To evaluate the models, we utilize the AIC and BIC statistics for the full models. The AIC and BIC provide goodness-of-fit estimates used to compare between nonnested models. To interpret these estimates, lower values are better, and a score difference of 10 or more between models reflects a strong model fit improvement. Summary statistics are presented in Table 6. Overall, three of the proposed models were significant. From this research, we have identified four variables associated with having a friends-only profile in Facebook. These variables are gender, friend network size, weak-tie expectancy violations, and conversant privacy practices.