RE: The science went over people's heads11 Oct 2020 17:30
The statistically significant phrase simply means that between the two samples there is a lot of overlap and they can't be confidently said to be from different populations. However... As the sample size decreased the gap has to get bigger before you can be confident. So on the larger combined study there is a difference as the sample size is bog enough to prove it, on the sub populations it's less certain that there is a difference. This is also significantly impacted by the spread of the outcomes for each sample, if the spread is large then the overlap is large and the gap needs to be bigger. If, for instance, the outcomes from the placebo had a wide spread, as might be expected from a disease who's symptoms vary from day to day, then it's could make a positive outcome seem to be less positive.