Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
APPROVED: Regina Branton, Major Professor Valerie Martinez-Ebers, Committee
Member Elizabeth Oldmixon, Committee Member Tony Carey, Jr., Committee Member Jae-Jae Spoon, Committee Member Matthew Eshbaugh-Soha, Chair of the
Department of Political Science David Holdeman, Dean of the College of
Arts and Sciences Victor Prybutok, Vice Provost of the
Toulouse Graduate School
RUN, WOMEN, RUN! FEMALE CANDIDATES AND TERM LIMITS:
A STATE-LEVEL ANALYSIS
Samantha Pettey
Dissertation Prepared for the Degree of
DOCTOR OF PHILOSOPHY
UNIVERSITY OF NORTH TEXAS
August 2016
Pettey, Samantha. Run, Women, Run! Female Candidates and Term Limits: A
State-Level Analysis. Doctor of Philosophy (Political Science), August 2016, 114 pp., 19
tables, 8 figures, chapter bibliographies.
This dissertation seeks to explain the puzzle in the state politics literature which
expects females to benefit from the enactment of term limits, but initial research finds
the number of female in office decreases after the implementation of term limits.
Examining this puzzle involves three separate stand-alone chapters which explore female
candidate emergence (1), success rates (2), and women-friendly state legislative districts
(3). The goal of the dissertation is to reconcile the puzzle while adding insight into how
female candidates behave at the state-level. Overall, I find that term limits increases
female descriptive representation by increasing the likelihood a female candidate will run
and win an election.
Copyright 2016
by
Samantha Pettey
ii
ACKNOWLEDGMENTS
I thank my family, especially my parents, for being the greatest pair of role models.
Also, for your unending support, faith, and love throughout my education, and life. To my
sisters Amanda and Rebecca: thank-you both for your support, and humoring my love of
politics since an early age by playing ‘political radio talk’.
I want to thank Regina Branton for her mentorship and guidance throughout my
graduate career and the dissertation process. I am also indebted to Jae-Jae Spoon, Elizabeth
Oldmixon, Valerie Martinez-Ebers, and Tony Carey Jr. for all the comments, feedback, and
thought-provoking questions, which significantly improved earlier drafts of this dissertation.
I also want to thank friends for understanding, and supporting this long journey.
And to friends here at University of North Texas: thank-you for creating a productive and
enjoyable environment. Many of you helped me along the way with prioritizing, feedback,
support, and laughs.
iii
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
LIST OF TABLES v
LIST OF FIGURES vi
CHAPTER 1 INTRODUCTION 1
CHAPTER 2 FEMALE CANDIDATE EMERGENCE 9
CHAPTER 3 FEMALE CANDIDATE SUCCESS 41
CHAPTER 4 WOMEN-FRIENDLY DISTRICTS 71
CHAPTER 5 CONCLUSION 110
iv
LIST OF TABLES
Page
1.1 Term-Limit Bans 4
2.1 Term-Limited States 24
2.2 Descriptive Statistics for Female Emergence and Term Limits 25
2.3 Distribution of Candidates and Term Limits 26
2.4 Difference in Differences 30
2.5 Likelihood Candidate is a Female 31
2.6 Predicted Probabilities for Democrats 33
2.7 Predicted Probabilities for Democrats in Open Seats 33
2.8 Predicted Probabilities for Republicans 34
2.9 Predicted Probabilities for Republicans in Open Seats 34
3.1 Descriptive Statistics for Female Success and Term Limits 56
3.2 Descriptive Statistics for Male Success and Term Limits 56
3.3 Logistical Regression of Democratic Candidate Success 59
3.4 Predicted Probabilities for Female Democrats 60
3.5 Logistical Regression of Republican Candidate Success 63
3.6 Predicted Probabilities for Female Republicans 65
4.1 State Legislative District Descriptives 86
4.2 Logistical Regression of Female Candidate Emergence 89
4.3 Logistical Regression of Female Success 93
v
LIST OF FIGURES
Page
Figure 3.1. Female Democrat Success 61
Figure 3.2. Female Republican Success 64
Figure 4.1. Core Democratic Female Emergence 90
Figure 4.2. Core Republican Female Emergence 91
Figure 4.3. Swing Democratic Female Emergence 92
Figure 4.4. Swing Republican Female Emergence 92
Figure 4.5. Core Democratic Female Success 95
Figure 4.6. Core Republican Female Success 95
vi
CHAPTER 1
INTRODUCTION
Term limits are one of the most significant institutional changes to take place since
the modernization of state legislatures (Kurtz, Cain and Niemi 2007). Term limits are an
institutional change that mainly started as citizen initiatives in the states. Citizens were
unhappy with their career politicians (at the state and Congressional-level) and wanted
term limits in order to remove longtime politicians from office. Since the implementation
of term limits is fairly new, continuing to study the effect they have on state legislators is
important.
