Teaching Children with Autism to Tell Socially Appropriate Lies

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2016, 49, 1–6 NUMBER 2(SUMMER)

TEACHING CHILDREN WITH AUTISM TO TELL SOCIALLY APPROPRIATE LIES

RYAN BERGSTROM AND ADEL C. NAJDOWSKI CENTER FOR AUTISM AND RELATED DISORDERS, INC.

MARISELA ALVARADO FOOTHILL CHILD DEVELOPMENT SERVICES, INC.

AND

JONATHAN TARBOX CENTER FOR AUTISM AND RELATED DISORDERS, INC.

This study used a nonconcurrent multiple baseline across participants design to evaluate the use of rules, role-play, and feedback for teaching 3 children with autism spectrum disorder to tell socially appropriate lies when (a) presented with an undesired gift and (b) someone’s appearance changed in an undesired way. The intervention was effective in teaching use of socially appropriate lies, and generalization to untrained people and gifts or appearances was observed. Key words: autism, deception, lie, perspective taking, theory of mind

Although lying is broadly considered to be (e.g., allowing someone to wear inappropriate problem behavior, there are situations in which clothing to work may be more detrimental than it is socially appropriate. For example, decep- hurting his or her feelings). Thus, the mediat- tion is used when keeping secrets, avoiding giving responses associated with telling a socially ing away surprises, telling jokes (avoid giving appropriate lie are primarily verbal responses away punch line), and bluffing in games. It is and are highly conditional on other environ- also socially appropriate to be deceptive when a mental contexts. Given the advanced condi- friend solicits an opinion on his or her unat- tional discriminations and language repertoires tractive new haircut or when a spouse asks needed to tell a socially appropriate lie, it is how he or she appears in ill-fitting clothing. not surprising that individuals with autism For most individuals, the determination of spectrum disorder (ASD) may experience diffi- how to respond in the latter case involves culty understanding and telling such lies (a) assessing the appearance of the individual, (Happé, 1994). (b) envisioning the emotional effect of stating Although there are curricula that exist for the truth, and (c) envisioning the consequences teaching deception to individuals with ASD for that individual of telling the truth or lying (e.g., Skills), there is a dearth of research on teaching adaptive deception skills to this popu- Ryan Bergstrom is now at The Help Group, Adel Naj- lation. Reinecke, Newman, Kurtz, Ryan, and dowski is now at The ABRITE Orgnization, Marisela Hemmes (1997) investigated the effects of a Alvarado is now an independent consultant, and Jonathan Tarbox is now at First Steps for Kids. procedure for teaching children with ASD to We thank April Love, Angela Persicke, Aimee Fuentes, hide objects in their hands from an observer and Talya Vogel for their assistance with this project. who guessed in which hand it was hidden. Address correspondence to Adel Najdowski, The ABRITE Organization (e-mail: [email protected]). Modeling and delivery of praise were used dur- doi: 10.1002/jaba.295 ing baseline. Treatment included the addition

