Analyzing the South Carolina Gubernatorial Election, Part 3

This is part of three posts analyzing the 2010 South Carolina gubernatorial election, in which Republican Nikki Haley won a closer-than-expected victory over Democrat Vincent Sheheen. The main focus of these posts will be to explore whether a racial effect accounted for Ms. Haley’s unexpected poor performance.

(Note: 

(Note: Note: I strongly encourage you to click the image links on this post when reading; they're essential to understanding what I'm saying. This is also part of a series of posts analyzing the 2010 midterm elections.)

 

Link to Map of South Carolina, 2010 Gubernatorial Election

 

The previous post mapped out the relationship between Democratic shifts in 2010 and white registration numbers. Here is the relevant map reposted:

 

Link to Map of South Carolina, 2010 Gubernatorial Election Comparison to Registered Voters

 

The post ended by noting that “So far this analysis has been relatively light on the statistical side of things.” It included a number of maps, but did not use any raw numbers.

This post aims to draw conclusions based on those numbers.

Let’s begin by translating the picture above into a graph:

Link to Graph of Relationship Between White Registration and Democratic Shift in the 2010 South Carolina Gubernatorial Election

 

This graph maps the relationship between how white a county in South Carolina is, and how much it shifted against non-white Republican candidate Nikki Haley in 2010.

If normally-Republican whites moved against Ms. Haley due to her race, one would expect the dots to be graphed in a roughly 45-degree diagonal line; the whiter a county, the more Democratic it would shift in 2010.

Clearly this is not the case in the graph above. There are a lot of very white counties that shifted strongly against Ms. Haley – but there are also a lot of very white counties that supported her more than they did Senator John McCain.

Indeed, the whitest counties seem to spread out into two groups; one group moves strongly against Ms. Haley, another actually shifts for her. One might speculate that the former group is composed of lower-income, rural whites and the latter is composed of higher-income, metropolitan whites.

To test this theory, the previous post adjusted for income by eliminating all the counties with a median household income greater than the state median (i.e. it got rid of the rich whites). Here is what the result looked like:

 

Link to Map of South Carolina, 2010 Gubernatorial Election Comparison to Registered Voters, Adjusted For Income

 

There seems to be a correlation here, as the previous post noted.

Here is how the relationship looks on a graph:

 

Link to Graph of Relationship Between White Registration and Democratic Shift in the 2010 South Carolina Gubernatorial Election

 

The group of white counties which shifted towards Ms. Haley has disappeared. Instead, one sees a much stronger trend: the whiter the county, the more strongly it moved against non-white Republican Governor Nikki Haley.

This only happens once high-income white counties are tossed out of the analysis. High-income Republican whites were very comfortable voting for non-white Republicans; low income Republican whites were less willing.

Interestingly, this pattern is not unique to South Carolina. In Louisiana, Republican Governor Bobby Jindal – a non-white individual of Indian descent – did extremely poorly amongst rural, low-income (Republican) whites while winning landslide support amongst high-income, suburban (Republican) whites. This caused Mr. Jindal to lose in his first attempt to run for governor.

Finally, one can test whether the effect above is statistically significant, or just the result of randomness.

Here is a regression analysis run on the 2010 South Carolina gubernatorial race:

Link to Regression Analysis of the 2010 South Carolina Gubernatorial Race

 

Regression analysis is something I am still not fully comfortable with, so bear this in mind as the analysis continues.

The regression attempted to use two variables – race and income – to predict whether voters would vote more Democratic in 2010. Specifically, it used the percent of white registered voters in a county and said county’s median household income.

The model states that every 10% increase in white registered voters results in a 3.65% greater Democratic shift against Ms. Haley (this is the Coefficient column at the bottom left).

More importantly, whiteness and income were statistically significant when placed together; there was a 0.1% chance that the effect of whiteness was random, and a 0.4% chance that the effect of income was random (this is the P>|t| column at the bottom center).

So the evidence is fairly strong that racially-based voting by low-income whites hurt non-white Republican Ms. Haley in 2010.

There is, however, a caveat. The above regression only explains 20% of the variance between the different degrees of Democratic shifts between different counties (this is the Adj R-Squared line at the top right). This means that 80% of the variance is not explained by race and income.

Racism probably hurt Ms. Haley in 2010, but it was far from the only factor.

--Inoljt, http://mypolitikal.com/

 

Analyzing the South Carolina Gubernatorial Election, Part 2

This is the second part of three posts analyzing the 2010 South Carolina gubernatorial election, in which Republican Nikki Haley won a closer-than-expected victory over Democrat Vincent Sheheen. The main focus of these posts will be to explore whether a racial effect accounted for Ms. Haley’s unexpected poor performance.

