Friday, July 21, 2006

How important were ‘moral values’ issues in the 2004 election? (continued) A look at the exit poll itself

J Langer and G Cohen, in 'Voters and Values in the 2004 Election', use data from the 2004 National Election Poll (NEP) exit poll itself. So they give tour through the data that caused all the fuss in the first place.

After explaining the many reasons why 'moral values' was a lousy choice for a 'top issue' in an exit poll, they show that it was chosen mainly by weekly churchgoing conservative Christians. They present summary data from a series of exit polls that show:
--increase support for Bush from 2000 to 2004 was from conservative white Protestants, and all Catholics, but not among all weekly churchgoing white Protestants,
--about 90% of conservative white Protestants have voted for the Republican presidential candidate since 1980 ( I see a slight upward trend),
--about 55% to 60% of other white Protestants and other Christians have voted for the Republican presidential candidate since 1980 ( I see no upward trend),
--most of increase in support for Bush in 2004 came from infrequent or non-churchgoers, not frequent churchgoers.

They present a statistical analysis of the NEP data to estimate the relationship between voting for Bush and choice of top issue on the exit poll. They included other respondent characteristics, so the effect of choosing one of the top issues could be estimated while holding other characteristics constant. The characteristics were self-identified evangelical Christian, weekly church attendance, Party ID, political ideology (liberal vs. conservative), and race (white vs. other). They did a simulation to estimate the average change in probability of voting for Bush for as a function of choice of top issue in the exit poll sample. The results are shown below.

change in probability of Bush vote
as a function of choice of top issue

--terrorism: 0.172
--economy: -0.113
--Iraq war: -0.099
--moral values: 0.081
--health care: -0.080
--education: -0.066
--taxes: 0.011

Choosing terrorism, economy and the Iraq war as the top issue had more impact than moral values. They also do an analysis that indicates that in the exit poll sample, the term moral values acted like a proxy for opposition to same-sex marriage and belief that abortion should be always or almost always illegal.

Votinglinks Comments: The authors recognize that using this data means they cannot unpack the vague term 'moral issues' in detail but they can take a look at how picking that buzzword itself as the top issue influenced the vote among those who turned out on election day. The main weakness of this paper is that they only report how the issues effect the probability of voting for Bush in the exit poll sample, but provide no info on how well the sample represents all the voters or subgroups.

I have not found any studies that indicate 'moral values', or same-sex marriage or reproductive rights were a dominant issue for all voters or any large political demographic, other than conservative Christians. The question of whether, in the close 2004 presidential election, the marriage and abortion issues may have made a difference in certain states is a different issue, and the next and concluding post in this series will look at two studies that tackle that.

G Langer and J Cohen. Voters and Values in the 2004 Election. Public Opinion Quarterly: 69(5) Special Issue 2005; 744-759.
http://poq.oxfordjournals.org/cgi/content/short/69/5/744?rss=1
(URL is for abstract, may need subscription to download paper, depending on from where you access)

previous post in this series