and a 3-1 shift in the sub population didn't actually happen because its mathematically impossible. Very one remember the trial would of hit statistical significance. If that 3-1 shift doesn't happen
Yes it could mathematically happen it's called variance. But it's more likely that there was some screw-up in the process. If there was actually no 3-1 difference in the 65 and over population that doesn't change the p value. So that would be one less reason to explain why the p value wasn't statistically significant. In other words, no actual skew is a bad thing.