Mormons and Evangelicals

Block voting has interested me a lot this election season. Particularly of interest is Romney’s support from Mormons (historically, an extremely conservative voting bloc) and Huckabee’s support of Evangelicals (historically, a relatively conservative voting bloc). In order to do a theoretical analysis, I used data from two sources.

First, for historical election and 2008 primary voting data I used Dave Leip’s Atlas of U.S. Presidential Elections at http://www.uselectionatlas.org/. This website is great in giving enormous amounts of data about past and current elections.

Second, for the religious data I used numbers from the Association of Religious Data Archive at http://www.thearda.com/. An issue does arise in looking at these two blocks of voters. Namely, Mormons are a well defined religious denomination. There is very little debate about who is a Mormon. Evangelicals on the other hand are not a defined religious denomination but a subset of Christianity that includes members of several denominations (although not necessarily entire denominations). Therefore defining who is an Evangelical poses a problem. The ARDA takes a denominational approach.

I wanted to determine answers to a few questions. 1) What are the voting habits of the block in comparison to the rest of the population? 2) What influence does the block have on the outcome of the election? 3) Was Romney’s support primarily because of the Mormon factor?

To begin I chose one heavily Mormon state (Utah – 66.45%) and two Evangelical states (Alabama – 40.59% and Oklahoma – 41.49%) to develop a baseline. Arkansas has a higher percentage of Evangelicals, however since it was the home state of Huckabee I elected to exclude it. Some may argue that based on this I should exclude Utah (since Romney lived there while running the Olympics), unfortunately, there isn’t another state that is as heavily Mormon (Idaho is next with 24.07%).

For each state, I determined what the non-block vote would theoretically be by assuming that the states bordering them would vote similar to the non-block population. Admittedly, this is more accurate for the Mormon block since the states surrounding Utah have a much smaller non-Mormon population. The states surrounding Alabama and Oklahoma have less of an Evangelical population, but still a significant percentage of Evangelicals.

I used Presidential election returns from 2000 and 2004 to develop a baseline. The non-block vote (Other) was compared with the Total returns for those years to obtain the block vote that would result in those returns.

Mormon Voting
	R	D	VAP	REG
Utah - Other	58.04%	38.70%	55.32%	75.81%
Utah - Total	69.18%	26.17%	54.25%	70.60%
Utah - Mormon	74.81%	19.84%	53.71%	67.97%

Evangelical Voting - Alabama
	R	D	VAP	REG
Alabama - Other	56.33%	41.89%	50.58%	66.05%
Alabama - Total	59.46%	39.22%	52.90%	69.30%
Alabama - Evangelical	64.05%	35.30%	56.30%	74.06%

Evangelical Voting - Oklahoma
	R	D	VAP	REG
Oklahoma - Other	56.33%	41.89%	50.58%	66.05%
Oklahoma - Total	62.94%	36.43%	51.90%	61.80%
Oklahoma - Evangelical	72.26%	28.73%	53.76%	55.81%

R is the percent that voted for the Republican Candidate, D is the percent that voted for the Democratic Candidate, VAP is the percent of the voting age population who voted, and REG is the percent of registered voters who voted.

From these numbers, I conclude that Mormons are slightly more conservative than Evangelicals as a whole, but that they are also slightly less likely to vote than non-Mormons.

With this information I looked at the 2008 primary returns for Utah, Alabama, and Oklahoma. I ran various scenarios of support for Romney and Huckabee with differing amounts of the block turnout. The following is what I found:

1) Even if Mormons turned out in massive numbers in Utah (90% to 10%) and supported Romney by a substantial margin (95%), for Romney to win 90% of the vote, he still would need 56% support from non-Mormons. This was echoed in the CNN exit poll.

2) With 60% Evangelical turnout (as in Iowa) and 50% of Evangelicals supporting Huckabee (as in Iowa), Huckabee would need the support of 26% of the other voters in Alabama and 9% of the other voters in Oklahoma to match his returns there.

