2 ms·
Yes, it's difficult to run macroeconomic experiments. The question still stands. What degree of evidence is sufficient to not be off-hand-dismissed with 'Corre
by vkou 3y ago
Yes, it's difficult to run macroeconomic experiments.
The question still stands. What degree of evidence is sufficient to not be off-hand-dismissed with 'Correlation does not imply causation'?
- ttyprintk 3y agoSince inferring causality is hard, we look for counterexamples that disprove it. Hence the reliance on randomization and the null hypothesis. A counterexample in social sciences might be that the corresponding relationship is found across methodologies and cultures. This turns a question of causal inference into questioning the power of overturning longitudinal data. Statisticians differ about the philosophy on that.