Among control variables in the first step of the regression
(Model 1 in Table 5), only the preference of online communication
had a positive association with approaches to online friendship:
Higher preference for online communication led to higher preference
for online friendship. Age, gender, and hours spent online
were not associated with approach to friendship. In the next step
(Model 2 in Table 5), we added the Internet addiction score and
found that, controlling for other variables, it was significantly
associated with approach to friendship, accounting for 3.9% additional
variance in approach-to-friendship variables.