From this then I can conclude a few things. Romney’s support is not primarily based on Mormon voters. There are enough non-Mormon voters to have given him the win in Utah, and there aren’t enough Mormon voters in other states to explain the support in other states (this is further supported by exit polls which show that Romney did have the support of a significant amount of Evangelicals and Catholics). Huckabee on the other hand had a majority of his support from Evangelicals, although exit polls do show that he did get support from other block groups. This is also evident from his wins in the South, Kansas, and Iowa which had significant Evangelical turnout according to exit polls.

One thing I have seen often mentioned on the blogs is the Mormon vote in states outside of Utah (particularly Nevada, Colorado, Wyoming, California, and Florida). So I decided to create a swing factor which is a product of the % registered to vote, the percent of dominant political party, and the percent of the block in the state.

State	% Mormon	Swing Factor
Utah	66.45%	33.79%
Idaho	24.07%	12.24%
Wyoming	9.54%	4.85%
Nevada	5.85%	2.98%
Colorado	2.15%	1.09%
California	1.56%	0.79%
Florida	0.47%	0.24%

An election would need to have a margin less than the Swing Factor for the block to have any real political clout. All of the primary/caucus margins were greater than the swing factor for these states; so, in essence, the Mormon vote (as a block) was not a deciding factor in any primary/caucus. There simply are not enough Mormons in other states even if we assume they are all Republican and that they all show up to vote to influence the elections by themselves. An interesting note, with the Florida swing factor of 0.24%, this means that during the 2000 election, there was enough Mormon clout to swing the election to Gore. However, based on the fact that Utah (ergo Mormons) overwhelmingly voted for Bush in that year, that would have been a near impossibility (besides if Mormons were for Gore, then Utah would have voted for Gore and the whole Florida mess would have been moot).

Governors vs. Senators

There is plenty of talk of Governors making better presidential candidates than Senators. Senators are said to lose while Governors win. The usual reason cited is executive experience. I decided to see for myself whether this was the case. I looked at all of the presidential elections since 1900. I divided candidates into four categories.

1) Senators - candidates whose last elected office was Senator (not necessarily a sitting Senator)
2) Governors - candidates whose last elected office was Governor (not necessarily a sitting Governor)
3) Presidents/Vice Presidents - candidates whose last elected office was President or Vice President (not necessarily a sitting President or Vice President)
4) All others (Congressmen, Supreme Court Justices, Businessmen) Senators won the presidency 2 out of 7 times (29%).

Governors won the presidency 6 out of 14 times (43%). Presidents/Vice Presidents won 16 out of 25 (64%). Using the Governors value as a baseline probability, and a cumulative binomial distribution, there is a 36% probability that the Senators would only win 2 contests out of 7 tries. Not a high percentage, but certainly not prohibitive; ergo, Senators are not really disadvantaged compared to Governors. In comparison, there is only a 3% probability that the Pres/Vice Pres would win 16 of 25. This indicates that the Presidents and Vice Presidents definitely have an edge over Governors or Senators. A couple of interesting tidbits. While Governors and Senators are statistically fairly even, Governors have received twice as many nominations as Senators. Only once have we had a Senator vs. Governor matchup, in 1920 and the Senator won.

Apples and Oranges

Comparing apples and oranges must be a required course for candidates in order to run for office. Here is a quick example. http://www.factcheck.org/elections-2008/myrtle_beach_blarney.html Huckabee continues to tout the roads as a great accomplishment. However just listening to the words should raise red flags. Going from worst to most improved doesn't mean anything. Going from worst to 4th worst is not something most people would brag about (apples). But comparing it to improvement (oranges) makes it more palatable. Of course if you start out at the very bottom, there is plenty of room for improvement. So what the truth is, is that the roads in Arkansas were awful before Huckabee, after Huckabee they are still awful, just not as much.

Regional Candidates

Now that Mitt Romney has decided to suspend his campaign, I thought it would be worthwhile to examine some of what I have read about regional candidates. On Super Tuesday, my wife and I discussed the image of the US with the states highlighted that each candidate had won. It gave the picture that Romney's support was in the west and midwest, McCain's support was in the Northeast and California, and Huckabee's support was in the South. So I decided to examine this further.

I took the percentages that each candidate received in the the states thus far. Then I divided them up regionally and found the average. To determine if someone was a regional candidate I assumed that they would have to have greater than 35% support in one region and less than 15% support in at least two regions.

