Reliability of seven measures of social intelligence in a sample of adolescents with mental retardation.

The team gave seven social-intelligence tests to 61 teens with mild or moderate intellectual disability. They checked if each test gave the same score when the same kid took it twice (test-retest), if two raters agreed (inter-rater), and if the items hung together (internal consistency).

The teens lived in residential schools in the Midwest. All tests were short questionnaires or role-play stories that asked things like 'What should you do if someone looks sad?'

Every test met the usual reliability bar. Internal consistency alphas ran .66–.90. Test-retest correlations were .76–.98. Inter-rater agreement hit .80–.98.

In plain words: the tests stay steady, different staff score them alike, and kids answer the same way a month later.

Kaiser et al. (2022) tried the popular SDQ in a similar teen-ID group and found weak reliability. The difference: the 1992 tests were built for ID, while the SDQ was built for typical kids and then dropped into ID. Same age, same label, but the tool origin matters.

Smith et al. (2010) later showed two self-report scales for social desirability also hold up in adults with ID. Together the three papers build a 30-year line of evidence that social-cognition tools can be reliable if they are written for the population.

Vassos et al. (2023) moved the reliability story to emotional development in adults. Same high alphas, different construct—showing the 1992 pattern repeats across age and domain.

If you need a quick, solid snapshot of social cognition for an IEP or treatment plan, these seven tests give you data you can trust. Pick one that matches your setting—role-play for clinic, questionnaires for classroom—and track progress every few months without second-guessing the numbers.