Age Invariance of the Cannabis Abuse Screening Test in a Probabilistic Sample of Cannabis Users

The Cannabis abuse screening test (CAST) is a short test used for screening those with problematic cannabis use. Recently, its invariance toward age was tested in an Israeli sample of cannabis users, but this study had some sample and methodological limitations: it was conducted in a volunteer sample aged 18-40 and considered the CAST items as continuous variables, although they are based on 5-point Likert scales. We thus tested the CAST invariance toward age (15-24, 25-34, and 35-64 years old) using a French probabilistic sample of 1,351 past-year cannabis users aged 15-64 and using appropriate methods for categorical items and survey weights. Factors retained (non-recreational use and problems) were the same as those mentioned in previous studies. Scalar invariance held for the "problems" factor but only partial scalar invariance was supported for the "non-recreational use" factor. Caution is thus needed when the CAST score is compared across age groups.