My attempt at predicting the winner of every remaining state using one metric

[Godhumor]

Last edited Tue Apr 19, 2016, 11:08 PM - Edit history (5)

Hypothesis: Simply by analyzing the percentage of a state's population that is black, I believe you can fairly accurately predict the "winner" of a state in upcoming primaries and caucuses. There will be misses, I am sure, but I'm bored, so let's see how this works out. I'll bump the thread and fill in actual results (With insightful commentary like, "Boy, that one was wrong!&quot as primaries go on...assuming people are interested in my validation or humiliation, as the case may be.

So let's begin with states that have already voted

Your key for numbers below:
State Rank for Black Pop. State % of Pop. that is Black

All numbers from 2010 Census

Bernie Wins
44 NH 1.22%
33 CO 4.28%
31 MN 4.57%
26 OK 7.96%
49 VT 0.87%
29 KS 6.15%
32 NE 4.50%
47 ME 1.03%
16 MI 14.24%
48 ID 0.95%
43 UT 1.27%

Avg Black Pop 4.28%
At +1 Standard Deviation 8.38%

Hillary Wins
40 IA 2.68%
23 NV 9.00%
5 SC 28.48%
6 AL 26.38%
12 AR 15.76%
3 GA 31.4%
25 MA 8.1%
10 TN 16.78%
18 TX 11.91%
9 VA 19.91%
2 LA 32.4%
1 MS 37.30%
11 FL 15.91%
14 IL 14.88%
19 MO 11.49%
7 NC 21.60%
17 OH 12.04%
35 AZ 4.16%

Avg Black Pop 17.79%
At -1 Standard Deviation 7.80%


Prediction Methodology: If a state's black population is less than Bernie's 1 St Dev number, I predict he wins. If it is more than Hillary's 1 St Dev number, I predict she wins.

So my straight up, no commentary predictions (Note: this isn't a prediction of margin of victory, just who comes out on top as the state's winner. I also do not include territories):

34 AK 4.27% Bernie Correct
38 HI 3.08% Bernie Correct
36 WA 3.74% Bernie Correct
30 WI 6.07% Bernie Correct
42 WY 1.29% Bernie Correct
13 NY 15.18% Hillary Correct
21 CT 10.34% Hillary
8 DE 20.95% Hillary
4 MD 30.1% Hillary
20 PA 10.79% Hillary
27 RI 7.5% Bernie
22 IN 9.07% Hillary
37 WV 3.58% Bernie
24 KY 8.2% Hillary
41 OR 2.01% Bernie
28 CA 6.67% Bernie
50 MT 0.67% Bernie
15 NJ 14.46% Hillary
39 NM 2.97% Bernie
46 ND 1.08% Bernie
45 SD 1.14% Bernie
** DC 50.7% Hillary

Now there will be some misses here, because the two data sets overlap in the 2nd Standard Deviation (Mean+2*StDev vs Mean -2*StDev), so the question will become which states and in which direction. That said, misses should favor Hillary as her Standard Deviation is over twice as wide as Bernie's (Wider standard deviation means more variation in the numbers. In this case, wider Deviation means Clinton has been more successful among a wider variation in black population than Bernie).

Now, time will tell how right this "based on one metric" model turns out to be.

[JonLeibowitz]

1. Why do you think black people in all states think about politics the same way?

do white people in all states think about politics in the same way?

Hint: Iowa and NH did not have the same outcome.

[Godhumor]

2. Part of the Standard Deviation

That is accounted for.

[JonLeibowitz]

5. It seems a little too crude of a model, in my estimation.

I mean, you can't correct for not knowing the expected value by increasing the variance (when assuming a Gaussian distribution, anyway, for a one-parameter distribution this is obviously not the case). That's not what those things mean.

I say this as a formal objection to the methods. The outcome may still be correct, but it doesn't make the model right. I still fully expect Clinton to be the nominee (and will oppose her candidacy until that day comes).

[Godhumor]

7. I am making a very simplistic prediction

Approximately 65% of Bernie's wins should be about 8% or less Black population.

