> The current paper used an existing data set of middle-aged adults for the first study, who completed survey questionnaires and provided blood samples the rese

> The current paper used an existing data set of middle-aged adults for the first study, who completed survey questionnaires and provided blood samples the researchers analyzed for CRP. The authors collected their own data for the second and third studies using similar methods for college students. Was the hypothesis pre-registered, or did they just check the survey results against a bunch of markers to see what popped? While the conclusion certainly seems plausible, it's hard to imagine that something as diffuse and hard to measure as social media use could be reliably correlated to a specific biomarker...