but rather to an alteration in the reinforcing with EO effects occurring initially and AO values of the various conditions via the EO effects emerging over time. Thus, presession being analyzed. Whether similar results can be manipulations may function as an EO or AO, obtained by manipulating other presession depending on the temporal relation between variables and their potential effects on re- presession exposure and the functional analysis. inforcers remains to be analyzed. However, such a functional relation remains to Our findings also contribute to an emerging be empirically demonstrated. understanding of stereotypy as a complexly The data also suggest that when researchers determined behavior. Stereotypy has often been are presented with undifferentiated patterns of presumed to function primarily as sensory behavior during initial functional analyses, reinforcement (i.e., nonsocial positive or nega- manipulation of presession attention may assist tive reinforcement; Lovaas, Newsom, & Hick- in identifying possible functions of problem man, 1987). However, the results of several behaviorrelatingtosocialattentionasapositive recent investigations (e.g., Kennedy, Meyer, reinforcer. Our data showed that an initially Knowles, & Shukla, 2000) suggest that stereo- undifferentiated pattern of stereotypy was typy can also function to occasion, maintain, changed to a highly differentiated pattern by avoid, or escape socially mediated reinforcers. 384 CHRISTINA F. ROANTREE and CRAIG H. KENNEDY Hence, researchers need to carefully assess the Kennedy,C.H.,Meyer,K.A.,Knowles,T.,&Shukla,S. (2000). Analyzing the multiple functions of stereo- potential functions of stereotypy prior to typical behavior for students with autism: Implica- developing behavioral intervention plans. In tionsforassessmentandtreatment.JournalofApplied addition,ourfindingssuggestthatMOsneedto BehaviorAnalysis,33,559–571. be considered in the analysis of stereotypy as Laraway, S., Snycerski, S., Michael, J., & Poling, A. (2003). Motivating operations and terms to describe functional variables in their own right. them: Some further refinements. Journal of Applied BehaviorAnalysis,36,407–414. Lovaas, I., Newsom, C., & Hickman, C. (1987). Self- REFERENCES stimulatory behavior and perceptual reinforcement. Berg, W. K., Peck, S., Wacker, D. P., Harding, J., JournalofAppliedBehavior Analysis, 20,45–68. McComas,J.,Richman,D.,etal.(2000).Theeffects McComas, J. J., Thompson, A., & Johnson, L. (2003). of presession exposure to attention on the results of The effects of presession attention on problem assessments of attention as a reinforcer. Journal of behavior maintained by different reinforcers. Journal AppliedBehavior Analysis, 33,463–477. ofAppliedBehavior Analysis, 36,297–307. Catania, A. C. (1998). Learning (4th ed.). Upper Saddle O’Reilly, M. F. (1999). Effects of presession attention River,NJ:Prentice Hall. on the frequency of attention-maintained behav- Iwata,B.A.,Dorsey,M.F.,Slifer,K.J.,Bauman,K.E.,& ior. Journal of Applied Behavior Analysis, 32, 371– Richman,G.S.(1994).Towardafunctionalanalysis 374. ofself-injury.JournalofAppliedBehaviorAnalysis,27, 197–209.(ReprintedfromAnalysisandInterventionin DevelopmentalDisabilities,2,3–20,1982) Received July 25,2005 Kennedy, C. H. (2005). Single-case design for educational Final acceptance February 7,2006 research.Boston: Allyn& Bacon. Action Editor,Jennifer J. McComas