Experience sampling self-reports of social media use have comparable predictive validity to digital trace measures

Abstract

Abstract Research agrees that self-reported measures of time spent with social media (TSM) show poor convergent validity, because they correlate modestly with equivalent objective digital trace measures. This experience sampling study among 159 adolescents (12,617 self-reports) extends this work by examining the comparative predictive validity of self-reported and digital trace measures of TSM, that is, the extent to which self-reported and digital trace measures of TSM correspond in their effects on self-esteem, well-being, and friendship closeness. Using an N = 1 method of analysis, we investigated the correspondence on a between-person, within-person, and person-specific level. Although our results confirmed the poor convergent validity of self-reported TSM reported earlier, we found that self-reports of TSM had comparable predictive validity to digital trace measures on all three levels. Because comparative predictive validity of self-reported TSM is crucial for investigating social media effects, our results have important implications for future research using self-reported TSM.