Carol Gray first developed social stories because individuals with autism spectrum disorders (ASDs) often have difficulty understanding and responding during social situations (Gray and Garand 1993). As defined by Gray (2000), “…a Social Story is a short story—defined by specific characteristics - that describes a situation, concept, or social skill using a format that is meaningful for people with ASD” (p. 13–1). While detailed information is provided about how to write social stories (Gray 2004) few well-controlled studies have examined their effectiveness, with the first experimental study published within the last decade (Kuttler et al. 1998). In a recent review, Nichols et al. (2005) found only ten experimental, peer-reviewed studies of social stories. Of these ten studies, only two actually focused on actual social skills (Barry and Burlew 2004; Thiemann and Goldstein 2001) while the other studies focused on daily living skills and decreasing disruptive behavior.

Thiemann and Goldstein (2001) included five boys with autism ranging in age from 6 to 12years. Each participant was grouped with two typically-developing peers. Target skills were chosen from a group of four possible skills: increasing contingent responses, securing attention, initiating comments, and initiating requests. A multiple-baseline-design-across-behaviors was used for each participant. The baseline phase consisted of social interactions between the child with autism and the two peers. The intervention phase consisted of sessions during which time the child with autism read a social story and then interacted with peers. In general, results demonstrated that the four social skills improved after the introduction of social stories. At least two social skills for each participant increased.

A study by Barry and Burlew (2004) investigated the effects of a social story on the choice-making behavior and appropriate play skills of two children with severe autism. “Holly,” a 7-year-old girl, had receptive language skills, but she did not initiate speech beyond yelling, “no.” “Aaron,” an 8-year-old boy, did not speak other than exhibiting echolalia, and he did not read. Aaron was able to respond to picture prompts. An ABCD multiple-baseline-design-across-participants was implemented. Target behaviors were making independent choices about where to play and exhibiting appropriate play behaviors at the play center (i.e., interacting with play materials or peers appropriately). The study included a baseline control phase (A), a social story intervention phase (B) that focused on choice-making and appropriate play with materials, and a social story intervention phase (C) that focused on appropriate play with peers. A fourth phase (D) consisted of reading the social story at the beginning of the school day, but the story was not read immediately before the play sessions. Results indicated that choice-making behavior increased steadily as demonstrated by reduced number of prompts necessary to have the child go to the play center. Aaron did not actually interact with any peers during the study, but he did engage in parallel play. As a result of Holly’s newly acquired play skills she was placed in a general education classroom where she immediately chose two girls in the class as friends. This study is unique because it focused on children with very limited language skills.

Since the Nichols et al. (2005) review was published, two additional studies have been published that examine the use of social stories on the social skills for children with ASDs (Delano and Snell 2006; Sansosti and Powell-Smith 2006). Both of these single-case research studies were well-designed and demonstrated promising results for the intervention. However, these four published studies include just 13 children, thus far more research needs to be conducted involving social stories and social skills.

Several other problems exist within the current literature. First, in research studies social stories are often combined with other interventions (e.g. schedules, timers, cues, corrective feedback, prompts) during the intervention phase, confounding the unique effects of the social story (Barry and Burlew 2004; Hagiwara and Myles 1999; Kuoch and Mirenda 2003; Kuttler et al. 1998; Lorimer et al. 2002; Thiemann and Goldstein 2001). Second, parents and siblings are rarely included in this research, thus overlooking a valuable and economical source of interventionists and peers. Third, the literature also lacks studies that use a placebo control to account for the extra adult attention received during a social story. Fourth, few studies have examined the maintenance of gained skills. Finally, the literature rarely discusses treatment acceptability of social stories as rated by parents or children. To some extent, the present study attempts to address all of these limitations. We hypothesized that the social story would be effective at decreasing a social skill excess (i.e., directions) for one participant. We also hypothesized the social story would increase a social skill deficit (i.e., compliments) for both participants.

