Yeah, some things are good, for some people. The market is up? Whee – but that doesn’t do much good for the majority of the population who have less than $1000 in the bank and aren’t buying a lot of stocks. In fact, it means that the disparity in nominal wealth between the stock-holding class and the wage-earning class is getting even larger. Heavily massaged unemployment numbers are lower now, true. That doesn’t mean that things are getting better for the “lower” classes. They aren’t. Pointing to these numbers to say that everything is just peachy economically denies the experience of a huge fraction of Americans.

You asked about a particular sub-group, white males. Even after the earlier discussion of the division fallacy, you still are insisting that the subgroup of white males is totally representative of the total electorate. Looking at the exit polling, this is simply false; white males are unrepresentative of the larger group of democratic primary voters. There is simple nothing inconsistent with the sub-group of white males being slightly bi-modal white the larger electorate being unimodal.

Secondly, as I noted before, your question is based upon a linear model running from very republican at one end to very progressive at the other, with Clinton being a “republican lite”, that is, on the scale between republicans on one side and Sanders on the other. Then you gave some numbers for the three groups you identify (republicans, Clinton, Sanders). The math is not difficult: you are assuming the three groups, as referenced to white males in the entire electorate, are substantially republicans: 55%, Clinton: 20%, Sanders: 25%. Notice that this distribution – which is yours, not mine – is in fact bi-modal; the center group is the smallest.

Now you have several choices. You can reject your linear model, and assume that either Clinton is more progressive than Sanders, or assume that the relationship between the three groups in not a linear one as you supposed in you original comment. This would mean that your original question is ill-framed since it is based on a false linear model.

Or you can reject your suggested data, and say that you made a mistake in the numbers you presented and that Sanders is not really more popular amount white male voters than Clinton is.

Or you can accept that the white male subgroup is non-representative and there is nothing unreasonable in taking it to be somewhat bi-modal.

Now what you cannot do, at least reasonably and rationally, is to present a bi-modal distribution and then reject an explanation of this distribution because it is bi-modal. You presented a bi-modal distribution, and asked for an explanation. You may be able to reasonably reject my explanation, but not because it accurately reproduced the bi-modal characteristic of the distribution you yourself provided.

So you are insisting that (1) the white male subgroup must mirror the characteristics of the larger group from which it is taken, which is a logical fallacy, (2) that the white male subgroup is in a totally representative sample of the larger electorate, which is emprically false, and (3) that an explanation of a bi-modal distribution cannot be accepted it if explains a bi-modal distribution, which is simple bizarre.

SInce you are obviously immune to both facts and reason, I suppose there is no point in continuing this conversation.

Nonsense. Go read the NYTimes research on Trump supporters.

The center is holding just fine in the USA. Employment is up, the market is up, gay people are having delightful weddings, poor people are getting health insurance, the unfortunate violence we are engaged in internationally is far. far less destructive than it was ten years ago, and there’s a really good chance that SCOTUS will once again become an institution that will protect citizen’s rights rather than diminish them. Thanks to the efforts of moderates, against vicious opposition.

You bring up fallacy of division– look up Nirvana fallacy.

Also look up “if if sounds like a Republican troll using every opportunity to slip in negative Republican talking points about Hillary, maybe it is not a duck.”

If there is a bimodal distribution as you describe, then each party would have been selecting extremist nominees in the past. But that is not what the data tells us– each party has traditionally selected moderates (or those who pretend to be moderate, in some R cases.)

So, you have to come up with some better argument. A “normal distribution” where moderation predominates is a well supported model.

Let’s look at a different ethnic group – African-Americans. We know in the general election the R:D split for them is somewhere in the neighborhood of 10:90. But in the democratic primaries so far the C:S split is around 83:17. But ‘if H is “Republican Lite” and B is “True Progressive”, shouldn’t it be the other way around?”

Your comment seems to divide the electorate up along a line from uber-republican at one end to uber-progressive on the other, with Clinton being slightly more progressive than the middle center and Sanders being on the far progressive end. You seem to discount the possibility that the white male segment is somewhat polarized with more “bunched up” at the two ends, where the republicans and Sanders are, while, for African-Americans, they tend to be more strongly bunched up around slightly left of the middle, where Clinton is. There is no reason to assume the various groups are evenly distributed along your republican-progressive scale.

If you go back and read #6, you will see that I said:

“Can anyone explain that? If H is “Republican Lite” and B is “True Progressive”, shouldn’t it be the other way around?”

You appear to be ignoring that second sentence.

I think you are missing the continuity of the discussion I was having with Jesse prior to that, and that’s what #18 was intended to clarify.

If we thought that the voting results were completely random, why would Greg make his model, or Jesse and I be arguing back and forth?