Note: Each line shows mean life satisfaction (ST016QO1NA, O-10 scale) at each weekday internet-use level for activities pooled within two groups: Group A (social media, gaming, browsing for fun) and Group B (news, content creation, learning, and communication). At each usage level, individual life satisfaction observations are pooled across all activities in the group; shaded bands show 95% confidence intervals. The sample is restricted to respondents with valid responses on all seven weekday ICT items (IC177 Q01-Q07). Two exclusions are then applied: (1) students reporting zero time on all seven weekday activities, and (2) students with missing life satisfaction or gender. Unweighted, These data reflect simple correlations, and we internet use. This logic would suggest that make no causal claims. How might these patterns would lessen, rather than increase, the steepness be more likely to use the internet, and that unhappyof life satisfaction declines associated with have? If happy people are more likely to make regions. There is much in common across regions, some use of the internet, that would shift the with Group A (social media, gaming, and browsing curves, without any obvious link to the rate of for fun) above Group B (communication, news, internet use. If already unhappy people are more learning, and content creation) at low levels of likely to turn to social media, greater unhappiness use, and then falling below at high rates of use. This suggests that social media, gaming, and