Objective: The aim of the present study was to evaluate the psychometric properties of the CAGE-Internet questionnaire, adapted from CAGE (a questionnaire for alcohol abuse), to screen individuals for problematic Internet use, among Turkish university students.
Methods: A cross-sectional online self-report survey was conducted at Turgut Ozal University in Ankara. The questionnaire included sections about demographic information, the Addiction Profile Index Internet Addiction Form Screening Version (BAPINT-SV) and CAGEInternet questionnaire. The CAGE is a 4- item, relatively non-confrontational questionnaire for detection of alcoholism. Similarly, CAGE- Internet questionnaire consists of 4items. It takes less than 1 minute to administer, is easy to learn, remember and replicate. The analyses were conducted with 88 participants. A series of factorial structure, convergent validity and internal consistency reliability analyses were conducted.
Results: Adequacy of the sample size was verified using Bartlett’s test of sphericity and the Kaiser-Meyer-Olkin (KMO) measurement of sampling adequacy. Bartlett’s test of sphericity was significant (Chi-Square=85.536, df=6, p<0.001) for the CAGE-Internet questionnaire, and the KMO measure of sampling adequacy was acceptable at 0.597. Only one component on the CAGE-Internet questionnaire reached the criterion of an Eigenvalue greater than one (2.10), and the variance accounted for by this component was 52.53%. The unidimensionality of the scale then was assessed with confirmatory factor analysis (CFA). Estimation of the model produced a good fit (χ2/ df=3.34/2=1.67; root mean square error of approximation RMSEA=0.088, goodness of fit index GFI=0.982, adjusted GFI=0.910, parsimony GFI=0.196, normed fit index NFI=0.962, comparative fit index CFI=0.984, incremental fit index IFI=0.984). All item-component loadings were in the “good” to “excellent” range. Thus, results from the PCA and the CFA suggest that the CAGE-Internet questionnaire assesses a one-dimensional construct. The Pearson product–moment correlation between the CAGE-Internet questionnaire and BAPINT-SV scores was moderate (r=0.62, p<0.001). Internal consistency for the CAGE-Internet questionnaire (coefficient α=0.679) examined by Cronbach’s alpha was also moderate.
Conclusion: These preliminary findings support the CAGE-Internet questionnaire as a reliable and valid problematic Internet usescreening instrument that measures a one-dimensional construct in Turkish university students. Further research is warranted with a higher number of university students, with clinical populations and with high-risk populations such as those in Internet cafes.