1 2 RYAN BERGSTROM et al. of edible reinforcement for independent displayed rule-governed behavior; and were able approximations toward the desired response. to learn via role-play. We selected participants Two participants learned the response during for inclusion based on parental indication that baseline, and one participant required edible their children made blatantly honest utterances reinforcement to learn the response. Although often interpreted as rude (e.g., asking, “Why that study lacked experimental control, results are you wearing a girl’s shirt?” when seeing a tentatively suggest that behavioral procedures boy wearing a pink shirt). might be effective for teaching deceptive skills. Therapists implemented one to three one- Another behavioral procedure, behavioral trial sessions per day during periods between skills training (BST), which involves the use of other regularly scheduled instructional therapy instructions, modeling, rehearsal, and feedback, programs. Baseline and generalization sessions has been demonstrated to be effective for teach- lasted less than 1 min, and training sessions ing social behavior in individuals with ASD. lasted 5 to 10 min. Specifically, Stewart, Carr, and LeBlanc (2007) and Peters and Thompson (2015) demon- Response Measurement and Interobserver strated that BST improved social skills related Agreement to attending to a conversational partner’s inter- Observers scored participant responding on est. BST has also been shown to be effective in each trial by assigning up to 3 points. To teaching children with ASD to detect and receive 3 points, a participant was required to respond appropriately to the deceptive state- tell a lie expressing approval (e.g., “I like it”), ments of others (Ranick, Persicke, Tarbox, & using a sincere tone (i.e., speaking in a higher Kornack, 2013). Specifically, in the Ranick pitched, nonmonotone voice) while smiling et al. (2013) study, participants learned to iden- (and not engaging in inappropriate facial tify when others were lying to them either expressions such as eye rolling). To receive (a) to take their things or (b) to leave them 2 points, a participant was required to tell a lie out. Although no previous research has been expressing approval using either a sincere tone conducted on using BST to teach individuals or smiling, but not both. To receive 1 point, with ASD to use deceptive statements, previous the participant was required to tell a lie expres- research is promising. sing approval with an insincere tone and with- The current study evaluated whether BST, out smiling. Failure to tell a lie expressing which consisted of rules, role-play, and feed- approval, regardless of tone or facial expression, back, could be used to teach children with resulted in a score of zero. ASD to use socially appropriate lies when given A second observer simultaneously collected an undesired gift and when someone’s appear- data on 57%, 54%, and 56% of sessions for ance changed in an undesired way. Leo, Kathy, and Carl, respectively, to calculate interobserver agreement. We compared obser- METHOD vers’ records of the presence or absence of the lie, tone, and facial expression on each trial. Participants and Setting Observers agreed on 100% of trials across all Two boys (5 and 9 years old) and one girl participants. (7 years old) who had been diagnosed with ASD participated. Each received one-on-one behavioral intervention for 8 to 30 hr per week Procedure in his or her home. Each spoke in full sen- We taught participants to tell lies in two tences using mands, tacts, and intraverbals; contexts. During gift sessions, an adult SOCIALLY APPROPRIATE LIES 3 presented the child a wrapped gift containing nice like, ‘Thanks! I like it!’” Appearance ses- either a nonpreferred or already owned item sions began by the therapist saying, “If some- and asked, “What do you think?” These items one is wearing something you don’t like or were determined by parental nomination and changes how they look, you need to make sure included board games, dolls, pens, stickers, or not to hurt their feelings by saying something puzzles with disliked character themes (e.g., nice if they ask you what you think. Something princess, Simpsons, Cars, popular singer); aca- like, ‘It looks good’ or ‘that’s cool.’” Note that demic workbooks; pink items (e.g., magic the therapist did not provide any instruction wand, candles); items to do with coloring (e.g., regarding tone or facial expression but did coloring books, crayons, colored pencils); and model the target response. plush toys. After stating rules during gift sessions, the During appearance sessions, an adult’s therapist initiated a role-play opportunity by appearance was altered in a way that the child presenting a gift and saying, “Hey! I got you a did not like based on parent report. When the present!” After it was opened, the therapist adult arrived at each participant’s home, the asked, “What do you think of it?” After stating adult expressed satisfaction with his or her new the rules during appearance sessions, the thera- appearance (e.g., “Check out my awesome new pist left the room and changed his or her shirt!”) and asked, “What do you think?” The appearance (e.g., put on a cowboy hat) and stimuli used to alter appearance included hair then returned and solicited the participant’s (e.g., styles, clips, large bows, blonde exten- opinion (e.g., “Check out my awesome hat! sions); pink clothing (e.g., shirt, sunglasses, What do you think?”). glove); baggy clothing; hats (e.g., sideways, The therapist praised correct responding cowboy); eyeglasses and eyeglass chains; unu- (i.e., gave a score of 3) if it occurred within 3 s. sual shoes; fake facial hair; and bright lipstick If no response occurred within 3 s, the thera- colors. Across sessions, gifts and changes in pist provided a rule reminder and model of the appearance were presented by experimenters, correct response (e.g., “You need to smile and therapists, or confederates; all are referred to as say something nice like, ‘Wow! This is awe- therapists hereafter. some!’”). If the participant response did not Baseline. Three to five therapists presented a receive a score of 3, the therapist provided feed- gift or change in appearance across sessions; back regarding the missing elements (e.g., they did not deliver feedback or other conse- “That was good, but remember you need to quences for participant responding. smile and sound excited.”). The therapist did Training. Training procedures (adapted from not require participants to repeat the response, the “deception” lesson in the Skills curriculum; but instead began the next session. The thera- (www.skillsforautism.com) included providing pist did not present any of the gifts or appear- descriptive rules, role-playing, and when ances presented in baseline during training; needed, corrective feedback. One therapist con- these were reserved for generalization sessions. ducted all rules and role-play sessions across After participants received 3 points across both gift and appearance contexts. Gift sessions three consecutive sessions, rules and role-play began by the therapist saying, “Sometimes you were no longer presented at the start of ses- might get a gift you don’t like or already have, sions. The therapist provided in situ contingent and you won’t like it. It was nice of the person feedback (CF) when participants responded to give you a gift, and you don’t want to hurt incorrectly or failed to respond within 3 s of their feelings, so even though you are not the prompt. Specifically, the therapist said, happy, you should smile and say something “Remember, even if you don’t like it, you need 4 RYAN BERGSTROM et al. to smile and say something nice so you don’t These results demonstrated that our inter- hurt my feelings.” To minimize reactivity, vention was effective in teaching all participants therapists conducted only one training session to use socially appropriate lies. These per day, and multiple therapists arranged gift outcomes were obtained quickly (each session and appearance trials. The therapist still praised consisted of one trial), and training resulted in appropriate responses. We terminated training generalization to untrained people and when correct responding to novel people and stimuli. These results add to the literature stimuli was elevated and stable. regarding the teaching of complex social To program for generalization, three to six skills to children with ASD. This is the first different therapists presented gifts and changes study of which we are aware that has targeted in appearance across the CF phase. Only one socially appropriate lying among children with therapist from baseline was involved in train- ASD. Our results are promising, but there ing; responding to the other baseline therapists remain questions to be addressed in future was assessed during a generalization assessment. research. Generalization assessment. After completion First, participants in this study had well- of training, therapists and scenarios presented established verbal repertoires, which likely in baseline were re-presented using identical facilitated their acquisition of these skills. procedures. All therapists and gifts from base- Future research should examine the prerequi- line were reassessed, but each gift was given by sites necessary to learn this skill and how the a different therapist. Again, none of these thera- speed of acquisition is affected by the presence pists were present during training, nor were or absence of particular skills. Second, we tested any of these gifts presented during training. for generalization across adults and stimuli, but we did not assess generalization across settings, which will be important to ensure that children respond appropriately to individuals with novel RESULTS AND DISCUSSION appearances in school or when they receive a Figure 1 displays the results of the evaluation gift outside the home. Third, we taught chil- for Leo, Kathy, and Carl. The gift scenario was dren to respond to the question, “What do you not included for Leo (top panel), because his think?” in both the gift and appearance ses- performance met mastery criteria during base- sions. It is possible that socially appropriate lies line for undesired gifts. During baseline, each came under the stimulus control of this ques- participant received scores of 0 points across all tion and facilitated high levels of responding sessions (both gift and appearance). During during the generalization assessment. Future rules and role-play, performances increased to research should include multiple question topo- scores of 3 within a few sessions. When preses- graphies to ensure that responding is not sion rules and role-play were discontinued under selective stimulus control. Fourth, it is (CF on the figure), participants continued to unclear which components of treatment were receive 3 points. Due to therapist error, the responsible for behavior change. Omission of CF-only phase was not implemented with Carl the CF-only phase for Carl before generaliza- for the appearance scenario. The therapist tion sessions were conducted indicated that this instead conducted a generalization session, and gradual fading of the interventions was not nec- Carl responded appropriately. Given his suc- essary to see generalization and maintenance of cess, generalization sessions were continued. All the targeted skill. Future research could con- participants received 3 points during generaliza- duct generalization probes to determine if fad- tion sessions. ing is necessary. SOCIALLY APPROPRIATE LIES 5

