Fig 1.
Flow diagram of sample selection.
Separately our two teams selected journals, then used all empirical articles in those journals as treated observations.
Fig 2.
Citations per article appear in the top row, for articles published before treated journals’ adoption of a data-posting policy.
The bottom row shows the difference in citations per article following the policy change. Error bars show 95% confidence intervals.
Table 1.
Effect of policy switch, treating years after publication simultaneously.
Fig 3.
Percentage of articles with posted data appear in the top row.
The middle row shows the cumulative citation advantage for articles with posted data as of November 2017, and the bottom row shows citations in year five post-publication. AER = American Economic Review, QJE = Quarterly Journal of Economics, AJPS = American Journal of Political Science, and APSR = American Political Science Review.
Table 2.
OLS estimates of citations and data sharing.
Table 3.
2SLS estimate of data sharing and citations.
Table 4.
Tests of the 2SLS exclusion restriction.