Informal Influences in Selecting Female Political Candidates Les

Transcription

Informal Influences in Selecting Female Political Candidates Les
Informal Influences in
Selecting Female Political Candidates
Les influences informelles
dans la sélection de
femme candidats politiques
Published in Political Research Quarterly 2011
13 Jan 2012
Dr Christine Cheng
Exeter College, Univ. of Oxford
@cheng_christine
christinescottcheng.wordpress.com
Dr Margit Tavits
Washington University in St. Louis
« ...ce sont les femmes dans le parti qui m'a
encouragé à participer plus profondément et
d'enregistrer sur une liste pour les élections
au Parlement cantonal.>>
- Candidate de l'Europe de l’Ouest*
“…it was the women in the party who
encouraged me to get more deeply involved
and to register on a list for the elections to
the cantonal Parliament.”
- Female candidate from Western Europe*
* Politics: Women’s Insight, Inter-Parliamentary Union 2000, p.75. Cited in Women, Politics and Power by Pamela Marie
Paxton, Melanie M. Hughes.
Agenda/Overview
Les informations de base et la question de
recherche
Les mécanismes causals
Les méthodes et les données
L'analyse et les résultats
Les implications politiques
Rang
Pays
% Femmes *
1
2
3
4
Rwanda
Andorra
Seychelles
Sweden
56.3%
53.6%
45.2%
45.0%
5
6
South Africa
Cuba
7
Rang
Pays
% Femmes *
34
Bolivia
25.4%
35
Iraq
25.2%
36
Sudan
25.1%
44.5%
43.2%
37
Laos
25.0%
38
Australia
24.7%
Iceland
42.9%
38
CANADA
24.7%
8
Finland
42.5%
39
Namibia
24.4%
9
Norway
39.6%
39
Vietnam
24.4%
10
Belgium
39.3%
40
Lesotho
24.2%
10
Netherlands
39.3%
41
Lichtenstein
24.0%
41
Poland
24.0%
11
12
13
13
14
Mozambique
Denmark
Angola
Costa Rica
Spain
39.2%
39.1%
38.6%
38.6%
36.6%
*
Chambre unique/basse
Inter Parliamentary Union Database (2010)
www.ipu.org/wmn-f/classif.htm
Les femmes au parlement
Pourquoi est-ce que c’est important?
Différentes priorités politiques
salaire minimum, accès aux avortements, crèches,
horaires flexibles
Les femmes représentent les intérêts des femmes
mieux sur le fond (Tremblay 2000)
Les femmes au parlement
Comment pouvons-nous expliquer la
proportion de femmes au parlement?
Les facteurs structurels: le système électoral, l'idéologie du parti,
les quotas de genre, etc.
Au Canada: Un manque de soutien par les partis politiques et un
environnement politique hyper-masculin
Aucune différence entre hommes et femmes dans l’argent amassé
pour les campagnes
Dans les démocraties avancées, les électeurs ne soient pas
discriminatoires entre les candidats hommes et femmes
Les influences informelles
Processus de sélection des MP
Les candidats potentiels  La sélection interne du parti
 Les candidats  La Campagne  Le représentant élu
Selecting local candidates
Selection takes place at local level, minimal
outside intervention
Candidates are chosen at local nomination
meetings
Voted upon by local party members if more than
one person is running
Local party executives are key in both contested
and uncontested nominations
Research Question
We already know that women:
Are more likely to see themselves as “not qualified”
Are less likely to be recruited, encouraged to run, be promoted
from within party
BUT If asked by gatekeepers, more likely to consider candidacy,
eliminating much of the gender gap
Gatekeepers are important!
Does the gender of party gatekeepers play a role in local
candidate recruitment?
Hypothesis: Local female party presidents are more
likely to field women candidates in their constituency
Relevant to other countries, other levels of government
Causal Mechanisms
Like recruits like
More likely to recruit & support people like
themselves (Caul 1999; Kunovich and Paxton 2005; Huddy 1994; Niven
1998; Sigelman and Sigelman 1982; Erickson 1997; But see Lovenduski and
Norris 1993; Tremblay and Pelletier 2001)
Social networking effect (Dow and Wood 2006)
Role Model effect (Beaman, Duflo, Pande, Topolova 2012)
Message: Women can succeed
Signalling effect
Welcoming environment; women are valued; “women
friendly” (Lawless and Fox 2005; Trimble and Arscott 2003; Costantini
1990; but see Carroll 1985)
Similar studies
Tremblay and Pelletier 2001
No evidence that Canadian female local party leaders
are more women-friendly
Niven 1998
Male party leaders preferred candidates like
themselves
Revealed preference
Political correctness
Unaware of bias
Gendered networks
Chocolate vs. carrots analogy
Methods & Data
Dependent variable: gender of candidates at constituency
level
Independent variable: gender of party presidents at
constituency level
Bloc, Liberals, Conservatives, NDP, Greens
Controls: % women 18+, Share of female candidates in riding
since 1980, Share of female party candidates, district
competitiveness, district ideology, party ideology, incumbency,
contested nomination, % English speakers, % college grads, %
Catholics, avg. household income, region
2004 & 2006 Election Data
Female
presidents
Female
candidates
N
Conservatives
20%
14%
261
Liberals
23%
25%
614
Bloc
30%
27%
92
NDP
30%
33%
515
Greens
23%
25%
158
Total
25%
26%
1651
Effect of local party president’s gender on candidate’s gender
Probit
coefficient
(SE)
0.174**
Contested nomination
0.081
Incumbent
-0.069
0.117
0.0166
0.087
Share of English
speakers
0.001
0.003
District ideology
0.002
0.003
Share of women 18+ in
district
0.063**
0.018
Average household
income
0.001
0.018
0.029
Share of female
candidates since 1980
0.968**
Share of female party
candidates since 1980
0.972***
Competitiveness
Share of college
graduates in
district
0.003
0.426
Share of Catholics
0.188
0.0001
0.003
-0.0002
0.003
Year 2006
0.066
0.073
Note: Dependent variable: Female Candidate.
**p ≤ .05, ***p ≤ .01
Female president
Probit
coefficient
(SE)
Effect of local party president’s gender on the candidate’s gender
Probit coefficient (SE)
Region dummies
West
0.025
0.171
East
-0.107
0.203
North
0.678
0.458
Ontario
-0.019
0.162
Party dummies
Bloc
-0.06
0.303
Conservatives
-0.368**
0.171
Liberals
-0.0005
0.149
NDP
0.132
0.136
Findings
Probit regression
Use predicted probabilities to understand results
Gender of local party president matters
6% more likely to have a female candidate if local party
president is female
Party, Affirmative Action, % Share of women 18+ in district
statistically significant
District ideology, District Competitiveness, Incumbency not
significant
Other control variables not significant
Correlation doesn’t imply causation, but the relationship is
robust!
Surprising Findings
Share of female candidates in district since 1980
matters
18% difference
Share of female party candidates in district since
1980 matters
34% difference
Exposure Effect, Demonstration effect
Policy Implications (Based on this study)
Gender composition of the candidate selectorate
(gatekeepers) is correlated with gender of
candidates
If we want more female candidates in the future:
Recruit more women into local party executives
Field more female candidates today!
Target constituencies with fewest female
candidates
Riding Competitiveness