Cala Luna logoCala Luna by Little Lynx Studios

Research · evidence review · 9-minute read

Do social stories work? What 30 years of research shows

An honest summary of the evidence on social stories: the meta-analyses, the largest randomised trial, what works best, and where the research is still thin.

By Oleksandr Prykhodko, founder of Cala Luna · Reviewed by the Cala Luna team · Last reviewed 24 October 2026

In short

  • Social stories are short, literal stories that describe one situation from a child's point of view. Carol Gray introduced them in the early 1990s for autistic pupils (Gray & Garand, 1993).
  • Most of the evidence comes from small single-case studies. Reviews find real but variable effects, strongest for reducing one specific problem behaviour and weaker for teaching broad social skills (Kokina & Kern, 2010); (Qi et al., 2018).
  • The 2020 US national review of autism interventions lists social narratives as one of 28 evidence-based practices (Steinbrenner et al., 2020).
  • The largest randomised trial to date (249 children, 87 English primary schools) found no meaningful change in overall social responsiveness, but children met their individual, teacher-set goals more often, at slightly lower cost than usual care (Wright et al., 2024).
  • The practical lesson is consistent: one situation, one clear goal, read shortly before the moment, with the child involved and understanding checked.

What a social story is

A social story is a short narrative, usually written for one child, that explains a situation the child finds confusing or hard: what happens, who is there, why people act as they do, and what the child might do. Gray's guidance asks authors to describe more than they direct, to use a patient and positive tone, to write literally, and to avoid promises that may not come true; words such as "usually" and "sometimes" are preferred to "always" (Gray & Garand, 1993). A good story answers the questions a child would ask, rather than giving a list of rules.

That design choice matters. A social story is not a behaviour chart in disguise: its first job is to share accurate information, on the assumption that some difficult behaviour comes from not knowing what will happen or what others expect.

How the research has been done

Most studies of social stories use single-case experimental designs. A small number of children, often one to four, are measured repeatedly before and after the story is introduced, sometimes with the start staggered across children or settings. When behaviour changes only after the story arrives, each time, that is evidence of an effect. These designs suit individual interventions but make it hard to estimate how well an approach works for children in general. Reviews usually summarise them with overlap statistics such as the percentage of non-overlapping data (PND): the share of measurements after the intervention that beat the best measurement before it.

What the reviews found

Early reviews. Reynhout and Carter reviewed the published studies of social stories for children with disabilities and concluded that effects were highly variable and that many studies had design weaknesses, including combining the story with other supports so that its own contribution was hard to isolate (Reynhout & Carter, 2006).

Meta-analysis. Kokina and Kern analysed 18 single-case studies published up to 2009. The average PND was 60%, below the 70% often used as a threshold for an effective intervention, so the overall effect was judged questionable. But the averages hide a clear pattern. Stories worked substantially better when they targeted the reduction of one behaviour than when they tried to teach broad social skills. They were also more effective when read shortly before the target situation, when they described a single behaviour rather than a long chain, when sessions were brief, when understanding was checked with comprehension questions, and when the child or a natural adult, such as a teacher, delivered the story (Kokina & Kern, 2010).

Systematic review against quality standards. Qi and colleagues assessed 22 single-case studies against What Works Clearinghouse standards. By visual analysis, only 5% of studies gave strong evidence of a causal relation, 27% moderate and 68% weak, so on that test social stories did not qualify as evidence-based. On overlap measures they did meet the threshold. Effects on reducing inappropriate behaviour were consistently strong, effects on increasing new skills and social communication were mixed, and children varied widely in how much they benefited (Qi et al., 2018).

National evidence reviews. The US National Professional Development Center review of 1990–2011 research (Wong et al., 2015) and its 2020 update covering 1990–2017 (Steinbrenner et al., 2020) both list social narratives as an evidence-based practice. The 2020 report defines them as interventions that describe social situations to highlight the relevant features of a target behaviour or skill and offer examples of appropriate responding. Video modelling appears on the same list of 28 practices.

The largest randomised trial: ASSSIST2

Randomised controlled trials of social stories are rare. The most important one, ASSSIST2, was a multisite cluster randomised trial in 87 primary schools in Yorkshire and the Humber, England, with 249 autistic children aged 4 to 11. School staff were trained to write and deliver social stories, and schools were randomly allocated to social stories or usual care (Wright et al., 2024).

  • Primary outcome. Teacher-rated social responsiveness (SRS-2) at six months differed by −1.61 points (95% CI −4.18 to 0.96, p = 0.22): slightly in favour of social stories, but not statistically significant.
  • Individual goals. Children in the social stories group met their individual, teacher-assessed behavioural goals more often than controls (difference 0.97, 95% CI 0.21 to 1.73, p = 0.012).
  • Costs. Social stories saved around £191 per child with similar quality of life, and had a 75% probability of being the preferred option at a £20,000 per QALY threshold.
  • Caveats. COVID-19 disrupted delivery, and the authors note that the results carry uncertainty.