Materials and Methods

Participants and Setting

Two families were recruited for research using flyers that were posted at a local clinic. Both participants had previously been diagnosed with Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS). Each of the participants met the criteria of a second grade reading level as indicated by their mothers. The study took place in participants’ homes and mothers implemented the study procedures.

“Mark” was a Caucasian male who lived with his biological parents and two younger brothers. He was 9years-10months of age and in the fourth grade. He was diagnosed with PDD-NOS by a pediatric neurologist one month before the study took place. Mark was polite and could carry on a conversation, although his eye-contact was poor. Mark smiled often and he appeared to be a happy most of the time. His national grade percentile ranks on the Illinois Stanford Achievement Test were: Reading 43%; Math 39%; Language 54%; Science 81%; Social Science 91%; and Listening 62%. His mother reported that Mark had difficulty making friends at school, and he attended a weekly social skills group for children with autism spectrum disorders. Mark’s mother had attended a social stories presentation by Carol Gray, and she had previously written social stories for Mark. She indicated that these stories had helped Mark in situations such as school and family gatherings. Mark had not used social stories recently, however, and he did not read social stories other than those used in this study during the month the study was conducted. Mark’s younger brother (8years-10months of age) took part in data collection play sessions with Mark. Mark’s brother was diagnosed with high-functioning autism at a young age and he received intensive early intervention with applied behavior analysis. At the time of the study, Mark’s brother was functioning at a high level and his mother did not have any particular concerns about his behavior. The study focused on Mark, because his mother was more concerned about his social skills.

“Logan” was a Caucasian male who lived with his biological parents and a younger brother. He was 12years-7months of age and in the seventh grade at the time of the study. Logan was diagnosed with PDD-NOS at age six by a pediatric neuropsychologist. He received special education services, but he was in regular education classrooms. Logan received resource support, and he had a personal aide in the classroom. He also could carry on a conversation, and his mother described him as very creative and imaginative. He tended to speak formally, often sounding like a “little professor.” He also tended to be a perfectionist, becoming frustrated when he did not perform as well as he thought he should. Logan’s mother reported that he had difficulty navigating the social environment at school, and he was often bullied. He attended a weekly social skills group for children with ASDs and was assessed by a school psychologist at 10years-10months of age. On the Wechsler Intelligence Scale for Children, 3rd Edition, Logan received a Verbal IQ of 105, a Performance IQ of 120, and a Full Scale IQ of 107. Logan’s verbal skills were within the average range and his nonverbal skills were within the superior range. His overall intellectual ability was within the average range. On the Wechsler Individual Achievement Test Logan received composite scores of 84 in Reading and 95 in Mathematics, which are in the low average and average ranges, respectively. Logan’s mother stated that he had not been exposed to social stories before. Logan’s younger brother (10years-6months of age) took part in the data collection play sessions with Logan.

Measures and Data Collection

The main dependent measures were frequencies of occurrence of target social skills. The goals for Mark were to decrease the excessive “directions” he gave to his brother and increase the “compliments” he gave for his brother’s ideas. The target behavior for Logan was also “compliments” and focused more on compliments within the context of being a good sport (see Appendix A for operational definitions). Data were collected over a period of 4weeks with the number of sessions per week varying between zero to three. Parents collected data by videotaping their childrens’ play sessions. Mark’s play sessions were 15min and Logan’s play sessions were 10min. The time differences were due to the parents’ preferences for length of play session. Three of Logan’s play sessions were longer than 10min so these sessions were divided into two sessions and coded as separate sessions.

At the end of the 4-week data collection period, the children were given a three-item questionnaire to assess intervention acceptability (e.g., “How much did you like reading your social story?”). This form included smiley-face and frowning-face pictures to enhance the meaning of the 4-point Likert scale for the first two questions. The third question was an open-ended question. A general questionnaire was given to parents to obtain descriptive information. The questionnaire obtained qualitative information by inquiring about positive and negative consequences of the study and subjective impressions of improvement. In addition, parents were asked if they would continue using the social story created for this study and if they planned to create more social stories. Suggestions for improvement in the study were also gathered.