Approximately 65% of Hillary's wins should be about 9% or more black population.

And I would argue that black population is by far the one demographic variable that can be considered fairly uniform to date, as Hillary has overwhelmingly won the black vote in every single state, including those that she has lost. Therefore, extrapolating the idea that she does better when the black population is higher is not a stretch nor a huge assumption, by any means.

[oldandhappy]

3. Lots to think about! Thanks

[NowSam]

4. Divide and conquer

This ugly narrative is tired. Dead Horse. Enough.

[JonLeibowitz]

6. Nah, I actually don't mind a thread like this. It's when the projection occurs that it gets ugly

I mean that when they start claiming this means Bernie is himself racist because AAs don't support him. Or when they criticize him for coming from lily-white Vermont (I mean, who do you want to represent VT, Richard Tarrant??)

[1StrongBlackMan]

41. No one, except Bernie supporters on DU, has said ...

Bernie, himself, is a racist. It usually comes in the form of a (deflective) question, e.g., "Are you saying Bernie is a racist?"

And when, people respond No one, except Bernie supporters on DU, has said Bernie, himself, is a racist ... we get to rinse and repeat.

[JonLeibowitz]

43. No. I have seen many posters say that he left diverse Chicago for lily white Vermont.

Why he did that is an exercise left to the reader but the implication is all too clear.

[1StrongBlackMan]

44. Is that not a factual statement ...

and is the "implication" not validated/supported by a dearth of "Civil Rights" protests in "less diverse" Vermont?

But to be clear, a lack of active participation in Civil Rights protests, does not a racist make.

[JonLeibowitz]

45. The implication is that he is racist.

A dearth of civil rights protests in Vermont doesn't support that at all. It's a connection that might sound right, but is ludicrous. You can't draw a line connecting them. Bernie himself has said he enjoyed living away from the city and enjoyed Vermont's countryside. That doesn't say anything about his civil rights commitment.

(Consider a complete hypothetical: does moving from South Africa where one protested to Utopia where there is no racism whatsoever support the idea that the person is a racist?)

That you would consider moving to Vermont implies something about his dedication to civil rights is absurd and makes me think less of your posts. It's a shame.

edit: I may have completely misread your post. I think I must have.

[1StrongBlackMan]

46. Only in the reflectively defensive mind ...

We both agree that a dearth of civil rights protests in Vermont doesn't support that that Bernie is a racist .... But it does signal a change in focus from what you cite ... No? Again, that change in focus does not mean he changed to a racist worldview.

(Consider a complete hypothetical: does moving from South Africa where one protested to Utopia where there is no racism whatsoever support the idea that the person is a racist?)

No. But, then again, it is only you making that straw man argument.

That you would consider moving to Vermont implies something about his dedication to civil rights is absurd and makes me think less of your posts. It's a shame.

I consider his moving to Vermont as suggesting (only) that he changed his focus ... actually, he changed his focus from racial justice to economic justice a couple of years before he decided to move, as evidenced by his frustration with SNCC for focusing on racial justice; rather than, his world view of economic justice. And, even that does not make him a racist ... it just means he de-prioritized racial justice in favor of economic justice.

And, BTW, I really don't care what you think of my posts because, if you were honest with yourself, you would admit that any post that does not cast Bernie as some hero, you would think less of.

[JonLeibowitz]

47. No. There are multiple Clinton supporters that I respect.

And no, I don't think it means he changed his worldview or what he valued. I think he always had his goals. There may not have been many racial justice causes to push in Vermont, but it doesn't mean he didn't care about the issues and fight where needed.

The whole division over race that has been employed in this campaign is very sad. Vermont isn't a place for people to run away from blacks, they're very progressive people (obama 2008, jackson 1988, sanders 1980-present).

[1StrongBlackMan]

48. Vermont may not be a place for people to run away from Black ..

people; but it, certain is not a state where anyone would go to interact with embrace or interact with Black folks, either.

The whole division over race that has been employed in this campaign is very sad.

There has never been a time, in US history when this division over race has not been present ... it's something Black people live with every day and white people can get to be irritated about being reminded of, on the few occasions it enters their consciousness.

