SOME 2008 MISSISSIPPI POLL RESULTS
This telephone survey was
conducted by the Survey Research Unit of the Social Science Research Center at
Mississippi State University with facilities that permit simultaneous use of
twenty telephones. The SSRC is directed by Dr. Arthur Cosby, and the Survey
Research Unit (SRU) is directed by Dr. John Edwards. A random sampling
technique was used to select the households, and a random method was employed
in sampling each individual within the household. No substitutions were
permitted, and up to ten callbacks were made. Calls were made from 5:00 p.m. to
9:00 p.m. on weeknights, and from 1:00 p.m. to 5:00 p.m. on Saturdays and 1PM
to 9PM on Sundays. A Computer Assisted Telephone Interviewing System (CATI) was
used to collect the data. Twenty-two undergraduate students in the Political
Analysis class made the phone calls.
Five hundred twenty eight adult
Mississippi residents were interviewed from April 2-22. The response rate was
40%. The results were adjusted or weighted by demographic characteristics to
achieve a representative sample. With 528 people surveyed, the sample error is
plus or minus 4.4%, which means that if every Mississippi resident had been
interviewed, the results could differ from those reported here by as much as
4.4%. The sample error is higher for examining only likely voters, since that
sample size is smaller.
The Director of the Survey was
Stephen Shaffer in the MSU Political Science Department. SRU Director John
Edwards, assisted by sociology Wolf Frese, was of great assistance in creating
the CATI questionnaire and providing the required survey expertise.
It is important to note that since this poll was conducted over six months before the November elections, it serves only as a benchmark for voter attitudes in April 2008, and cannot be used to forecast the likely election results in November.
Question wordings:
“If the presidential election was held today, and the candidates were New York Senator Hillary Clinton the Democrat and Arizona Senator John McCain the Republican, whom would you vote for? Clinton or McCain?”
“If the presidential election was held today, and the candidates were Illinois Senator Barack Obama the Democrat and Arizona Senator John McCain the Republican, whom would you vote for? Obama or McCain?”
“If the special election for United States Senator was held today, and the two candidates were Democrat Ronnie Musgrove and Republican Roger Wicker, whom would you vote for? Musgrove or Wicker?”
April 2-22 Poll Results, Inclusive Likely Voter Definition (sample size = 488, sample error is 4.6%; percentages total 100% across each row):
|
Which election contest? |
Democrat Vote |
Republican Vote |
Undecided or refused |
|
Presidential (Clinton vs. McCain) |
42.2% |
48.1% |
9.7% |
|
Presidential (Obama vs. McCain) |
40.8% |
51.0% |
8.2% |
|
Senate (Musgrove vs. Wicker) |
46.9% |
39.3% |
13.8% |
Note that these results are so close that they are within the margin of error in every case except for the Obama-McCain matchup. Therefore, even if these elections were held today, our poll could not predict who would carry Mississippi.
Question wording, two part question: “Generally speaking, do you consider yourself a Democrat, Republican, Independent, or what? (If Democrat, ask:) Do you consider yourself a strong or not so strong Democrat? (If Republican, ask:) Do you consider yourself a strong or not so strong Republican? (If Independent, ask:) Do you think of yourself as closer to the Democratic party or the Republican party?” This 2-part question was combined into a 7 point, party identification scale. The number of Mississippians in each of these 7 groups, and the vote preferences of each of these 7 groups follow (for election preferences, undecideds and refused were excluded from analysis):
|
|
Strong Dems |
Weak Dems |
Indep. Dems |
Pure Independents |
Indep. Reps |
Weak Reps |
Strong Reps |
|
% of people in each group (N= 507): |
21.6% |
15.5% |
11.0% |
8.8% |
12.6% |
12.0% |
18.5% |
|
% backing Clinton over McCain |
87% |
74% |
65% |
63% |
25% |
5% |
13% |
|
% backing Obama over McCain |
82% |
80% |
76% |
28% |
15% |
4% |
14% |
|
% backing Musgrove over Wicker |
85% |
90% |
80% |
61% |
36% |
17% |
8% |
By combining the independents leaning to a party with the strong and weak partisans, we can see how evenly split Mississippians are. Democrats comprise 48.1% of the sample, Republicans comprise 43.1% of the sample, and pure Independents comprise 8.8%. This results in a mere 5 point Democratic party identification margin, which is comparable to the 5 to 7 point Democratic margin in the 1992-1998 Mississippi Polls.