Research Design

Since “Mark” had two target behaviors, the study employed a multiple-baseline-across-behaviors design for him, and the “directions” behavior included a maintenance phase. For the behavior of “compliments,” the study employed a multiple-baseline-design-across-participants.

Procedures

After parents expressed interest in the study, they met with the principal investigator to further discuss participation in the study. During an initial meeting the study was explained, informed consent was signed, and possible target behaviors and play scenarios were discussed. Parental input was used to decide that Mark and his brother would play with cars, and Logan and his brother would play a popular children’s card game (“Yu-Gi-Oh!”).

Baseline phase

A second meeting was held with each parent to explain the procedures of the baseline phase. Each parent was given a tip sheet for completing the baseline activity and an experimental procedures checklist. The tip sheet encouraged parents to choose a certain time of the day that they believed they would be likely to observe problem behaviors. Parents were asked to find a quiet room and have their children read one page aloud to them from a favorite book (a “non-social” story). During baseline, a “non-social” story was used as a placebo in order to control for receiving adult attention and reading a story before the activity. Next, they were to turn the camera on and have the siblings play together for ten to 15min. The parents were asked not to be involved in their children’s interactions unless the children were in physical danger. At the end of the play session, the parents were asked to provide a small reward to each sibling for participating in the play session. The reward was not contingent upon performing the target behaviors.

Intervention phase

During the social story phase, the procedure remained the same as the baseline phase except for the substitution of social stories (see Appendices B and C) in place of the “non-social” story. For each story, three comprehension questions were added to verify the children’s understanding of the main themes. Parents were instructed to ask the child the questions and if the child was not providing an appropriate answer, the parent was to provide an answer. Mark gave excessive directions to his sibling, so the behavior of “directions” was chosen as a target behavior to decrease and a social story was written and titled, “Giving Just a Few Directions Makes Playing Fun.” The story met Gray’s (2004) guidelines and had a total of 17 sentences over 8 pages. Standard 8 1/2 × 11 cardstock was folded in half to create a booklet. The story included nine photographs of Mark and his brother playing with their cars. The photographs were included to increase visual interest and to include a special interest of Mark’s (i.e., the cars). The social story for Mark was discussed with his mother to assure that she was in agreement with the story and to elicit suggestions.

A second social story was also created for Mark, which aimed to increase the prosocial behavior of “compliments.” The “compliments” social story titled, “Listening to Others’ Ideas” and was a total of 15 sentences over 7 pages. The story included six new photographs of Mark and his brother playing with their cars. Mark’s mother was asked to provide suggestions for the story. The “directions” social story was no longer read during the next phase and the “compliments” story was introduced.

For Logan’s intervention phase, an individualized social story was written with the aim of increasing his use of compliments. His social story was titled, “Being a Good Sport.” Thus, the operational definition for compliments for Logan had a greater emphasis on sportsmanlike behavior during games. His social story was ten sentences over seven pages. Cardstock was used to construct the booklet and clip-art and “Yu-Gi-Oh!” images from the Internet were used to illustrate the story. Pictures and references to “Yu-Gi-Oh!” were included in order to incorporate Logan’s special interest.

After all the sessions were completed, the children and mothers answered the acceptability questions. The participants and siblings were each given a university-logo water bottle and the parents were presented with a gift card for a local restaurant to thank them for their participation. As another token of appreciation, parents were given the Jenison Autism Journal, which included the most up-to-date social story criteria and suggestions for writing social stories (Gray 2004).

Interobserver agreement

Reliability checks were completed randomly for the baseline and intervention phases for each of the three behaviors. The first author was a graduate student studying clinical-child and school psychology, and she rated the frequency of behavior for all of the sessions. An undergraduate research assistant, who was blind to condition, rated 33% of the sessions for each behavior. The research assistant was a junior and a psychology major. She had previous experience with children with autism as an applied behavior analysis therapist. The reliability was calculated by dividing the smaller frequency by the larger frequency and multiplying by 100%.

