Author ORCID Identifier

Date Available


Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation


Agriculture, Food and Environment


Family Sciences

First Advisor

Dr. Alexander T. Vazsonyi


Based on a sample of 1,012 late adolescents and young adults (ages 18-26), the current three interrelated studies tested a series of specific questions and hypotheses focused on understanding Internet Addiction.

Study 1 sought to directly compare psychometric properties of four of the most widely used Internet addiction scales based on citation metrics (Internet Addiction Test, Problematic Internet Use Questionnaire, Generalized Problematic Internet Use Scale 2 and Chen Internet Addiction Scale), and to aggregate the best items across all scales into a new measure using bifactor IRT analyses. The resulting 10-item Aggregated Internet Addiction Scale (AIAS) consists of the best performing items from the original scales, representing each of Griffiths’ six conceptual criteria, and matches each original scale in performance (reliability, test information, relationship with covariates) with considerably lower item count. This offers an important alternative to scholars seeking a reliable and valid measure which is consistent with Griffiths’conceptual work.

Study 2 tested a set of hypotheses based on a behavioral model of Internet addiction in which online activities are used and reinforced as a maladaptive coping strategy to manage stress. If Internet is used to disengage from stressors, it may become associated with the desirable outcome (reducing stress), while not actually solving it; rather, it would exacerbate the underlying problems and therefore create more stress. Stress was found to be positively associated with Internet addiction symptoms; in addition, this effect was mediated by reliance on problem disengagement coping strategies. Self-control was found to moderate the relationship of stress and coping strategies (with exception of problem disengagement); however, it did so in an unexpected direction. At lower levels of self-control the relationship of stress and reliance on engagement strategies becomes more positive, and conversely, more negative for emotional disengagement The moderation, however, had little impact on the indirect effect.

Finally, the third study tested a series of hypotheses based on Uses and Gratifications theory (UGT) and the existing literature on predictors of Internet addiction. Specifically, it compared low self-esteem (LSE), perceived social self-efficacy (PSSE), and self-control (SC) as predictors of Internet addiction. Building on UGT, specific online behaviors were examined as mediators of the relationship between known predictors and

Internet addiction. Social networking was hypothesized to mediate the pathway from LSE and PSSE to Internet addiction; and the use of internet for entertainment was hypothesized to mediate the path from SC to Internet addiction. Using structural equation model SC was found to be the strongest predictor of Internet addiction, both directly and indirectly through entertainment use. After controlling for SC and entertainment online use, social networking and the hypothesized underlying predictors (LSE and PSSE) had no remaining explanatory power for Internet addiction. There three predictors overlapped to some degree, where SC explained the largest amount of unique variance.

Digital Object Identifier (DOI)