e. Empirical social scientists often do econometric regression analysis on real-world data to ascertain the direction and magnitude of people’s responses to different policies. As computational analysis has become less costly, such researchers are now able to run literally tens of thousands of different regressions, using combinations of different variables and empirical specifications, whereas in the past they have had to limit themselves to a few regressions.

Suppose that researchers have prior beliefs about what an empirical investigation might show.

How might you view statistically significant empirical results reported in research papers more skeptically as a result of knowing about confirmation bias?