McCain Romney Huckabee Paul
Northeast 43% 37% 9% 7%
South 32% 22% 36% 4%
Midwest 28% 33% 23% 10%
West 23% 50% 11% 11%
Average 32% 35% 20% 8%

From the numbers I have above, only Huckabee meets my definition of a regional candidate. He won the majority of the states in the South and had a good showing in the states he didn't win. However, his support in the Northeast and the West is dismal, and even in the Midwest, his victory in Iowa is what is boosting his average (without Iowa, his Midwest average is 18%). What I found more interesting were the numbers for Romney and McCain. Romney had strong support in each of the three regions he lived in (Massachusetts, Michigan, Utah), and was still able to draw a significant number of votes in the South against a regional candidate. His West average is skewed upward by the 90% landslide in Utah (look for my article on the Mormon vote for more on this). However, he still won the majority of states in the West. McCain like Romney has fairly strong support all over the country. His strength is the Northeast with a significant support component in the South. His West support has been concentrated in California and Arizona (his home state).

I don't think that anyone would argue that McCain is not a regional candidate. However the question remains, is Romney? I would say no. His numbers compare very favorable to McCain, and the overall average is similar to McCain's (32% vs 35%). So what is the difference? McCain has significant support in the more populous states in each region. In the Northeast, McCain won NY, NJ, and CT compared to Romney's MA and ME. In the Midwest McCain won IL and MO while Romney won MI and MN. In the South McCain won FL, Romney didn't win anything. In the West McCain took CA and AZ while Romney had UT, CO, MT.

One other piece of evidence to look at is the number of states versus the percent of support.

McCain Romney Huckabee
0%-10% 1 0 7
10%-20% 4 3 9
20%-30% 5 8 3
30%-40% 9 9 6
40%> 8 7 2

From this we can see that the majority of states have Huckabee in the 0%-20% support range. McCain's majority is in the 30% and above, while Romney's is the 20%-40%. Excluding Utah, Romney and McCain garnered greater than 10% support in every state that has been contested. What this all tells me is that Romney did a good job of gathering a strong support for himself across the country. It wasn't enough to beat McCain who also has widespread support. Huckabee has not shown that he can get significant support outside of the South.

Next would be to look at how the numbers shake out with Romney out of the Race. A simple analysis of Romney voters going half for McCain and half for Huckabee yields the following regional result:

McCain Huckabee
Northeast 61% 27%
South 43% 47%
Midwest 44% 39%
West 48% 36%

McCain still wins three of four regions. If we assume Huckabee is able to court 2/3 of the Romney vote the outcome is different.

McCain Huckabee
Northeast 55% 33%
South 40% 50%
Midwest 39% 45%
West 40% 44%

Huckabee wins three of the regions. That being said, McCain still has strong enough support that he can win another 500 delegates and lock in the nomination. Since Huckabee has not shown much support in the West and Northeast, I would consider it very optimistic to assume that he even gets half of the Romney votes in those regions. All this is well and good, but in the end a state by state analysis would be better and nothing beats the actual results from the primaries and caucuses. If I was a betting man, my money would be on McCain.

Interesting Use of Poll Statistics

I have been amazed at how much the RealClearPolitics (RCP) poll average is quoted. This number is basically meaningless. What they do is compile all of the polls and then take an average of the most recent several polls. In some cases I have seen the polls that are averaged being weeks apart. So for instance say McCain was up by 20 points three weeks ago in one poll. However more recent polls show only a 3-4 point spread between McCain, Romney and Huckabee. Because of their averaging method, the RCP average may still have McCain up by 5-8 points. This could be the case even if Romney or Huckabee is ahead in the more recent polls. Additionally, each poll has a different methodology of selecting their sample and asking the questions. So trying to compare what the Rasmussen poll means in relation to the SurveyUSA poll is not a simple task of looking at the numbers. A better way to look at the polling information is to look at each poll over a period of time. This is especially effective for those tracking polls that do a rolling sample over a few days. This way one can see if Obama has "momentum" or if Clinton's support is remaining steady. In the end, a poll doesn't tell the story like election returns.