Also note how partisan people are in terms of their votes. Strong Democrats vote at least 82% for Democratic candidates, and Strong Republicans vote at least 86% for Republican candidates. Even Independent Democrats vote at least 65% for Democratic candidates, while Independent Republicans vote at least 64% for Republican candidates. Obviously, the success of each party and their candidates in turning out their supporters to vote on election day will be a very important factor in the election outcomes.
Other Issues: Popular Governor, Unpopular President
Two factors that may or may not have some impact on the November elections are the relative popularity of Mississippi’s Republican governor and the relative unpopularity of President Bush.
Question wording: “Now I'm going to ask you to rate the job performance of a few political figures and institutions. Rate each of them as excellent, good, fair, or poor. What about President George W. Bush? What about Governor Haley Barbour?” (In the table below, percentages total 100% across each row.)
|
Officeholder: |
Excellent |
Good |
Fair |
Poor |
N Size |
|
Barbour |
20% |
35% |
35% |
10% |
(494) |
|
Bush |
9% |
24% |
33% |
34% |
(516) |
In the Mississippi Polls conducted since 1981, only Governor Winter was able to remain above 50% in terms of popularity throughout his term of office, with popularity defined as the proportion of voters having opinions who rated the governor’s performance as either excellent or good. This is the third of three times that the Mississippi Poll has measured Barbour’s popularity, and each time it has remained above 50%. The 55% excellent-plus-good rating in the 2008 poll is particularly impressive, since it includes non-voters as well as voters (non-voters are somewhat more Democratic than voters).
President Bush’s 33% approval rating is the lowest ever measured by the Mississippi Poll for a Republican president. Reagan’s low point was a 40% mark in 1982 during a national recession, and former President George Herbert Walker Bush had a low of 49% in the recession year of 1992. Only Democratic President Clinton reached a similar low to the current President Bush, of 32%, in 1994, the year of the national GOP congressional landslide. Clinton was rated 38% favorably in 1996, and then reached a 47% mark in the 1998, 1999, and 2000 polls.
Other Issues: Race, and Ideological Perceptions
Factors that may be helping Republicans in the presidential race while helping Democrats in the U.S. senate race are ideological perceptions and the role of race. Notice how both Democratic presidential candidates are viewed as significantly more liberal than the average Mississippian, as were Democratic presidential nominees in previous elections. A “Mississippi Democrat” such as Musgrove (studied in previous polls when he was an officeholder) is generally viewed as more moderate, compared to these national level Democrats.
Consequently, there are big racial differences in voting at the presidential level, as Democratic presidential candidates are unable to win more than 28% of the white vote, while racial divisions for the senate are not as strong, since Musgrove is so far drawing 37% of the white vote. Racial divisions are especially great in an Obama-McCain race, where currently 95% of African Americans back Obama while 83% of whites back McCain.
The table below indicates the percentage of each race backing the Democratic candidates (the Republican vote is 100 minus these figures):
|
Election contest: |
Whites |
African Americans |
Racial Gap |
|
Clinton-McCain |
28% |
88% |
60 |
|
Obama-McCain |
17% |
95% |
78 |
|
Musgrove-Wicker |
37% |
84% |
47 |
Question wording: “What about your political beliefs? Do you consider yourself: very liberal, somewhat liberal, moderate or middle of the road, somewhat conservative, or very conservative?”
Question wordings: “Please label the following political figures as very liberal, somewhat liberal, moderate (or middle of the road), somewhat conservative, or very conservative.”
“Democratic Presidential hopeful Hillary Clinton.”
“Democratic Presidential hopeful Barack Obama.”
“Republican Presidential hopeful John McCain.”
