2 ms·
Three 95% confidence intervals not quite overlapping with no effect is plenty evidence to believe this is true unless someone does a bigger study showing otherw
by londons_explore 3y ago
Three 95% confidence intervals not quite overlapping with no effect is plenty evidence to believe this is true unless someone does a bigger study showing otherwise...
- kriops 3y agoCorrect, but if and only if every involved party is completely trustworthy. Which would be naive, and is why independent replication is important.
- refurb 3y agoFor a retrospective analysis? No. And for a study showing an effect of 1% change (maybe) That’s no different than zero. No, it’s not anything close to “believe this is true”.