Evaluating the Psychometric Properties and Measurement Invariance of the Addiction-Engagement Questionnaire Using a Longitudinal Sample

To date, number of scales have been used to study online gaming addiction. Aside from their quantity, their uncertain psychometric properties are among the issues that this field of research faces. This study aims at a more detailed assessment of the frequently used Addiction-Engagement Questionnaire (Charlton & Danforth, 2010) that can uniquely discriminate pathological and non-pathological intensive gaming. An online questionnaire was administered thrice during a period of one year to a sample of online gamers (N = 5080). A series of factor analysis models was used to re-evaluate the original suggested models and formulate new models. Measurement invariance of the final selected model was evaluated in terms of preferred genres (MMORPG and MOBA), age groups, and waves of data collection. The original model did not exhibit a solid fit to data. Few changes were suggested as remedies to improve model fit. The revised model keeps the original two-factor solution, however, we found little support for the so-called peripheral components of addiction with the exception of tolerance. The revised model was found to be invariant over the selected groups. With only few modifications, the Addiction-Engagement Questionnaire was found to be usable for survey-type research purposes, showing a mid-high internal consistency and test-retest reliability.