Not necessary, because we know that opinion researchers choose their population sample for a national poll on the basis of national census data. Even if there were other ways to get a population sample, those other ways aren’t relevant, because opinion researchers do in fact use national census data.

Not necessary, because even if there are many who would participate in a census if it were made voluntary, we still know that the participation rate in a voluntary census would be much lower, and the respondents can still be less representative of actual national demographics.

Necessary, because if it were not true — if the group of people who would participate in a voluntary national census would NOT differ in demographic characteristics from the group who would participate in a mandatory census — then we have no reason to think the national census data would be less accurate in representing national demographics. In order to conclude that polls based on a voluntary census would be less accurate, the author must assume there would be some difference in the demographic results from gathering census data from volunteers than from those mandated to participate.

Not necessary, because even if the group of people who refuse to participate in an opinion poll have DIFFERENT demographic characteristics as the group who doesn’t currently participate in the national census, making the national census voluntary can still result in a less accurate representation of demographics compared to a mandatory national census.

Not necessary, because even if the percent of people who participate in the mandatory census changes a lot from year to year, we still know that if the census were made voluntary, fewer people would participate. And that lower participation can still result in a set of demographic characteristics for the national population that is less accurate.