The previous post can be found here, and the next post can be found here.

(Note: Note: I strongly encourage you to click the image links on this post when reading; they're essential to understanding what I'm saying. This is also part of a series of posts analyzing the 2010 midterm elections.)

 

Link to Map of South Carolina, 2010 Gubernatorial Election'

 

How to Find a Racial Effect

The purpose of this series of posts is to determine whether or not Ms. Haley’s relatively weak performance was due to a racial effect.

In order to due this, it’s necessary to define what to look for. In this case, it would be normally Republican voters abandoning Ms. Haley due to her race.

Now, South Carolina is a state in which less than 5% of the population is neither white nor black; minorities other than blacks play a negligible role in the state’s politics. It is also a very racially polarized state, like most places in the Deep South. Blacks vote Democratic; whites vote Republican.

There is one final factor to take into account. When Republican Bobby Jindal ran for governor in 2003 and faced racially-based opposition by (white) Republicans, such opposition was not evenly distributed. The Republicans who abandoned Mr. Jindal tended to be predominantly from rural, relatively lower income areas. This is something that is not especially surprising, although it conforms to some unfortunate stereotypes.

For these reasons, an examination of Republicans who abandoned Ms. Haley for racial reasons would look specifically at areas with lower-income whites. These areas would be expected to shift more Democratic than the norm.

Democratic Shifts

To begin this post, let’s examine the places where Republicans improved upon their 2008 performance, and the places where Democrats improved upon 2008.

Naturally, given that Ms. Haley did worse than Mr. Sheheen, one would expect Democrats to have relatively more improvement.

This turns out to be the case:


Link to Map of South Carolina Shifts, 2008 Presidential Election to 2010 Gubernatorial Election

 

Here one sees a very interesting regional pattern, a pattern that I did not expect when making this map.

The northern parts of South Carolina moved strongly Democratic in 2010. The sole exception is York County, which for whatever reason shifted Republican (there is, strangely enough, very little that differentiates this county with others in the region; nor did either Ms. Haley or Mr. Sheheen represent the county as politicians before 2010).

On the other hand, the coastal regions actually supported Ms. Haley more than they did Senator John McCain.

This is a very interesting regional divide; it is something that is entirely hidden by normal partisan patterns.

Whites

Now, let’s take a look at white registration figures:


Link to Map of South Carolina Registered Voters

 

This map shows what percent of South Carolina’s registered voters are white. The information is mandated by the Voting Rights Act, given South Carolina’s history of preventing minorities from voting, and can be found at this website. It is also quite useful for the purposes of this analysis. (For fun: compare this map to President Barack Obama’s performance).

In order to make comparisons easier, the same color scale was used in this map as in the previous map. The whiter a county’s voter population, the bluer the county on the map.

If white Republican voters rejected Ms. Haley due to her race, then the whitest counties here would also have the strongest Democratic shift (i.e. the colors in each map would roughly match).

Let’s compare the maps:


Link to Map of South Carolina, 2010 Gubernatorial Election Comparison to Registered Voters

 


There is a bit of a match, but not much. A lot of very white counties shift strongly against Ms. Haley, but a lot of them also shift strongly for her (especially along the coast).

One can reasonably conclude that a lot of white voters – i.e. Republicans – remained loyal to Ms. Haley despite her Indian heritage.

This is not entirely unexpected. Mr. Jindal also retained a large amount of white support, mainly amongst suburban and wealthy whites.

Adjusting For Income

Where Mr. Jindal did especially poorly – and why he lost the 2003 gubernatorial election – was amongst rural, lower income whites in Louisiana.

Let’s therefore shift this analysis by adjusting for income; in other words, by focusing upon lower-income counties in South Carolina.

South Carolina’s median household income was $42,580 as of 2009, according to Census Data (which can be accessed here).

One can therefore adjust for income by restricting the analysis only to those counties in which median household income was below the state median.

This is what happens:


Link to Map of South Carolina, 2010 Gubernatorial Election Comparison to Registered Voters, Adjusted For Income

 

This looks like a far stronger relationship. In the poorer parts of South Carolina, it appears that the whiter the county, the more against Ms. Haley it shifted.

It seems that we have found something here.

So far this analysis has been relatively light on the statistical side of things; it kind of looks like there is a pattern in the map above, but perhaps there isn’t one. How likely is it that this could have occurred by chance?

The next post will answer this question.