[1StrongBlackMan]

20. Empirical observation is not divide and conquer.

[Godhumor]

40. So far, I think the evidence shows this model works n/t

[1StrongBlackMan]

42. Yes ... It does. But, or and, ...

It is a real positive ... It is instructive to Black youth that their vote really does count, no matter what some would tell them.

[Retrograde]

8. What happens if you redo your predictions

using the % non-White population as opposed to % Black? The Western states have relatively low percentages of people identifying as Black, but much larger percentages of people identifying as Hispanic (especially Mexicans of Native American descent) and Asian than much of the country - Whites are close to a minority in California.

Why isn't Arizona in your model?

[Godhumor]

9. I just missed AZ

My bad.

And I chose black because it is the demographics with the clearest difference in support between Hillary and Bernie.

[RandySF]

10. I have to quibble with California.

You don't take into the Latino vote into account or the fact that this has been a Clinton stronghold since the 90's.

[RandySF]

11. Same goes with New Mexico.

Same demographics as AZ.

[Godhumor]

12. Like I said, misses will happen and, due to StDev spread, will favor Clinton

But I'm not invalidating the rest of the predictions by changing out for a couple states.

For the record, I think Hillary wins CA and NM, but that is personal feeling not the the"one metric model".

[Raissa]

13. Math is interesting-

Looking forward to following hits and misses.

If anyone has any charts that highlight other demographic splits that would be interesting too.

[frustrated_lefty]

27. Demographic split by education.

This examines the proportion of delegates awarded to each candidate as a function of the percentage of the population with a college degree in each state. Census data was obtained here: https://en.wikipedia.org/wiki/List_of_U.S._states_by_educational_attainment and appears to be drawn from the 2012 census.



The data is presented as a scatter plot with trend lines for each candidate and basic stats.

Assuming the trend to be correct, the following wins would be expected, with some being too close to determine (TCTD).

Clinton
Wyoming
Pennsylvania
West Virginia
Kentucky

TCTD
North Dakota
South Dakota
Wisconsin
Missouri
New Mexico

Sanders
Montana
Oregon
Connecticut
Maryland
Delaware
New Jersey
District of Columbia
Indiana
New York
Rhode Island
California

There's obviously bound to be miscalls with such an overly-simplistic model, but I'm curious to see how it plays out.

[Donald Ian Rankin]

14. Do you have a scatter plot?

If so, could you post or link to it, please?

If not, generating one seems an obvious way to illustrate and test this, but I'm too lazy to do so myself.

[DemocratSinceBirth]

15. Similar to Alan Abramowitz's model

http://www.centerforpolitics.org/crystalball/articles/a-simple-model-for-predicting-hillary-clintons-vote-in-the-march-15-democratic-primaries/

https://en.wikipedia.org/wiki/Alan_Abramowitz


You should run another model using the entire non white vote as your independent variable.

I still think these models are overly determinative.

[Godhumor]

16. 3 for 3 so far n/t

[GeorgiaPeanuts]

17. Never have I seen a campaign and its supporters so obsessed with making this election about race

[hardboiled]

18. I have never seen a race...

block vote with such a high frequency as the blacks.

[Godhumor]

19. There is a significant demographic divide

A huge one, in fact.

[yardwork]

21. I'm not understanding where the Standard Deviations came from.

Can you explain more about the standard deviations?

Thank you for this post. This is what I used to love about DU.

[obamneycare]

22. I did some similar work, looking at state-level predictor variables

While black population (%) was the strongest single predictor that I looked at, there were other factors that were also significantly correlated -- some almost as strongly as black population (%). It's worth noting that "white population (%)" was not technically among them.

