Exercise

Regressions with small coefficients and small confidence intervals

Does a regression coefficient of 0 with a small confidence interval / small standard errors imply that the outcome and treatment variables simultaneously cause each other?

Instructions

50 XP

Possible answers

Yes, it shows the variables cancel each other out.

No, it just shows there's no relationship between these variables in the data.

It shows "misspecification" and we can't learn anything