--Inoljt

 

Analyzing the South Carolina Gubernatorial Election, Part 1

This is the first part of three posts analyzing the 2010 South Carolina gubernatorial election, in which Republican Nikki Haley won a closer-than-expected victory over Democrat Vincent Sheheen. The main focus of these posts will be to explore whether a racial effect accounted for Ms. Haley’s unexpected poor performance.

The next post can be found here.

(Note: I strongly encourage you to click the image links on this post when reading; they're essential to understanding what I'm saying. This is also part of a series of posts analyzing the 2010 midterm elections.)

Link to Map of South Carolina, 2010 Gubernatorial Election

It was the October, 2010 in South Carolina. Nikki Haley, Republican candidate for South Carolina governor, was cruising. She was a conservative candidate – endorsed by none other than Sarah Palin herself – running in a conservative state, in the best Republican year in a generation.

Opinion polls showed the Republican politician leading by double-digits. Even the most pessimistic gave Ms. Haley a high single digit lead.

On election day, however, Ms. Haley won by only 4.5%:

Link to Map of Margins South Carolina, 2010 Gubernatorial Election

What could have accounted for Ms. Haley’s poor performance?

Several factors come to mind. Ms. Haley was not an uncontroversial candidate; her positions were conservative even for South Carolina. The Democratic candidate, Vincent Sheheen, might have been an unnaturally talented campaigner. And there is always the factor of randomness to take into account. There were hundreds of races in November; the polls would inevitably be inaccurate on one or two, and this race just happened to be one of them.

Or perhaps there is another explanation – a particularly ugly one, but one that lurks at the back of everybody’s head. Ms. Haley was an woman of Indian heritage running to govern South Carolina, a state with not exactly the most innocent racial history. Throughout the campaign, Ms. Haley was subject to attacks that implicitly played up the racial angle: she had had affairs with white men (unfortunately for the accusers, this attack doesn’t work as well against women), she wasn’t Christian or was only pretending to be one, and so on.

It is not unimaginable that a sort of Bradley effect took place in South Carolina, that a number of normally steadfast Republicans balked at voting for the first non-white and female governor in history.

This is a serious accusation, and therefore needs serious evidence. The next post will therefore begin an extensive examination of whether Ms. Haley’s race undermined her performance.

--Inoljt

 

 

The Worst Republican Senate Candidates of 2010, Part 2

This is the second part of two posts analyzing patterns in the 2010 Senate midterm elections. The previous part can be found here.

The previous post presented a table ranking the worst Republican candidates in the 2010 midterm elections. The model used to create the table is also explained in the previous post.

Let’s take a look at this table once again:

State                Margin (R) Cook PVI Overperformance

South Dakota    100.00%      8.9%        91.10%

North Dakota     53.91%       10.4%      43.51%

Kansas              43.72%       11.5%      32.22%

Iowa                 31.05%        -1.0%     32.05%

Idaho                46.25%       17.4%      28.85%

Oklahoma         44.50%        16.9%      27.60%

Florida              28.69%        1.8%       26.89%

South Carolina  33.83%        7.8%        26.03%

New Hampshire 23.22%        -1.6%       24.82%

Arizona             24.14%        6.1%       18.04%

Alabama            30.47%        13.2%     17.27%

Ohio                  17.44%        0.7%      16.74%

Georgia             19.31%        6.8%       12.51%

Arkansas           20.96%        8.8%       12.16%

Missouri            13.60%        3.1%       10.50%

Illinois              1.60%          -7.7%      9.30%

Louisiana          18.88%        9.7%       9.18%

Utah                 28.79%        20.2%     8.59%

Indiana             14.58%        6.2%       8.38%

North Carolina   11.77%        4.3%       7.47%

Wisconsin          4.84%        -2.4%      7.24%

Pennsylvania     2.02%        -2.0%       4.02%

Kentucky          11.47%       10.4%      1.07%

Washington       -4.73%       -5.0%      0.27%

Alaska              11.94%        13.4%    -1.46%

Colorado          -1.63%        0.2%       -1.83%

California         -10.01%      -7.4%     -2.61%

Nevada            -5.74%        -1.3%     -4.44%

Connecticut      -11.94%     -7.1%      -4.84%

Delaware         -16.58%      -7.0%      -9.58%

Oregon            -17.98%      -4.0%     -13.98%

New York (S)    -27.84%     -10.2%    -17.64%

Maryland         -26.44%      -8.5%      -17.94%

West Virginia   -10.07%      7.9%       -17.97%

Vermont          -33.41%     -13.4%    -20.01%

New York        -34.10%      -10.2%    -23.90%

Hawaii            -53.24%      -12.5%    -40.74%

Total/Average  5.54%          2.3%         8.08%

(Note: The data in Alaska and Florida refer to the official candidates nominated by the parties, not the independent candidates – Senator Lisa Murkowski and Governor Charlie Crist – who ran in the respective states).

