Digital divide is one of the most ill-defined concepts in Internet-related social science investigations. Based on the review of literature of digital divide on conceptualization and operationalization of the concept, it was conceptualized that an empirically-tested digital divide index should encompass three key constitutive components--levels of analysis, inequality types and types of ICT involved. The inequality divide can be measured based on the established approach in assessing the distributional inequality, i.e., the Gini coefficient. Access divide and usage divide (includes online time, information acquisition, email use and online games playing) indexes were constructed from the survey databank.

It was found that the Internet access divide is the largest one while information acquisition is the least. The greatest inequality is found in online games playing among the four usage divides. In addition, a closer examination on the Internet access divide shows that the divide drops from 0.3941 in 2002 to 0.2795 in 2007 whereas no clear pattern is found among the four Internet usage divides. The findings might indicate that the Internet access divide and Internet usage divide are not parallel in terms of the magnitude of divide. In addition, the evolutive pattern of Internet access divide, in some sense, might mirror the socio-economic status (Jin & Cheong, 2008).