What's the matter with What's The Matter With Kansas?

 Yes, I stole that title, but if I can no longer remember where I got it, does that diminish the gravity of the crime?  Decide for yourself.  I could, of course, google it.  I am on a computer right now, but I am lazy, and I would rather simply bloviate a bit.  Anywho, once upon a time, before anyone started making fun of the title, a gentleman named Thomas Frank wrote a very bad book called What's The Matter With Kansas?  Quality and fame are rarely related, unless inversely, and so it was with Frank's book.  Here is the precis.  Republicans use "cultural" issues, like abortion, to trick voters into voting against their economic interests, which is why a state like Kansas votes GOP, even though it "should" vote Dem.

The book was wrong in oh so many ways, and in fact, you could use it to teach methodology as "don't do this."  The most flagrant problem was what we call "ecological inference," which is a highfalutin term for a subtle, but important concept.  If you have a hypothesis about one level of aggregation, you cannot test that hypothesis with data aggregated to a higher level.  If you have a hypothesis about individuals, for example, you cannot test that hypothesis with data about states.

There is a long, distinguished and infamous history of this error.  Consider Emile Durkheim.  One of his most famous studies, lo' those many years ago, was about suicide rates.  He hypothesized about religion and anomy, but he did not have individual-level data.  He had country-level data.  Oopsies!  Ecological inference!  Ecological inference continues in many fields, including voting behavior, errors abound, and if I felt like writing a history of the ecological fallacy, it wouldn't be a blog post, it would be a fucking book.  And nobody would read it.  Kind of like this blog!  But at least I can bang out these posts while sipping my morning coffee and swingin' to some groovy jazz.  (Horace Silver this morning, if anyone cared.)

So Frank.  He had a claim about individual voters, but he just looked at states, and blathered.  Bullshit, all bullshit.

Fortunately, a real scholar-- Andrew Gelman-- came along and debunked the whole thing, but alas, it was a bunch of math, so nobody read it, even though his title was even better:  Red State, Blue State, Rich State, Poor State.  Here's the deal.  In wealthier states, like Connecticut, the association between income and vote choice is relatively weak, but in poor states, like Mississippi (and Kansas), the association is actually much stronger.  It is simply that everyone has a higher baseline probability of voting GOP.

Think back to algebra, and the equation for a line.  In regular, old algebra, they give it to you as y = mx + b.  In "linear regression," we write it as y = a + bx + e, or something like that, where e is the error term.  Same thing.  In linear regression, the b term is the slope, and the a term is the y intercept, or, the constant.  You need both.  Both matter.  For voters in poor states, there is a higher baseline probability of voting GOP (the constant), but the probability of voting GOP is more responsive to income (the slope).  Yes, in those poor states, being poor makes one more likely to vote Dem.

Translation:  Thomas Frank was completely wrong, and absolutely full of shit.  This is what happens when you look at ecological data.  Do you see what happens, Larry?

That said, Kansas is not exactly Dem-friendly territory.

So Dobbs.  Why did Dobbs happen?  Because conservative activists have been working towards it through court appointments for decades.

Money is bullshit.  It does nothing.

There is no boogeyman of an interest group with a three-letter acronym to demonize.  Those, too, are bullshit.  They do nothing.

Your lefty-bullshit conspiracy theories about how and why you lose?  Those are bullshit.

Dobbs was a very unpopular ruling, though.  Even among a lot of Republicans.  It happened because of pluralism.  A committed minority, acting strategically, won.  Robert Dahl.

But in Kansas?  They put it up for a vote.

You'll notice that we aren't seeing Big Lie freakouts about this one.  They'll reorganize, re-strategize, and get there.  The committed minority, as small as the total-ban minority is, cannot win via popular vote.  They can take over the Supreme Court, through the GOP, and again, that has fuckall to do with money or three-letter-acronym interest groups.  But they can't win referenda.  So, next!

But I'm getting off track here.  My basic point is that if you bought into that Thomas Frank bullshit, you would really think that the vote had been rigged!  Why isn't Thomas Frank calling for the Cyberninjas to investigate?!  Oh, right.