Social Media and Scientific Careers · data explorer
Each line is the estimated effect of joining in each year before and after entry, relative to the year before joining. Pick a figure, then switch series on and off to compare them.
Each point is the estimated difference between joiners and their comparison group in that year, relative to the year before joining, which is zero by construction. Shaded bands are 95 percent confidence intervals. Points to the left of the dotted line show whether the groups were already drifting apart before anyone joined; points to the right are the estimated effects.
Hover over a panel to read each estimate, its interval, and, for the matched designs, how many entry cohorts identify it. Late event times rest on fewer cohorts: eight years after entry, the headline estimate comes from the six cohorts that joined between 2007 and 2012.
The splits by platform use measure tweeting in 2020–22, years after most people joined. They show who gains, not why.