Demographics of white Democrats in Mississippi

[SingingAnalyst]

As I'm sure we're all aware, Mississippi whites vote overwhelmingly for the Republicans(+90% of whites in 2016). But what are the demographics of the handful of whites who vote Democratic? Are a significant number holdouts from the Solid South of yore who haven't gotten the memo? Are they demographically similar to white Democrats elsewhere(disproportianately young, college educated, urban, poor, and female)? Are they disproportionately in mixed-race marriages/partnerships?

Up until about 10 years ago there were, but those people have mostly either died off or switched over to voting Republican. White Democrats in MS are mostly actual liberals nowadays, and are disproportionately young and from out of state.

IIRC (and this is half-remembered, possibly outdated information) Mississippi has one of the lowest rates of interracial marriage in the country, though the few who are are probably more likely to vote Democratic.

Probably true. In 2000, MS had the lowest proportion of people identifying as mixed-race (0.6%) of any state.

[SingingAnalyst]

What were the percentages of the white vote that Trump and Hillary got statewide?

Roughly 10% for Clinton, 88% Trump, I believe.

[SingingAnalyst]

Obligatory reminder that in 2012, Romney only won the under-65 crowd in MS by 5 points:



Relatively young (i.e. non-ancient) whites in Mississippi are probably no less Democratic than whites in, say, Georgia. They're not that much of an anomaly. You're still looking at a majority-white electorate (anywhere from 50/50 to 55/45) even in the 18-29 grouping, meaning that anywhere from 20-25% of those whites are voting Democratic.

It's pretty obvious when looking at both the exit polls and general demography of the state by age that older blacks do not really vote at all in Mississippi (the legacy of the times in which they were raised). This is the predominant problem with flipping MS, but once they - and their white cohorts that are voting 95% R - pass on, the state becomes much more competitive.

Do exit polls actually weight by age and race together? Do we know that those age categories in the exit polls are racially representative of what those age categories actually look like? I'm thinking of the result from the Florida exit poll this year that had Trump winning 25-29 year olds by 6%, which seems completely inexplicable unless they got the weighting within the age group wrong.

There's also potentially the issue of past exit polls in states like MS, AL & LA underestimating white Democratic support for some reason: it's something I encountered when doing my 2012 county breakdowns and when just delving into 08/12 exit polling data for these states, and have written about on here in the past.

There were instances where the exit polls did not match the confirmed demographic turnout whatsoever; instances where I maxed out black turnout beyond what was reported and gave 100% of the black vote to Obama and yet there was still a discrepancy that could only be explained by an underestimation of white Democratic support in these areas. I don't necessarily buy the 10% for MS & LA, nor the 15% figure for AL.

When I calculated the White vote for 2016. There were no counties where I got negative Clinton support in Alabama. I initially got two such counties in Mississippi, but both of them have corrections facilities in them (and thus a population that's over 20% incarcerated). Thus the actual electorate is probably Whiter than my model lets on. In Louisiana, I got three such Parishes. Two of them had corrections facilities. I tried to make ad hoc adjustments to correct for this electorate issue.

The other one was La Salle Parish, which was only very slightly negative. I suppose its possible that there's just some statistical noise going on, and the relatively small Black population there is slightly more Republican for some random reason, slightly less likely to turn out, or slightly more inclined to support third parties, (or some combination of all of these) for whatever reason.

I don't know if we can say anything definitive about 2012 from this data, since I suppose it's possible that the White vote swung to Trump in the interim, but I thought I'd mention it.

Overall, I'm inclined to think the exit polls are broadly accurate on this. I remember messing around on Dave's Redistricting, and managing to draw a 85.1% White district in Mississippi where Obama only got 15.4% of the vote in 2008.

10% of MS whites for Clinton may be a bit low. But 12-14% I can believe.

Tishomingo Co, MS (2.6% Black) voted 86-12 Trump; Itawamba Co, MS (6.5% Black) voted 87-11 Trump; and George Co, MS (8.8% Black) voted 88-10 Trump. Thus, in these counties, white support for Clinton was in the single digits. I'm sure whites in places like Hinds, Rankin, Hancock, and Oktibbeha counties were somewhat more Democratic.