How should the effectiveness of Social Stories to modify the behaviour of children on the autistic spectrum be tested? Lessons from the literature.

Jonathan and colleagues read every Social Stories paper they could find. They asked one question: do these studies prove the stories work?

The team looked at how each study picked kids, set up the story, and measured change. They did not run new kids; they graded the old work.

Most papers used tiny groups and no control. Some had no clear goal or way to score behavior.

Because of these holes, the review says we cannot yet trust Social Stories. It calls for RCTs or strong single-case designs before teachers adopt the tool.

Lord et al. (2005) said the same thing one year earlier about all autism psychosocial work. Jonathan et al. narrow that warning to Social Stories only.

Tromans et al. (2018) later counted 529 autism RCTs and found most are still small. The 2006 plea for bigger, tighter trials is still alive.

Shaked et al. (2004) showed that poor matching of kids skews meta-analyses. Jonathan et al. add that poor design plus poor matching makes Social Stories evidence shaky.

Mammarella et al. (2022) looked at school FBIs and found only 7 of 55 studies checked real-world fit. Jonathan’s group wants the same ecological check for Social Stories.

If you write or approve Social Stories, treat them as experimental until an RCT backs them. Use single-case methods with clear baselines, blind scoring, and social-validity checks. Share your manual so the next team can replicate. This stance protects kids from wasted time and guards your program from weak evidence claims.