Background: The prevalence of social anxiety disorder (SAD) is frequently higher in younger age groups and people with other anxiety or mood disorders; however, it is unclear whether these groups have a higher risk for developing SAD or are simply more likely to endorse diagnostic criteria than other people with similar levels of social anxiety. Explicitly testing the assumption all people respond to structured diagnostic interviews in comparable ways (measurement invariance) is essential in ensuring systematic response biases do not create spurious group differences. This research aims to systematically test whether age, comorbidity status, or types of social fears affect responses to a structured diagnostic interview.
Design and methods: Responses from 1755 participants in a large-scale survey of mental health in Australia screening into the social phobia/SAD section of the Composite International Diagnostic Interview were used. Three series of multigroup confirmatory factor analyses for categorical data systematically tested for increasingly strict levels of measurement invariance.
Results: Overall, patterns of responding to diagnostic criteria were comparable across the groups, supporting assumptions of measurement invariance.
Conclusions: Establishment of invariance supports the interpretation of differences between age, comorbidity status, and types of social situations feared as genuine differences in experience as opposed to measurement biases.
Keywords: age; comorbidity; interaction; measurement invariance; performance; social anxiety disorder.