Procedural integrity

Parents completed a self-report checklist of their compliance with the experimental procedures after each session. The experimental procedures checklist was a calendar that included the steps for completing the study procedures. During the baseline phase the checklist had four steps: (1) Read story, (2) Camera on, (3) Activity, and (4) Reward. During the intervention phase the checklist had five steps: (1) Read story, (2) Comprehension questions, (3) Camera on, (4) Activity, and (5) Reward. Procedural integrity was calculated by dividing the number of correct steps by the total number of steps and multiplying by 100%.

Results

Interobserver Agreement and Procedural Integrity

Mean interobserver agreement for directions was 65.1%. Interobserver agreement for compliments for both children was 100%. The procedural agreement for Mark was 100% and the procedural agreement for Logan was 97.1%. The 97.1% agreement was due to one missed step in the baseline phase. The parent stated that she and her son forgot to read a page from his favorite book before the play session. In addition, the videotapes showed that the mothers never interfered in their children’s playtime during the sessions.

Social Skills

As shown in Fig. 1, during the baseline phase Mark had a stable and flat trend, and he exhibited consistently high frequencies of directions (M = 77.7). When the social story for directions was introduced, the frequency of directions dropped immediately with a decreasing trend (M = 43.7). From baseline to intervention, this was an overall 43.8% decrease in directions. After three sessions of using the social story for directions, it was withdrawn, and directions continued to decrease (M = 36.0). During the first six sessions, Mark did not give his brother any compliments (M = 0). During the Social Story phase, there was an increasing trend of compliments (M = 2.0).

Fig. 1
figure 1

Frequency of Social Skills during Baseline and Social Stories for Mark (top 2 panels) and Logan (bottom panel). Mark’s sessions were 15 min and Logan’s sessions were 10 min. Logan’s data for the following sessions were collected on the same day: 2 and 3; 5 and 6; 7 and 8

Full size image

During the baseline phase, Logan exhibited only one compliment, and the trend was stable and flat (M = 0.11). After the social story was introduced, Logan’s compliments increased dramatically to seven. While compliments in the Social Story phase were higher, compliments declined over the course of the phase (M = 2.25). Overall, there was a 19.5% increase in compliments from baseline to intervention.

Treatment Acceptability

Mark and Logan both indicated that they “somewhat liked” reading the social story. In addition, they “somewhat liked” learning something new. Mark stated that his favorite part about the social story was looking at the pictures. He also liked getting to play with his cars after he read the social story. Logan stated that his favorite part about the social story was “the part on how to lose like a good sport.”

Both mothers reported that their children willingly read the social stories. Mark’s mother stated that as a positive consequence of the study, she learned more about social stories. Logan’s mother expressed that a positive consequence of the study was being able to have her son focus on a particular skill before an activity. She also liked being able to use an existing situation to work on social skills. Both parents were unsure if their children’s social skills had improved and they felt more time would be needed after the intervention to observe the effects. Both parents planned to continue using the social stories created for this study with their children. Mark’s mother estimated that she would use the story three to four times a week, while Logan’s mother believed she would use it approximately once a week before similar activities. Mark’s mother stated that as issues arose, she would write more social stories. Logan’s mother also expressed a desire to use the social story with her younger son who did not have an autism spectrum disorder. Logan’s mother also wanted to write a social story for Logan to address saying “no” to peers when pressured to engage in inappropriate activities.

Discussion

Children with autism spectrum disorders face many social challenges. Numerous interventions have been developed to improve the social skills of children with autism. Social stories are currently a popular intervention; however, with limited research it is important to further investigate this intervention. This study was designed to examine the effects of social stories on the behavior of two children with PDD-NOS.

