Transcript Exercise 7 - People Server at UNCW

(EXERCISE 7)
THE IMPACT OF GENDER
ON 2004 PRESIDENTIAL VOTE CHOICE
CONTROLLING FOR MARITAL STATUS
Roger C. Lowery
PLS 401, Senior Seminar
Department of Public & International Affairs
UNC Wilmington
3 March 2009
1
Bivariate Hypothesis
• Theory:
– Men are more likely than women to support the use of
force in domestic and foreign affairs.
– This gender gap, since the Vietnam Conflict, has
produced higher support for Republican candidates
among men than women voters.
– By 2004, the Iraq and Afghan Conflicts had
overshadowed Bush’s earlier compassionateconservative image among women voters.
• H1: Men were more likely than women to favor Bush in
the 2004 presidential election.
3 March 2009
2
Table 2:
2004 Presidential Vote Choice by Gender
Gender
Cells contain:
-Column percent
-Weighted N
2004
Presidential
Vote Choice
1
Male
2
Female
ROW
TOTAL
1: Bush
53.9
204
46.8
192
50.2
395
2: Kerry
46.1
174
53.2
218
49.8
392
COL TOTAL
100.0
378
100.0
409
100.0
787
Summary Statistics
Tau-b =.07
Chi-square probability = 0.04
3 March 2009
3
Bivariate Findings
• H2 (men were more likely than women to favor Bush in the
2004 presidential election) is supported by the sample data
in Table 2 (because both necessary & sufficient conditions are
observed):
1.
The gender-gap voting pattern predicted by H2 is
observed in the sample data (although the taub of 0.07
indicates that the gender gap was weak in the sample
voters).
2.
In addition, this sample finding is statistically significant
and thus can be extrapolated to the national electorate.
The χ2 probability of random-sampling error is less than
0.05 (it is 0.04).
3 March 2009
4
Multivariate Hypothesis
• Theory:
– Marriage tends to attract socially and politically
homogeneous spouses.
– Marriage tends to produce socially and politically
homogeneous spouses.
– Therefore, the strength of the gender gap in
voting is conditional upon the marital status of
voters.
• H3: the impact of gender on presidential vote choice
is stronger within non-married than married voters.
3 March 2009
5
Table 3a
2004 Presidential Vote Choice by Gender
(Single, divorced, or separated)
Gender
Cells contain:
-Column percent
-Weighted N
2004
Presidential
Vote Choice
1
Male
2
Female
ROW
TOTAL
1: Bush
47.2
57
33.8
46
40.1
103
2: Kerry
52.8
63
66.2
91
59.9
154
COL TOTAL
100.0
120
100.0
137
100.0
258
Summary Statistics
Tau-b =.14
Chi-square probability = 0.02
3 March 2009
6
Table 3b
2004 Presidential Vote Choice by Gender
(Married or widowed)
Gender
Cells contain:
-Column percent
-Weighted N
2004
Presidential
Vote Choice
1
Male
2
Female
ROW
TOTAL
1: Bush
57.0
147
53.3
145
55.1
292
2: Kerry
43.0
111
46.7
127
44.9
238
COL TOTAL
100.0
258
100.0
272
100.0
530
Summary Statistics
Tau-b = .04
Chi-square probability = 0.41
3 March 2009
7
Multivariate Findings
• H3 (the impact of gender on presidential vote choice is
stronger within non-married voters than married and widowed
voters) is supported by the sample data (because both
necessary & sufficient conditions are observed):
1.
The strength of the gender gap did change as predicted in
the partial-table subgroups: the taub was moderately strong
(0.14) within the non-married (single, separated, or
divorced) voters and weak (0.04) among married and
widowed voters.
2.
The gender gap was statistically significant within the nonmarried voters (χ2 = 0.02) but not the married and widowed
voters (χ2 = 0.41).
3 March 2009
8
Substantive Implications
• Although weaker than in previous presidential
elections, the gender gap was still present in 2004 –
but only in non-married voters.
• Because: 1) there is an increasing number of nonmarried voters and 2) there are more female than
male voters, the gender gap is a not-insignificant
problem for the Republican Party to address.
• However, because the current Democratic president
is faced with both the Iraq and Afghan wars, his
success or failure in resolving those conflicts will
impact the gender gap in his support.
3 March 2009
9
Methodological Implications
• Because the future of both parties lies in
the voting loyalties of younger voters, the
impact of period and life-cycle effects on
the gender gap in voting needs to be
analyzed.
• Because sectional divisions in party
loyalties have increased in recent years,
sorting out regional effects on gender-gap
voting needs to be examined.
3 March 2009
10