Abstracts of the 2nd International Conference on Behavioral Addictions — March 16–18, 2015, Budapest, Hungary

Abstract

Background and Aims: A relatively large body of studies reported bivariate associations between sociodemographic variables (e.g., age), intensity of internet usage (e.g., hours spent online) and internet addiction (IA). However, only few actually employed more robust statistical approaches (e.g., path analysis) to understand these complex relationships and their possible causal pathways. In the present study, a full structural equation modelling (SEM) model using year of first use of the internet and hours of internet usage per week as mediators of the relationship between age and IA was devised. Methods: A total of 1,105 internet users (Mage= 33; S.D. = 12.31), 36.7% (n = 405) from the US, 30.4% (n = 336) from India, 24.6% (n = 272) from the UK, and 8.3% (n = 92) from other countries were recruited online. In addition to collecting sociodemographic information, participants filled out a newly developed brief measure of IA based on the DSM-5 criteria. Results: After analyzing the full SEM model, partial mediation presented the best overall fit for the data (Satorra-Bentler scaled Δχ2 = 114,62; Δdf = 1; p < .0001), with year of first use of the internet and the hours per week spent surfing mediating the relationship between age and IA. Furthermore, the present model accounted for 18% of the variability in IA. Conclusions: Although diagnosing IA on grounds of non-standardized criteria is not correct from a methodological standpoint, the variables here analyzed may play an important role in the development of IA.