Why can't I post this?
Yeah, this variance is already covered by the fact that we are counting years and not days, minutes or second of living. If your sample is big enough this would make no difference at all (statistical theory "central value theorem"). If I take 1000 people I will have probably 2 or 3 having birthday every day. So one would always be wrong about the ages if the data is older than 1 day. Is like "The most that can be expected from any model is that it can supply a useful approximation to reality: All models are wrong; some models are useful" https://en.m.wikipedia.org/wiki/All_models_are_wrong And looking for the age spread (10-63) it is very clear to me that 1 year difference would cause no change in the results (assumed that the samples were independent, so no one answering the first survey answered the second), then we need to begin to talk about covariance, a beautiful subject. Look into "Fishers information matrix" if you wish a journey in that direction. The original papers from Fisher are free to download at the University of Adelaide (Australia).
Vous avez souvent des questions comme celle-ci ?
Apprenez de manière plus efficace, gratuitement :
