Alternatively, when you pose questions using push-poll style phrasing, you get push-poll style skewed responses that are not necessarily indicative of how many

Alternatively, when you pose questions using push-poll style phrasing, you get push-poll style skewed responses that are not necessarily indicative of how many people would really support specific policies being enacted. That's why push polls are generally considered bad practice and a tool of people more interested in getting their preferred answer rather than discovering what people really think. This is the somewhat rare case of a push poll designed to elicit a response those running the poll think is bad, so they can complain about it, but the principle is the same. "Would you support X to prevent [Y that most people think is bad]?" is a garbage question, because saying "no" immediately suggests the person answering the question doesn't actually object to "Y". You personally may be comfortable implying to a stranger that you don't care about reducing domestic violence, but enough people are weirded out by the implication that they'll answer in a way that allows Cato to rile up their target audience. Of course supporting actual policies can work the same way, but polls aren't really a good way to measure this because there is zero cost in answering a poll compared to actually supporting something like a new law. It's one thing to tell a pollster you support a measure designed to combat domestic violence because you don't want to give the impression you don't care about the issue, it's another to support the measure in reality because now you have to consider all the other implications. Also this is hardly my biggest objection, but Cato predictably labels the Democratic party as "Democrat" in their results chart. In my experience this is remarkably accurate Shibboleth for obnoxious right-wing political actors and fundamentally unserious people.