1) asking five years apart only confirms if the bias is stable/changing. It doesn’t correct anything.

2) asking more people gives you an answer closer to the ‘true answer’ (i.e. the mean answer if you asked ‘everyone’). It corrects for sample error, not directional bias.