Why_question

I don’t think it’s that bad. Not sure why web bad is an issue but voluntary/self selected can lead to representativeness issues. I think they are trying to get around that by asking for family members info and treating each family member as separate data points. I’m not sure they are doing this but in theory they can also just not count the responder and just count the family members. There are also statistical methods I assume they know about that can be used to increase representativeness for potentially biased samples such as weighting by known parameters etc. if you are interested you can probably email them to ask for specific methods. I think this is an imperfect but decent start and am kind of excited about this: I think this is the first time people are even trying to get real data about this subject

I agree that without data the conversations dont mean that much. Apparently some other people agree too – there is a study going on now trying to collect real data – see link I posted above. Accordingly to the article they are basically just asking about age, gender and whether people got married. What do you think they should be measuring?