The Impact of Candidate's Sex On Voter Choice

• Published date: 01 March 1981
• Author: Laurie E. Ekstrand,William A. Eckert
• DOI: 10.1177/106591298103400107
• Date: 01 March 1981
• Subject Matter: Articles

THE
IMPACT
OF
CANDIDATE’S SEX ON
VOTER
CHOICE
LAURIE E. EKSTRAND
Westat, Inc., Rockville, MD
WILLIAM A. ECKERT
Arthur D. Little, Inc., Washington, D.C.
AALL
LEVELS of government women are underrepresented in elec-
tive positions. Despite recent increases in the number of public of-
-~. ~
fices held by women, the proportion of women compared to men in
elective office is small and decreases with the status of the officer Among
reasons cited for this is the reluctance of the public to vote for a female
candidate. The results of public opinion polls, and psychological and
sociological research lend support to this proposition by demonstrating the
public’s perception of women as less capable or less desirable as leaders than
men.2 While this does not prove that sex is a factor in voter choice, it does
suggest that such discrimination may exist.
Several recent studies have examined directly the effects of candidate’s
sex on voter choice using actual election returns.3 Among these is a study of
congressional voting by Darcy and Schramm,’ who found that the average
percent of vote received by female candidates when incumbency and party
were controlled was not substantially different from that received by their
male counterparts. The sex of the candidate did not seem to be a factor in
voter choice in congressional contests. The authors conclude that the lack of
significant female representation in Congress is largely a function of the
absence of female candidates rather than sex discrimination at the polls.
Darcy and Schramm’s conclusions are valid only if the average voter
appeal of male and female candidates is identical except for the one variable,
sex. This is difficult to demonstrate considering the different political
socialization and recruitment processes experienced by males and females.
Life experiences (including education, employment history, military service
record, etc.) of the average male and the average female are hardly similar!
If the average female candidate is &dquo;better qualified&dquo; for office than the
average male, then the Darcy and Schramm finding of no difference in the
percent of votes received is actually indicative of discrimination against
female candidates. Conversely, if the males are actually &dquo;better candidates,&dquo;
then equality in votes received indicates bias against the male candidates.
Only when sex of candidate is the single element of difference in electoral
contests can the effects of sex on voter choice be discerned.
NOTE: The authors would like to thank the editors and reviewers of the Western Political Quar-
terly for their helpful comments on an earlier draft of this article.
1
Albert K. Karnig and Oliver B. Walter, "Election of Women to City Councils," Social Science
Quarterly 56 (March 1976): pp 605-13.
2
For examples see Hazel Erskine, "The Polls: Women’s Roles," Public Opinion Quarterly 35
(Summer 1971): 275-90; Myra Marx Ferree, "A Woman for President? Changing Re-
sponses : 1958-1972," Public Opinion Quarterly 38 (Fall 1973): 390-99; Inge K. Broverman,
Susan Vogel, Donald Broverman, Frank Clarkston and Paul Rosenkrantz, "Sex Role
Stereotypes: A Current Appraisal," Journal of Social Issues 28 (1972): 59-78; Phillip
Goldberg, "Are Women Prejudiced Against Women?" Trans-action 5 (1968): 28-30; Irene
H. Frieze, Jacquelynne E. Parsons, Paula B. Johnson, Diane N. Ruble and Gail L. Zellman,
Women and Sex Roles (New York: Norton and Company, 1978), 335-56.
3 For example see Jorgen S. Rasmussen, "The Role of Women in British Parliamentary Elec-
tions,"
Journal
of Politics 39 (November 1977): 1044-54, and Karnig and Walter, "Election
of Women to City Councils."
4
Robert Darcy and Sarah Slavin Schramm, "When Women Run Against Men," Public Opinion
Quarterly 4 (Spring 1977): 1-12.

---

79
In another study, the emphasis was on detecting race and class differ-
ences in support for female candidates. Bernstein and Polly examined
precinct-by-precinct returns from an at-large city council race in Dallas in
1969.5 None of the candidates were incumbents. Two candidates were male;
two female. One candidate of each sex was Mexican American. The other
two were non-Mexican American whites. Precincts were classified as pre-
dominantly black if at least 75 percent of the residents were black, and
predominantly white if at least 75 percent of the residents were white (using
census tract data). Racially mixed precincts were excluded from the analysis.
Once precincts were classified by race, they were further categorized by class
based on the average value of owner occupied homes. Precincts with average
home values less than $15,000 were classified as lower class. Those with
average home values above $15,000 were considered middle and upper
class.
Analysis of the percent of votes received by the female candidates in the
precincts classified by race and class lead Bernstein and Polly to conclude
that regardless of race, residents of lower class precincts voted less for
female candidates than residents of middle and upper class precincts.
Among lower class precincts, women did better in black precincts than in
white. This analysis suffers from the same difficulty as the Darcy and
Schramm research. It cannot be safely assumed that the candidates were
equal except for their sex.
Although Bernstein and Polly’s analysis encounters some difficulty, it
raises a very important point: the significance of candidate’s sex to voter
choice may vary within the population. Some groups may be more likely to
use candidate’s sex as a significant voting cue than others. Bernstein and
Polly’s findings link economic status and race to vote for female candidates.
The previously cited Darcy and Schramm article suggests a relationship be-
tween political party, candidate’s sex, and voter choice. Democratic women,
it appears, do slightly better than their Republican counterparts as congres-
sional candidates.
Other research has also explored this issue. Myra Marx Ferree exam-
ined the relationships between a variety of respondent characteristics and
...