Studying the effect term limits have on female state legislators is arguably the most
important contribution this dissertation makes to the field. Studying women in office is
important since females are underrepresented in government. Despite making up an equal
portion of the population, women do not have political parity. Descriptive representation is
important since studies find women politicians bring a different viewpoint and style to the
legislative table. Further, this difference is significant because scholars find women politicians
are better at politically representing women (Swers 2002, Osborne 2012). If term limits have
an impact on descriptive representation, positive or negative, the consequences are important
to understand and study.
Below, I provide a brief discussion of the term limit movement in the U.S. Then, I
followup with an example from a state, Florida, to show how the dissertation tries to explain
the positive effect term limits seem to have on females. Lastly, I provide a roadmap for each
substantive chapter by briefly discussing the theory and findings.
1
Term Limits
Generally speaking there are three main arguments supporters of term limits use
to emphasize why term limits are good. The first deals with elections, the second with
politicians and the last argument relates to lobbyists and special interest groups. The first
argument claims elections will become more competitive if there are term limits. The ratio-
nale comes from the advantage incumbents receive when running for reelection. Incumbents
have name recognition and usually more money at their disposal, which helps better their
chances of winning reelection. In the official voter guide, supporters of the term limits in
California described state legislative turnover as less than The British House of Lords and
the Soviet Legislature (Ballotpedia 2016).
On the other hand, opponents of terms limits argue by limiting the term of politicians,
citizens are losing their right to choose. If a politician is successful at their job and popular
with constituents, taking this person out of the election is bad for the democratic process.
Term limits can prevent the best, or most preferable candidate from seeking office again.
The second argument relates to elections and the types of candidates in the race.
Those in favor of term limits argue career politicians are bad. Therefore, term limits can
help eliminate career politicians and bring in a different type of legislator— one who is
a citizen, not a politician (Price 1992). Supporters of term limits argued taxpayers and
consumers are ruled by a small, elite class. And what is best for democracy is a ruling class
of people, by the people (Ballotpedia 2016).
Opponents argue opening up the pool of candidates is not a good thing because term
limits incentivize the wealthy to run. Term limits incentivize the wealthy to run since being
a legislator in a term limited state is only a temporary job. Most people cannot afford to
2
leave their job to work a few years, but a person who is wealthy may be able to temporarily
leave their job. Therefore, the new politicians will be wealthy and ‘out of touch’ with the
majority of their constituents.
The last argument supporters make is that special interests have too much control
over politicians. This is bad for the democratic process because the people are ignored.
Supporters argue that term limits will stop the close relationships politicians have with
lobbyists. This would force politicians to do what is best for their constituents and not what
is best by special interests.
On the other hand, opponents of term limits argue the opposite—state legislators will
become more reliant on special interests and lobbyist because the lobbyist are constant and
more familiar with the system. The argument is term limits create so much turnover that
many norms within the chamber are lost and committee chairs become less qualified. Since
legislators lack experience, they are forced to rely on lobbyists.
