Relationship between loneliness and internet addiction: a meta-analysis

Background In the digital age, the Internet has become integrated into all aspects of people's work, study, entertainment, and other activities, leading to a dramatic increase in the frequency of Internet use. However, excessive Internet use has negative effects on the body, psychology, and many other aspects. This study aims to systematically analyze the research findings on the relationship between loneliness and Internet addiction to obtain a more objective, comprehensive effect size. Methods This study employed a comprehensive meta-analysis of empirical research conducted over the past two decades to investigate the relationship between loneliness and Internet addiction, with a focus on the moderating variables influencing this relationship. This meta-analysis adopted a unique approach by categorizing moderating variables into two distinct groups: the objective characteristics of research subjects and the subjective characteristics of researchers. It sheds light on the multifaceted factors that influence the relationship between loneliness and Internet addiction. Results A literature search in web of science yielded 32 independent effect sizes involving 35,623 subjects. Heterogeneity testing indicated that a random effects model was appropriate. A funnel plot and Begg and Mazumdar's rank correlation test revealed no publication bias in this meta-analysis. Following the effect size test, it was evident that loneliness was significantly and positively correlated with Internet addiction (r = 0.291, p < 0.001). The moderating effect analysis showed that objective characteristics significantly affected the relationship. However, subjective characteristics did not affect the relationship. Conclusions The study revealed a moderately positive correlation between loneliness and Internet addiction. Moreover, this correlation's strength was found to be influenced by various factors, including gender, age, grade, and the region of the subjects. However, it was not affected by variables such as the measurement tool, research design, or research year (whether before or after COVID-19).