The hypothesis that directions would decrease with use of a social story was supported with Mark. After the social story was introduced, Mark’s directions decreased and stabilized. The results are consistent with other findings that social stories can be effective in decreasing problem behaviors (Brownell 2002; Kuoch and Mirenda 2003; Kuttler et al. 1998; Lorimer et al. 2002; Scattone et al. 2002). In addition, the decreased level of directions was maintained, and continuing to decrease, after the intervention was withdrawn. This most likely indicates that the social story was no longer needed to decrease the behavior.

The other hypothesis that compliments would increase when using a social story was supported for both children. Mark’s compliments increased gradually after the social story was introduced, but there was not immediate change. Logan’s compliments increased immediately. This increase was not maintained during the intervention phase, and the compliments continued to decrease during the Social Story phase. Thus, the results are somewhat mixed. Overall, the results indicate that prosocial skills were learned, but it appears that the short intervention did not allow enough time for the skills to stabilize. Adams et al. (2004) and Thiemann and Goldstein (2001) also demonstrated increases in prosocial skills but had longer intervention periods, which may explain why the gains were more pronounced.

Results of the acceptability questions showed that the children somewhat liked reading the social story, but the intervention did not receive the highest rating possible from the children. The parents reported anecdotally that the children willingly read the social stories. Incorporating colorful images and children’s interests may have helped increase the children’s willingness to read the stories. Both mothers stated that the social stories intervention was an acceptable method of intervention. In addition, they both indicated that the intervention was likely to be effective and to result in continued improvement. Overall, both children and parents had a positive reaction to social stories. These positive evaluations are important because it shows that children enjoy reading social stories and parents are likely to consistently implement a social story intervention. Our results corroborate the findings of the two other studies that addressed intervention acceptability (Adams et al. 2004; Scattone et al. 2002). Both mothers were unsure if their children’s social skills had improved, however, they indicated that they planned to continue using the social stories created for this study and write more social stories. Logan’s mother also stated that she would like to use the social story with her other son, who does not have an autism spectrum disorder. The parents’ intent to use social stories in the future provides further support for the acceptability of the intervention.

The current study has some strengths in comparison to previous research. Specifically, this study included children diagnosed with PDD-NOS, which have been rarely included in the literature. This was the only study to involve siblings and one of the few to involve parents. In addition, the study controlled for adult attention by including the “non-social” story in the baseline condition. The far majority of previous studies did not control for this variable. In the current study, children read the stories aloud, which encouraged them to be engaged in the story, whereas most of the literature involved adults reading to children. Intervention acceptability was assessed in the current study, while most studies do not address acceptability. One final strength was that no additional interventions were added during the intervention phase. This allowed for the effects of the social story to be assessed independently of other variables.

The study also has certain limitations, which have implications for future research. One limitation was the length of the intervention. The stories were used for only three to four sessions. Although effects were seen, there may not have been enough sessions to gauge the stability of the behaviors. In addition, the intensity of the intervention was low. The sessions were completed over 4-weeks, with the sessions spaced over uneven intervals. Future research could focus on increasing the length or intensity of intervention. The current study did not assess for long-term effects after cessation of the intervention or for generalization outside of the training environment. Future research could examine the long-term effects and generalization. Another major limitation was that the primary observer of this study was not blind to condition. Although a blind observer was used for reliability checks for 33% of the data, future studies would benefit to have a blind primary observer. In addition, the interobserver reliability was low for the directions. The current study did not compare the social stories intervention to another intervention. A comparison of social stories to other interventions would be an interesting direction for future research. Finally, a limitation of the general social story literature is that females and minorities are extremely underrepresented and research with these populations is warranted. Research with other disabled populations or with typically developing children (Burke et al. 2004) may also prove fruitful.

Social stories are economical and straightforward, and can be created and implemented by a variety of individuals such as parents, teachers, and aides. Another benefit of social stories is the brevity of intervention; it only takes a few minutes to read a story. Social stories also offer intervention flexibility, and can be created for almost any topic or situation and they can be easily modified as needed. With promising research results, social stories appear to be a viable intervention option.