Internet Addiction Disorder： Categories or Dimensions？

马文超,边玉芳,骆方

DOI: https://doi.org/10.16187/j.cnki.issn1001-4918.2012.05.015

2012-01-01

Abstract:Internet Addiction Disorder （IAD） is getting increasing concerns because it brings about many negative influences from both physical and mental aspects. However, an important but open question is whether the internal structure of IAD is dimensional or categorical. A false assumption of nature of latent variable is likely to cause biased results. This paper focused on the latent structure of IAD-categories, dimension or combination. We analyzed 2511 Chinese middle school students＇ responses in Internet Addiction Test （Young, 1996） using the Rasch Model, Latent Class Model （LCM） and Mixed Rasch Model （MRM）. To determine the appropriate number of latent classes, LCM and MRM with different numbers of classes were fitted to the data. The results show that MRM with two latent classes （MRM-2L） can fit data in terms of Q index （Rost ＆von Davier, 1994） and is the best-fitting model according to information criteria （BIC and CAIC）. Thus, the data of IAD should be described by MRM-2L allowing for qualitative differences between two groups （difference in kind） as well as quantitative differences in each group （difference in degree）. Furthermore, the features of classifiCation using MRM are discussed and comparisons are made with existing classification using cut-off raw scores. Finally, based on the new classification method, we provide some practical recommendations for assessment and treatment of IAD.