can't win. If your sample isn't diverse enough it isn't representative. It it is representative, its likely too small to get a reliable effect due to the number

can't win. If your sample isn't diverse enough it isn't representative. It it is representative, its likely too small to get a reliable effect due to the number of confounds.