Not if the samples are skewed. For example, the people who get the care are from stable environments with financial means. After universal childcare is implemented, we start measuring these things in the broader population that has fewer access to resources generally.
The assumption here is that only people with means got care and were surveyed. I am not sure that this is the case. Moreover, you can correct for those factors, and, I assume, any statistician worth their salt are.
Given the reproducibility crisis, particularly in the social sciences, I wouldn’t put too much weight into the skill or honesty of the people doing that work (and statisticians they are not - more like people with a humanities background who take some statistics courses and then do numerology)
