Validation and measurement invariance of a modified version of the marijuana craving questionnaire-short form across community and clinical samples

Introduction The Marijuana Craving Questionnaire 12-item Short Form (MCQ-SF) was developed to assess marijuana craving. Few validations exist, and none have been conducted for Spanish-speaking countries. This study aimed to validate a modified version of the MCQ-SF, the Modified Marijuana Craving Questionnaire (MMCQ), in Spanish. Methods A sample of 373 Chilean adults (mean age: 26.10 years, 57.37% male, 71.31% secondary education) who reported Marijuana use completed the MMCQ, Impaired Control Scale (ICS), Cannabis Withdrawal Scale (CWS), and a sociodemographic and substance use characteristics questionnaire. We performed confirmatory factor analysis, using weighted least squares mean and variance adjusted, to test different factor structures. A unifactorial model was selected as the best-fitted solution and tested for measurement invariance by sex (male/female) and sample type (general population/clinical population). Convergent validity was examined through correlations between the MMCQ and the ICS, CWS, and age of onset of marijuana use. Results Four and two-factor models exhibited issues related to discriminant validity, as their factors were highly correlated, whereas a unifactorial model with one correlated residual showed the best fit. Measurement invariance using the selected model was met across sex and type of sample. The MMCQ correlated positively with the perceived control dimension of the ICS and the CWS, and negatively with age of onset. Conclusion The unifactorial MMCQ, with one correlated residual, is a valid scale for research and clinical assessment of marijuana craving that can be used among the Spanish-speaking population.