Carroll College Election Day Exit Poll Results

Friday, December 19, 2014

On November 4, students in Assistant Professor of Political Science Dr. Street’s class on “Elections, Political Parties, and Public Opinion” conducted an exit poll in Lewis and Clark County. The 14 students designed a survey of 11 questions, and collected responses from 957 helpful citizens on Election Day. Most of the responses were gathered in Helena, but students also traveled to East Helena, Lincoln and the Helena Valley. The students wrote research papers on the results from the exit poll, which are summarized by topic here.

1) Urban and Rural voters

Analysis by Hannah Hafner, Connor Hausauer and Dominic Salle

We split Lewis and Clark County voters into those living in rural, suburban and urban precincts. The survey asked how people voted in four races: the US Senate and House, and the Montana Senate and House. We measured the average number of these four votes that went to Democrats in each kind of precinct (almost all votes went to either Democrats or Republicans). We found more support for Democratic candidates in urban areas than in either suburban or rural precincts. Figure 1 shows the average number of votes for Democrats cast by voters in each kind of precinct. Vertical lines through each bar show 95% confidence intervals.

Figure 1: Support for Democrats in rural, suburban and urban precincts

2014 Political Poll

We also found that government workers were significantly more likely to vote for Democrats. However, this does not explain the differences between rural, suburban and urban precincts. In the urban areas, even the people who do not work for the government voted for Democrats at higher rates than either government workers or non-government-workers in the other precincts.

2) Political “independents” vote much like partisans

Analysis by Kacey Gollehon and Cori Losing

Almost one third of the voters described themselves as “independent” or as having no party preference, rather than as Republicans or Democrats. But even the people who did not describe themselves as partisans tended to vote consistently for either Democratic or Republican candidates. Figure 2 shows the percentage of people who voted a strict party line (all four votes for either Democrat or Republican candidates), by self-described partisanship. Vertical lines through each bar show 95% confidence intervals.

Figure 2: Percentage who voted a strict party line, by partisanship

2014 Political Poll

 

The strong tendency to vote along party lines is surprising in Montana, since the state is known for ticket-splitting. In recent decades, Montana has tended to elect Democrats to the US Senate but the state’s Electoral College votes have gone to Republican presidential candidates.

3) Education and ideology

Analysis by Anna Hoerner, Mark Schmutzler and Hunter Shima

We compared survey respondents in different ideological camps. To do this we used information on partisanship and ideology. We created five groups. At the extremes are the liberal Democrats and the conservative Republicans. In the middle are people who describe themselves as ideological moderates and independents. We also included people who lean to one side or the other. For instance, those who “lean left” are either moderate Democrats or liberal independents.

We studied whether people at different points on this ideological spectrum have different social backgrounds. In particular, we looked at whether they work for the government, and their typical levels of education. We found modest differences in the proportion of government workers. We found a clearer pattern for education. Figure 3 shows the percentage of people with four-year degrees in the different ideological groups. Vertical lines through each bar show 95% confidence intervals.

Figure 3: Education by ideology, from Liberal Democrat to Conservative Republican

2014 Political Poll

As the figure shows, we found the clearest differences between those in the center and left, and those on the right. In each of the three center and left categories, there were more people with college degrees than without, whereas in the two categories on the right the numbers at each level of education were about equal.

4) Education and partisanship

Analysis by Michael Blum, Grady Holt-Seavy and Jared McCauley

We studied how education is related to partisan identities. Specifically, we tested whether more educated people are more or less likely than people with less education to be partisans, independents, or to have no party preference. We found that more educated people are more likely to describe themselves as independents, and less likely to have no party preference.

These relationships are displayed in Figure 4. Around 24% of those with at least a four-year college degree described themselves as independents, compared to 18% of those with less education. And around 7% of those with a BA or more said they had no party preference, compared to 13% of those with less education. Vertical lines through each bar show 95% confidence intervals. Despite the partial overlap in the confidence intervals, each of these differences is statistically significant at a probability of less than 0.05.

Figure 4: Education, independents and those with no party preference.

2014 Political Poll

5) Presidential approval in non-Presidential elections

Analysis by Kristel Barry, Erin Gordon and Jack Rogers

We studied whether presidential approval is linked to voting decisions in elections that do not actually have the President on the ballot. We measured how many of the votes in the four elections we studied—the US Senate and House, and the Montana Senate and House—went to Democrats, and how this varied for people who said they approved or disapproved of the job that Obama is doing as President.

Figure 5: Presidential approval and the average number of votes for Democrats

2014 Political Poll

As is clear from Figure 5, we found a very strong relationship. 19 out of 20 voters who said that they approved of Obama’s work cast all of their votes for Democrats. In contrast, on average only around one in six of the people who disapproved of Obama cast a vote for a Democrat.

6) LR126 – Proposal to end Election-Day voter registration

This measure was supported by 43% of voters statewide, and by 41% of voters in Lewis and Clark County. As Figure 6 shows, we found majority support among Republicans (58%) but not among Democrats, Independents or those who expressed no general preference for a political party. Vertical lines through each bar show 95% confidence intervals for the survey estimates.

Figure 6: Support for LR126 by partisanship

2014 Political Poll

We also found majority support for LR126 among people who described themselves as ideological conservatives (59%) but not among moderates (24%) or liberals (7%). No age group provided majority support for the proposition.

Acknowledgements

We are very grateful for the support and advice of Audrey Dufrechou and Lisa Kimmet, and for comments and questions from Lisa Kimmet and Secretary of State Linda McCulloch. We also thank Dr. Jeremy Johnson for advice on the survey.