There has been concern about the effects of high levels of internet use on the mental well-being of young people. This has generally been phrased in terms of a displacement hypothesis, that the extent of internet use and mental well-being are directly proportional. This linear model has been contrasted with a Goldilocks Hypothesis, proposed by Przybylski and Weinstein. This supposes that moderate levels of internet use may be the least harmful, conforming to a curvilinear relationship. Here these hypotheses were tested on a sample of 1140 adolescents (42% boys, 58% girls) aged 12-18 years, in 12 schools from Southern Thailand. We first report levels of internet use, and of cybervictimization, taken as one important aspect of mental well-being. We then assess the relationship of four factors of internet use (frequency, time spent, number of places accessed, number of activities) with (a) being a victim of cyberbullying, and (b) being a frequent victim; taking these as indicators of mental well-being. For (a) there was limited evidence of a Goldilocks effect on two out of four measures. For (b) the evidence did support a Goldilocks effect for all four measures, but these were under-powered analyses and the findings did not reach statistical significance. If substantiated on larger samples, a curvilinear relationship between aspects of internet use and cyberbullying would suggest a 'safe zone' for adolescent internet use, bringing its benefits while minimizing risks of cyberbullying. In the future, similar research should use larger sample sizes or longitudinal measures when exploring nonlinear trends and include other aspects of mental well-being.
Keywords: Goldilocks Hypothesis; ICT use; Thailand; curvilinear trend; cyberbullying; cybervictim; internet; victimization.