2 ms·
Problem being of estimating NNT to see the effect. If the effect size is small you will need huge numbers to detect it Plus unfiltered air is elsewhere, as a c
by AstralStorm 7y ago
Problem being of estimating NNT to see the effect. If the effect size is small you will need huge numbers to detect it
Plus unfiltered air is elsewhere, as a confounder.
- dhshahsndeisjwn 7y agoThis is my hobby horse! A confounder would be an unobserved (unmeasured) variable that influences both the allocation to treatment and the outcome. In this case unfiltered air elsewhere is contaminating the magnitude of the treatment effect (in a conservative direction) but wouldn't undermine the validity of a properly conducted experiment.