Neat finding, and a good reminder that there's a world of difference between "no detected effect" and a precisely estimated zero.
Reposted from John Holbein
Does more schooling make people more likely to vote?

Yes!

Many natural experiments have been underpowered. Scholars have relied on a lack of statistical significance incorrectly as evidence for no effect.

When you pool estimates, the effects are consistent with education increasing voting.

Comments