Rules, Role-play Baseline& CF CF Only Generalization Assessment 3

1 Leo (Appearance) 0

1 Kathy (Gift) 0

Points 1 Kathy (Appearance) 0

1 Carl (Gift) 0

1 Carl (Appearance) 0 1 6 11 16 21 26 31 Sessions

Figure 1. Performance for Leo (top panel), Kathy (second and third panels), and Carl (bottom two panels). Filled circles represent baseline and generalization sessions, and open circles represent training sessions. CF = contingent feedback. 6 RYAN BERGSTROM et al.

Finally, we believe that our results provide and Developmental Disorders, 24, 129–154. doi: support for the use of behavior-analytic inter- 10.1007/BF02172093 Peters, L. C., & Thompson, R. H. (2015). Teaching chil- vention for skills referred to by some as theory dren with autism to respond to conversation partners’ of mind (Baron-Cohen, 1995). These skills interest. Journal of Applied Behavior Analysis, 48, involve perspective taking and are commonly 544–562. doi: 10.1002/jaba.235 identiﬁed as deﬁcits of individuals with autism. Ranick, J., Persicke, A., Tarbox, J., & Kornack, J. A. (2013). Teaching children with autism to detect and Although many of the skills thought to contrib- respond to deceptive statements. Research in Autism ute to theory of mind cannot be directly Spectrum Disorders, 7, 503–508. doi: 10.1016/j. observed or taught, these results provide sup- rasd.2012.12.001 Reinecke, D. R., Newman, B., Kurtz, A. L., port that observable behaviors thought to Ryan, C. S., & Hemmes, N. S. (1997). Teaching account for a theory of mind can be addressed deception skills in a game-play context to three ado- using behavioral intervention. lescents with autism. Journal of Autism and Develop- mental Disorders, 27, 127–137. doi: 10.1023/ A:1025835706522 Stewart, K. K., Carr, J. E., & LeBlanc, L. A. (2007). Eval- REFERENCES uation of family-implemented behavioral skills train- ’ Baron-Cohen, S. (1995). Mindblindness: An essay on ing for teaching social skills to a child with Asperger s – autism and theory of mind. Cambridge, MA: Bradford disorder. Clinical Case Studies, 6, 252 262. doi: Books/MIT Press. 10.1177/1534650106286940 Happé, F. G. E. (1994). An advanced test of theory of mind: Understanding of story characters’ thoughts Received August 10, 2015 and feelings by able autistic, mentally handicapped, Final acceptance October 19, 2015 and normal children and adults. Journal of Autism Action Editor, Jeffrey Tiger