| dc.description.abstract | This Collective Health doctoral thesis of the Collective Health Graduate Program of the Universidade do Vale do Rio dos Sinos - UNISINOS aimed to evaluate the psychometric properties of the short version of the Internet Addiction Scale (IAT), including its internal consistency, construct validity and test-retest reliability in a sample of health students from a Brazilian university. In addition, the prevalence of excessive internet use (IA) and its associated factors in this population of college students were also assessed. This study is part of a larger cross-sectional school-based study (university), consisting of a representative sample of university students in the health area conducted in the municipalities of Rio Verde (southwestern region), Aparecida de Goiânia (metropolitan region of Goiânia) and Goianésia (northern region) located in the state of Goiás in the year 2018. The thesis is composed of 2 papers. In the first paper the construct validation of the scale was performed through factor analysis, through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) of the 12 items of the IAT and for the evaluation of the internal consistency of the factorial model Cronbach's α alpha coefficient was used. The original scale has 20 items and assesses attitudes about the extent of Internet use. From a census sample of 1,582 medical students from the Brazilian Midwest, two random samples of 350 college students were selected for exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), respectively. Another 347 students participated in the test-retest stability evaluation. The two-factor solution with twelve items found in the EFA proved satisfactory, obtaining Cronbach's α of F1=0.84 and F2=0.81. The final model showed good fit (RMSEA=0.089; CFI=0.955; SRMR=0.049) and there was evidence of weak to moderate stability of the short scale. In paper 2 we assessed overuse using the short version of the Internet Addiction Test (AIT) Scale. Associated factors included sociodemographic, socioeconomic, academic, behavioral, and psychosocial characteristics. Poisson regression with robust variance was used to analyze the data. We observed, that in both female and male students, as age increases, excessive internet use decreases (IRR 1.67, 95% CI:1.30 - 2.14). Non-white male students (RP 0.96, 95% CI: 0.50-1.86), who lived alone (RP 0.77, 95% CI: 0.55-1.07), from economic class B (RP 0.91, 95% CI: 0.67-1.22), who failed (RP 0.96, 95% CI: 0.50-1.86) and who had very severe distress (RP 4.08, 95% CI: 2.76-6.02), used the internet more excessively. In female college students, those who were single (PR 1.39, 95% CI: 0.98-1.98), who were in medical school (PR 1.68, 95% CI: 1.25-2.26), who failed (PR 2.04, 95% CI:1.50-2.79), who were sedentary (PR 1.13, 95% CI: 0.93-1.36), who had made harmful use of alcohol (PR 3.12, 95% CI: 2.41-4.04), used the internet excessively. It was also observed that students who had very severe distress (RP 4.96, 95% CI: 3.15-7.82) and low support (RP 1.30, 95% CI: 0.95-1.78) used the web more excessively. Conclusions: The IAT short version showed evidence of validity and reliability, allowing the assessment of Internet use pattern in college adults with a reduced number of items. Moreover, we found a high prevalence of internet dependence among college students (23% in the overall sample) with significant association with younger age group, medical school, very severe distress measured by the K10 scale, and with low social support. When stratified by gender the associations were different for men and women. | en |
