A longitudinal study of the effects of internet use on subjective well-being

ABSTRACT

This study examined how internet use is related to subjective well-being, using longitudinal data from 19 nations with representative online samples stratified for age, gender, and region (N = 7122, 51.43% women, M_age= 45.26). Life satisfaction and anxiety served as indices of subjective well-being at time 1 (t1) and then six months later (t2). Frequency of internet use (hours online per day) at t1 correlated with lower life satisfaction, r = – .06, and more anxiety, r = .13 at t2. However, after imposing multivariate controls, frequency of internet use (t1) was no longer associated with lower subjective well-being (t2). Frequency of social contact by internet and use of internet for following rumors (t1) predicted higher anxiety (t2). Higher levels of direct (face-to-face plus phone) social contact (t1) predicted greater life satisfaction (t2). In multivariate analyses, all effect sizes were small. Society-level individualism-collectivism or indulgence-restraint did not show a direct effect on outcomes nor moderate individual-level associations. Results are discussed in the framework of the internet as a displacement of social contact versus a replacement of deficits in direct contact; and as a source of positive and negative information.

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1. The scales, while not highly reliable in line with recommendations (e.g., Hu & Bentler, 1999), are adequate for the purposes of the research, and better fit may not be possible due to the use of different languages used to formulate the question items.

2. We tested across nation’s measurement invariance of the Life satisfaction and Anxiety scales. Based on multi-sample confirmatory factor analysis, configural equivalence (factor structure is the same across groups in a multi-group confirmatory factor analysis) and metric equivalence (factor loadings are similar across groups) using cultural region as multi-group reference were tested. However, goodness of fit indices was not satisfactory, and measurement invariance was not confirmed.

3. In any case, analyses carried out with these variables showed similar results in the Individualism-Collectivism effects reported here.

4. Effect sizes are considered according to Cohen’s (1992) guidelines.

5. In the online Appendix are the models for each nation separately.

6. Factor analysis was carried out in waves one and two, and a one factor solution was rejected, excluding the presence of a common method bias. We used general means of all questions as a covariate to correct for responses bias and results were similar. These analyses are not included.

7. Given our limited sample size at level 2, we had to perform the analyses separately for this new measure. Complementary analyses using other dimensions, like power distance did not find significant effects.

Additional information

Funding

This research was supported by Grant FA2386-15-1-0003 from the Asian Office of Aerospace Research and Development. It was also partially supported by the grant PSI2017-84145-P from the Spanish Ministry of Economy Industry and Competitiveness (MINECO) to the first author.Parts of the data (time 1 well-being and social media variables) were used, but analyzed in a completely different manner, in a manuscript currently under review.