No, there’s no reason to think the sample size which votes is an unbiased reflection of the total sample. It’s very explicitly self selecting. Interesting area to consider though.

It's one of the most insidious problems with political polling - people who feel alienated and disconnected from society are seemingly less likely to respond to surveys so they inherently go undercounted. Kind of like the dark matter of the political world - we can tell that it's there, because we feel the effects, but we can't observe it directly.