Ideological perception questions were not asked for the U.S. senate candidates. However, such questions were asked in previous years’ polls for Musgrove, for when he was lieutenant governor (1998) and governor (2000, 2002 polls). For comparison purposes, we also include the perceptions of previous Democratic presidential candidates, asked in previous Mississippi polls.
The values below are “means” or averages for the ideological variables, all of which are coded as 1 for very liberal, 2 for somewhat liberal, 3 for moderate, 4 for somewhat conservative, and 5 for very conservative.
2008 Mississippi Poll:
Clinton’s perceived ideological mean = 2.21
Obama’s perceived mean = 2.08
McCain’s perceived mean = 3.59
Average Mississippian’s own mean = 3.52
Previous Democratic presidential nominees, 1988-2004 Mississippi Polls:
Kerry = 2.12
Gore = 2.35
Clinton (1996) = 2.20
Clinton (1992) = 2.50
Dukakis = 2.20
Ronnie Musgrove’s perceived ideology in previous Mississippi Polls:
Lieutenant Governor Musgrove (1998 poll) = 3.02
Governor Musgrove (2000 poll) = 2.71
Governor Musgrove (2002 poll) = 2.68
There are some important caveats about the election questions, which should be taken as merely instructive rather than as definitive. Other polling organizations encounter similar problems in modern times.
First, even after weighting for the ratio of number of adults in the household to the number of different household telephone numbers, four demographic groups were undersampled when comparing the sample with census data of Mississippi adults from the 21st century. High school dropouts comprised only 8% of the sample (compared to 22% of the population 25 years of age and older), adults 18-34 years of age comprised only 17% of the sample of adults (compared to 32% of population), men comprised 36% of the sample (47.5% of population), and African Americans comprised 23% of the sample of adults (33% of adult population). The sample had to be weighted four additional times so that the size of each of these groups was within 1% of the population sizes. However, there is no guarantee that members of these demographic groups included in the sample are completely representative of the same demographic group members not included in the sample. Therefore, while we use the fifth weighted variable to achieve a sample that is very representative of these four demographic groups’ presence in the population, there is no guarantee that the sample is as representative of characteristics of the population that are not measured by the census, such as political attitudes and candidate preferences. The proliferation of cell phones exacerbates this problem, as an increasingly number of people, particularly the young adults, have dropped their land lines.
Second, our response rate has fallen to only 40%. Therefore, adults in over half of the households contacted lacked the time or the desire to participate in the survey. We cannot determine whether people in households refusing to participate held similar political views to those who were willing to participate. It is likely that they are less politically knowledgeable, and therefore less likely to vote in November, but they may also be very time stressed because of being two career families having higher incomes.
Third, since this poll was conducted over six months before the November elections, it serves only as a benchmark for voter attitudes today, and cannot be used to forecast the likely election results in November. Furthermore, any predictive value that an early poll may hold has become even less valuable in recent elections. For example, in the 1984 and 1988 presidential elections, the Mississippi Poll’s April results produced a margin within 2 points of the November results. Over the next four, most recent presidential elections, the November results differed significantly from the April responses on three occasions. Republicans had nearly a 2-1 advantage in the April 1992 poll, but won by only a 9 point margin in November. In April 2000, Republicans held only a single digit margin over the Democrats, but won by a 17 point margin in November. The most recent 2004 election saw a divergence in forecasting success depending on how likely voters were defined. A restrictive definition produced a 2-1 April advantage for Republicans, while a less restrictive definition of likely voters produced a 24-point Republican margin in April that was similar to the 20-point November margin. Only in 1996 were the November results very consistent with the April poll responses.
Fourth, the most recent election of 2004 was the only one where how stringent the definition of likely voter was significantly affected the reported strengths of the presidential candidates. In that case, the April results that were closest to the November results were those for the less stringent definition of likely voters. Therefore, because of these most recent 2004 results, and because of the historically low 40% response rate of the 2008 survey, plus the increased need to weight the sample for a greater number of demographic groups, we rely on a less stringent definition of likely voters to report our 2008 results than we have historically used. Likely voters were defined as those indicating that they were at least “probably” likely to vote, were at least “a little” interested in upcoming elections, and were willing to complete the entire survey of 63 questions.