There are six possible outcomes which are possible here. This post will look at each outcome.

Outcome #1: A Republican candidate, running in a red state, wins while overperforming.

This outcome was by far the most common in the November elections: indeed, 18 Senate races fit this category. In a way this is not too surprising: the definition of overperforming here is doing better than the state’s Cook PVI (how a state would be expected to vote in a presidential election in the event of an exact tie nationwide). The average Republican should have “overperformed” in this sense, given how Republican a year it was.

Another factor is incumbency. Red states generally had Republican incumbents. Facing little serious competition in a Republican year and benefiting from their incumbency status, these people were probably expected to overperform – and they did.

Outcome #2: A Republican candidate, running in a red state, wins while underperforming.

Technically this did not happen once in this election. The race that comes closest is Alaska , where Republican candidate Joe Miller did better than the Democratic candidate while doing worse than Alaska ’s political lean (on the other hand, he still lost to Independent Lisa Murkowski).

This is actually quite surprising. There were twenty-one Senate contests in red states – and in just one (or zero, depending on how you count) did the Republican underperform while still winning.

In fact, this outcome is quite rare, for whatever reason, throughout American politics. If a Republican underperforms in a red state, he or she usually loses. Rarely does a Republican candidate underperform in a red state but still win (another variant along the same theme: out of the counties Senator John McCain won, he almost always improved on Republican performances in 1992 and 1996). Why this happens is something of a continuing mystery to this blogger.

Outcome #3: A Republican candidate, running in a red state, loses while underperforming.

This was another rare occurrence in the 2010 Senate elections. Only two states fit this category: West Virginia and Colorado . The performance of Democratic candidate Joe Manchin is especially remarkable. Mr. Manchin was the only Senate Democrat to win in a deep red state this year, and his name stands out as an outlier everywhere in the table.

Outcome #4: A Republican candidate, running in a blue state, wins while overperforming.

There are five states that fit this category: Illinois , Iowa , New Hampshire , Pennsylvania , and Wisconsin . These account for three of the Republican pick-ups this cycle. Interestingly, four of these states are in the Midwest , where Democrats were pummeled this year.

Among these states, Illinois stands out the most. It is the only deep blue state that a Republican candidate overperformed in. Although much of this is due to other factors – the continuing Blagojevich scandal, the weakness of the Democratic candidate – credit goes to Republican Mark Kirk for an outstanding overperformance.

Outcome #5: A Republican candidate, running in a blue state, loses while overperforming.

This is another outcome that, for whatever reason, rarely seems to happen in American politics; if Republicans overperform in blue states, they generally tend to win.

In 2010 this happened in exactly one state: Washington , where Republican candidate Dino Rossi did 0.27% better than the Cook PVI, but still lost.

Outcome #6: A Republican candidate, running in a blue state, loses while underperforming.

This was the second-most common outcome in 2010; ten states fit this category. These states tended to be the bluest states in America . The fact that Republicans tended to underperform a state’s political lean in the deepest-blue states is another strange pattern in American politics. This is something that the previous post analyzes extensively.

All in all, the table reveals a lot of surprising patterns – things which were not expected when this blogger initially made it. And as for the worst Republican candidate in 2010? That was Campbell Cavasso of Hawaii, who won a mere fifth of the vote against the Democratic institution Daniel Inouye.

--Inoljt

 

The Worst Republican Senate Candidates of 2010, Part 1

This is the first part of two posts analyzing patterns in the 2010 Senate midterm elections. The second part can be found here.

The 2010 congressional midterm elections constituted, by and large, a victory for the Republican Party. In the Senate Republicans gained six seats. While this was somewhat below expectations, it was much better than Republican hopes just after 2008 – when many expected the party to actually lose seats. The Senate results provide some interesting fodder for analysis. The table below indicates which Republicans Senate candidates did the worst in 2008. It does so by taking the Republican margin of victory or defeat in a given state and subtracting this by the Cook PVI of the state (the Cook PVI is how a state would be expected to vote in a presidential election in the event of an exact tie nationwide). Given that Republicans won the nationwide vote this year, the average Republican candidate would be expected to do better than the state’s PVI. A bad Republican candidate would actually do worse than the state’s PVI. Let’s take a look at this table:

State                Margin (R) Cook PVI Overperformance

South Dakota    100.00%      8.9%        91.10%

North Dakota     53.91%       10.4%      43.51%

Kansas              43.72%       11.5%      32.22%

Iowa                 31.05%        -1.0%     32.05%

Idaho                46.25%       17.4%      28.85%

Oklahoma         44.50%        16.9%      27.60%

Florida              28.69%        1.8%       26.89%

South Carolina  33.83%        7.8%       26.03%

New Hampshire 23.22%        -1.6%     24.82%

Arizona             24.14%        6.1%       18.04%

Alabama            30.47%        13.2%     17.27%

Ohio                  17.44%        0.7%       16.74%

Georgia             19.31%        6.8%       12.51%

Arkansas           20.96%        8.8%       12.16%

Missouri            13.60%        3.1%       10.50%

Illinois              1.60%          -7.7%      9.30%

Louisiana          18.88%        9.7%        9.18%

Utah                 28.79%        20.2%      8.59%

Indiana             14.58%        6.2%        8.38%

North Carolina   11.77%        4.3%        7.47%

Wisconsin          4.84%        -2.4%       7.24%

Pennsylvania     2.02%        -2.0%       4.02%

Kentucky          11.47%       10.4%       1.07%

Washington       -4.73%       -5.0%      0.27%

Alaska              11.94%        13.4%     -1.46%

Colorado          -1.63%        0.2%       -1.83%

California         -10.01%      -7.4%     -2.61%

Nevada            -5.74%        -1.3%     -4.44%

Connecticut      -11.94%     -7.1%     -4.84%

Delaware         -16.58%      -7.0%     -9.58%

Oregon            -17.98%      -4.0%     -13.98%

New York (S)    -27.84%     -10.2%   -17.64%

Maryland         -26.44%      -8.5%     -17.94%

West Virginia   -10.07%      7.9%       -17.97%

Vermont          -33.41%     -13.4%    -20.01%

New York        -34.10%      -10.2%    -23.90%

Hawaii            -53.24%      -12.5%    -40.74%

Total/Average  5.54%          2.3%           8.08%

(Note: The data in Alaska and Florida refer to the official candidates nominated by the parties, not the independent candidates – Senator Lisa Murkowski and Governor Charlie Crist – who ran in the respective states).

This table reveals some fascinating trends. There is a very clear pattern: the worst Republican candidates ran in the bluest states – and the bluer the state, the more the Republican underperformed. This does not just mean that these Republicans lost, but that they lost by more than the average Republican was supposed to in the state. Republican candidates did worse than the state’s PVI in thirteen states; nine of these states had a Democratic PVI.

There seems to be a PVI tipping point at which Republicans start underperforming: when a state is more than 5% Democratic than the nation (PVI D+5). Only one Republican in the nine states that fit this category overperformed the state PVI (Senator Mark Kirk of Illinois ).

Something is puzzling about this pattern. It is true that states like Connecticut or Maryland will probably vote Democratic even in Republican victories. The Cook PVI predicts that Democrats will win by X% in the event of a national tie in the popular vote. One would thus have expected Republican candidates to do better than this in 2010, given that 2010 was the strongest Republican performance in a generation.

Yet this did not happen. In a lot of blue states Democrats actually did better than the Cook PVI would project them to do - that is, said blue states behaved like the Democrats had actually won the popular vote, which they certainly did not in 2010. The bluer the state, the stronger this pattern.

There are a couple of reasons why this might be. The first thing that comes to mind is the money and recruiting game. The Republican Party, reasonably enough, does not expect its candidates to win in places like New York and Maryland . So it puts less effort into Republican candidates in those states. They get less money – and therefore less advertising, less ground game, and so on. Nobody had any idea who the Republican candidate in Vermont was, for instance. That probably contributes to Republican underperformance in deep-blue states.

The second factor might be a flaw in the model the table uses. Democratic and Republican strongholds, for whatever reason, behave differently from “uniform swing” models. In almost all the counties President Barack Obama won, for instance, he improved upon President Bill Clinton 1992 and 1996 performance – despite the fact that Mr. Clinton won by similar margins in the popular vote. This holds true from San Francisco to rural Mississippi . In the 2010 Massachusetts special Senate election, the most Democratic areas of Massachusetts swung least towards Republican Senator Scott Brown. The fact that the worst Republican candidates ran in the bluest states fits the pattern.

The table presents another startling pattern, which will be discussed in the next post: there are surprisingly few Republicans who did worse than they were supposed to in red states.

--Inoljt

 

 

 

Diaries

Advertise Blogads