(Note: Higher R² value, and lower p value indicates a stronger correlation. Statistical significance is usually defined at p < 0.05 or p < 0.01)


Variable_________________|__ R² __|_____p <_____
Black %______________________ .740 ____ 0.000000001
HS Graduation rate_____________ .730 ____ 0.000000001
Historically Black protestant %____ .656 ____ 0.0000001
Home internet access ___________ .618 ____ 0.000001
Religious "nones" %____________ .505 ____ 0.00001
Adult English literacy %_________ .415 ____ 0.0001
Evangelical %_________________ .385 ____ 0.001
Latitude______________________ .354 ____ 0.001
White %_____________________ .116 _____ 0.06


Which is to say that the prevailing truism that Bernie Sanders only does well in "lily-white northern states", is not as well-supported as the more accurate finding that Bernie Sanders does well in "more literate, high school-educated, internet-penetrated states".

[Fumesucker]

34. " Bernie Sanders does well in "more literate, high school-educated, internet-penetrated states". "

The ones that routinely vote for Democrats in the general.

Give them the candidate they don't prefer and depress the turnout in those states that really matter.

Brilliant electoral strategy, just brilliant.

[MineralMan]

23. Looks like you put a lot of work into that.

I'll look forward to comparing your predictions against actual election results. If you're right, it appears that Hillary will be the nominee. That will be a very unpopular prediction.

Thanks for posting your analysis.

[hellofromreddit]

24. Did you determine a beta for your model?

[Godhumor]

25. Exactly how much time do you think I spent on this?

It is a look at standard deviation, nothing more, so, no, I didn't calculate Type I or Type II errors or run a regression analysis or figure out significance factors.

[yardwork]

26. Kick for interesting model.

[Godhumor]

28. 4 for 4 n/t

[SidDithers]

29. Nice..nt

Sid

[FSogol]

30. So far, so good. n/t

[Godhumor]

31. 5 for 5 (This predicts state winner by vote % not delegate count) n/t

[yardwork]

32. Kick.

[DanTex]

33. Nice. Looks to me like KY is within 1 Std Dev of both means. Is there a reason why you have it

down as Hillary instead of Bernie?

[Godhumor]

35. Closer to the mean as a % for Hillary

8.2 is right on the very edge of the standard dev for Bernie while Hillary had a bit more room. A little arbitrary, bit I can justify the why.

[DanTex]

36. Make sense. I think you have an outside shot at getting 100% correct.

I'd say that this model's biggest vulnerabilities would be IN, KY, CA and NM, but still, it's interesting that you can get so much accuracy with just one variable.

[Godhumor]

37. I'm recalibrating for May after the 26th

Basically, adding the states already predicted into the standard deviation calculation to get more accurate for the next wave. My guess is a few of those states will flip.

Interestingly enough, Harry Enten of 538 over the weekend went much deeper than this model and discovered that 72% of Clinton vs Sanders delegates is correlated to POC population of a state. 72%!

That is why my half-assed version is working so far.

[Bread and Circus]

38. Interesting. What I have read is that it's not "black people" voting for Clinton but rather

..."older black women" who vote for her which has given her a large number of Democratic votes.

By percentages, from what I have read, older black men don't vote as much and younger black voters are more evenly split.

It seems Hillary's firewall is mainly comprised of older black women, which is a very nice constituency to have, because they reliably vote.

What is interesting me psychologically is how set in stone this voting bloc has become for Clinton so much so you feel confident developing the political model you have.

I wonder what they see in Hillary Clinton to feel such kinship.

[Godhumor]

39. And 6 for 6. N/t

[yardwork]

49. Predicts a RI win for Bernie on Tuesday.

[yardwork]

50. .

[Godhumor]

51. It does, indeed.

I also think it is his best shot next week (Though it could easily be the model's first miss.).

I'll be recalibrating the model for May after next week (Will fold completed primaries into the numbers) and see if it shifts any out states.

[yardwork]

52. Our posts aren't kicking your thread.

This is the second time I've noticed this problem with a thread no longer bouncing to the top of the page when somebody posts in it.

You might want to ask Skinner about this in ATA. The previous OP did so and he said they'd look into it.

Your thread is interesting. People would like to read it.

[Godhumor]

54. Thick it triggered the age of post cutoff

Might need to make a new thread for next week, then...

[Agnosticsherbet]

53. I did not see this until now.

Very informative post.