Factor structure and a multiple indicators multiple cause model of internet addiction test: the effect of socio-demographic and internet use variables

Abstract

Purpose: This study aimed to identify factor structure of a bilingual version (English and Malay) of Internet Addiction Test, a measure of severity on pathological Internet use (PIU); examine the measurement invariant by adding the socio-demographic and Internet use variables as covariates. Method: Pencil-paper questionnaire was distributed to undergraduates in Univerisiti Teknologi Malaysia (UTM). A total of 1120 valid responses were collected. EFA and CFA were first applied to identify the factor structure of IAT. Multiple indicator multiple cause (MIMIC) model was used to examine the differential item functioning (DIF) and the effect of covariates on the latent factors. Result: A four-factor model demonstrated moderate fit to the data for 17-item of IAT with significant high factor loading. Comparison of MIMIC model with and without DIF indicated that the most meaningful and significant DIF effect were on three items showing direct effect from Chinese to IAT18, IAT19 and IAT20. By controlling the DIF effect, some covariates still present the significant impact on latent factors of IAT, which were female, time spent online, effect on study, years of Internet use experience, general users, other users and Chinese. Conclusions: the risk factors of PIU were male, time spent online, years of Internet use experiences, game users, SNS users, perceived less effect on study. More research is encouraged to examine the racial group difference, including the risk of PIU and item response bias.