The authors concluded that social stories did not change broad social responsiveness, anxiety, depression or parental stress, but may help with specific behavioural goals, are low-cost and are acceptable to parents and staff. That fits the earlier single-case literature remarkably well.

What this means in practice

  • One situation, one goal. Write for a specific moment, such as a haircut or the school gate, with one small, observable goal. Broad goals like "better social skills" are where the evidence is weakest.
  • Read it shortly before. Timing matters: read the story in the days before and again shortly before the situation.
  • Keep it short and literal. Describe more than you direct, and avoid promises.
  • Involve the child and check understanding. Let the child read parts, point to pictures and answer simple questions.
  • Combine it with practice. A story prepares; small, real steps do the changing.
  • Watch the individual child. Responses vary a lot. If a story is not helping after a few weeks, change the goal or the approach.

How Cala Luna uses this research

Our free social stories follow these findings: each covers one moment, is short enough to read in two minutes, describes more than it directs, and ends with a small calm plan to practise. Our personalised cartoons add a second evidence-based practice, video modelling: children watch a model succeed at the situation before trying it. Video modelling has been shown to help children learn skills faster than live modelling in some studies (Charlop-Christy et al., 2000), with gains that are maintained and transfer across people and settings (Bellini & Akullian, 2007). Early comparisons of animated and filmed models suggest neither is clearly better; in one study of eight autistic children, three learned more from the animated video, two from the human video and three showed little difference (Bloh et al., 2025).

We are careful about what this does and does not mean. These studies tested social stories and video modelling, not our cartoons. Our approach is research-based, and we have not run a trial of it.

Limits of the evidence

  • Most studies involve autistic children; there is much less research with anxious children or children with ADHD.
  • Samples are small, and studies with positive results are more likely to be published.
  • Many studies combine social stories with other supports, making their own effect hard to separate.
  • Outcomes are often rated by teachers who know which children received the stories.

References

  1. Gray, C. A., & Garand, J. D. (1993). Social stories: Improving responses of students with autism with accurate social information. Focus on Autistic Behavior, 8(1), 1–10.
  2. Reynhout, G., & Carter, M. (2006). Social Stories™ for children with disabilities. Journal of Autism and Developmental Disorders, 36(4), 445–469. doi.org/10.1007/s10803-006-0086-1
  3. Kokina, A., & Kern, L. (2010). Social Story™ interventions for students with autism spectrum disorders: A meta-analysis. Journal of Autism and Developmental Disorders, 40(7), 812–826. doi.org/10.1007/s10803-009-0931-0
  4. Qi, C. H., Barton, E. E., Collier, M., Lin, Y.-L., & Montoya, C. (2018). A systematic review of effects of social stories interventions for individuals with autism spectrum disorder. Focus on Autism and Other Developmental Disabilities, 33(1), 25–34. doi.org/10.1177/1088357615613516
  5. Wong, C., Odom, S. L., Hume, K. A., et al. (2015). Evidence-based practices for children, youth, and young adults with autism spectrum disorder: A comprehensive review. Journal of Autism and Developmental Disorders, 45(7), 1951–1966. doi.org/10.1007/s10803-014-2351-z
  6. Steinbrenner, J. R., Hume, K., Odom, S. L., et al. (2020). Evidence-based practices for children, youth, and young adults with autism. National Clearinghouse on Autism Evidence and Practice, University of North Carolina. ncaep.fpg.unc.edu/
  7. Wright, B., Bell, K. J., Blackwell, J. E., et al. (2024). Impact of Social Stories on social and emotional health of autism spectrum primary school children: the ASSSIST2 RCT with economic evaluation. Health Technology Assessment, 28(39). doi.org/10.3310/JBTM8017
  8. Charlop-Christy, M. H., Le, L., & Freeman, K. A. (2000). A comparison of video modeling with in vivo modeling for teaching children with autism. Journal of Autism and Developmental Disorders, 30(6), 537–552. doi.org/10.1023/A:1005635326276
  9. Bellini, S., & Akullian, J. (2007). A meta-analysis of video modeling and video self-modeling interventions for children and adolescents with autism spectrum disorders. Exceptional Children, 73(3), 264–287. doi.org/10.1177/001440290707300301
  10. Bloh, C., Bacon, L., Begel, B., Madara, K., & Koller, B. (2025). Comparing human video modeling to animated video modeling for learners with autism. The Analysis of Verbal Behavior, 41(2), 262–279. doi.org/10.1007/s40616-025-00224-y

See every study we draw on, with its design and limits, in our research library. Spotted an error or a study we should add? Email hello@calalunakids.com.