Okay, this was an argument about how many people are paying attention to the presidential race now. My position was that only a relatively small proportion of Democratic voters are paying attention at this stage of the Presidential election process. You took the opposite position: that “a wide group of people must be paying attention.” Your reason for taking this position was that
So, basically, your claim is: ‘large changes in favorability prove that a wide swath of Democratic voters are paying attention now.’
Without getting into it back then (five days ago), I was trying to suggest that your understanding of how polling works as a model of reality wasn’t quite correct. To begin with, your premise was shaky: you claimed “bumps and troughs” in favorability for the various Dem candidates had occurred after a debate (the first set), and that in itself was a questionable interpretation of the data generated after that event. Certainly there were no large changes in favorability after the second set of debates. As DSeid had pointed out:
So the claim that there were large changes in people’s views of the candidates was doubtful in itself. But where the claim really fell apart is the theory that bumps and troughs (even if they had happened) prove that “a significant number of people” are paying attention to the Democratic race now.
It would be perfectly possible to have large swings in the favorability of candidates at the same time that relatively few people are paying attention.
For example: A representative random sample of registered Democrats who had voted in the last election and had watched the debates might show big swings in the favorability numbers for the top candidates—but prove nothing whatsoever about the topic of contention between you and me: ‘what proportion of Democratic voters are paying attention now?’
And the reason that poll would prove nothing about the relative sizes of the groups of voters (the ‘paying attention now’ group as compared with the ‘not paying much attention now’ group) is that the sample included only people who have a collection of traits that make them a minority in August 2019: they are registered; they voted in the last election; AND they watched the debates.
This collection of traits describes the narrow band of Democrats that I posit are paying attention now. It does not describe a “wide group” of people.
So you could see wide swings in favorability in a poll without being able to draw any conclusion whatsoever on the question ‘is it just a small proportion of voters–‘woke Twitter’ and such–who are paying attention now, or is it a wide group of voters paying attention now?’
The crucial thing to know in interpreting a poll is: what’s the nature of the representative random sample polled?
…and I should have said so five days ago, when we first started arguing. I guess I was some combination of short-on-time and lazy. So, anyway, here it is.
I don’t like standing on his damn lawn anyway!