Consecutive versus Lifetime Bans
The limits placed on legislators varies by states and are either consecutive or lifetime
bans. The latter is much more strict on legislative terms and does not allow legislators to
serve in office after a maximum year is reached. For example, California, enacted term limits
in 1990 as a lifetime ban on legislators. In California, a legislator can only serve up to twelve
years in either chamber. The time can be split between both chambers or served all in one
but a legislator cannot serve more than twelve years.
Consecutive bans are generally less strict than lifetime bans. In general, the limit on
the legislator is for a certain number of years. Most states do this by chamber; i.e. serve
eight consecutive years in the lower chamber then the legislator is allowed to serve another
3
eight in the upper chamber (or vice versa). Different from a lifetime ban, after a set time
period (usually two years) away from the legislature, the legislator can run again. If the
legislator runs again and wins, the clock is reset. Table 1.1 below lists all the states with
their ban limit in years as well as the type.
Limit in Years Consecutive Lifetime
6 House / 8 Senate – MI8 Total NE –8 House / 8 Senate AZ, CO, FL, ME, MT, OH, SD MO12 Total – CA, OK12 House / 12 Senate LA NV16 Total – AR
Source: NCSL 2014: ‘The Term-Limited States’
Table 1.1: Term-Limit Bans
Florida
The 2016 Florida State Legislature is ranked 25th amongst all 50 states in terms of
the proportion of women serving in the legislature (both chambers). Females make up 25%
of the legislature; 23% in the lower chamber.
In 1992 Florida voters overwhelming supported term limits on their state officials.
76.8% of voters were in favor of a ballot initiative making a constitutional amendment to
impose term limits. A statewide campaign from term limits supporters used the slogan
“Throw the Rascals Out” and “Eight is Enough”. The Florida Constitution now states that
legislators (and cabinet members) are limited to a term of eight years in a given chamber.
The limits in Florida are consecutive and therefore once a legislator is term out of one
chamber they can run for election in the other chamber and serve another eight years.
Sen. Rene Garcia (R) recently said, “We are a representative democracy and we
should be making sure that it is the elected officials who move agendas forward, and not the
4
lobbyists,” (Miami Herald, 2015). A concern by the legislators in Florida is the amount of
power unelected persons have on the political process. In particular, opponents such as Sen.
Garcia argue special interests donate to campaigns and have more control over the agenda
and policy-making in the legislature.
Further, opponents argue there are vast issues and changes that have taken place,
not for the better, on the daily operations and functioning of the chambers.
SJ 0902 and HJR 711: Term Limits, were jointly filed on 11/18/2015 in the Florida
Senate and House, respectively. The bills seek to amend the state constitution and increase
the number of years a legislator can serve from eight to twelve years. Thus far there has been
no floor action on either of the bills. Yet, this is not the first time legislators have tried to
extend terms from eight to twelve years. In 2005, both chambers overwhelming supported,
and passed, a legislatively referred constitutional amendment (LRCA) to appear on the
2006 ballot (Ballotpedia 2016). The ballot measure was repealed before the election when
voters began calling their representatives because they were unhappy with the possibility of
extending legislative term limits. Ultimately, the proposal never made the 2006 ballot so the
actual sentiments of voters can only be presumed (Tallahassee Democrat, 2006).
Female Descriptive Representation in Florida
The last election year before term limits took effect in Florida was 1998 and 50 females
ran for office in the lower house. In 2000, the first year of impact, 61 females ran for office.
Statewide, term limits created 83 vacancies in the 120 person house. While the number of
females emerging for office increased in the first year term limits took effect, the number
of female candidates emerging since the initial year has decreased and ranges from 44-53
females running in a given election year. Pre term limits, the number of females running
5
for the office ranged from a low of 35 in 1990 to a high of 56 in 1996. The mean percent of
female candidates pre term limits is about 25% and post term limits this number increases
to 27%. The emergence chapter discusses my theory as to why there is a post term limit
increase in the number/percent of women running for office.
In terms of female descriptive representation, the number of women in the lower
chamber pre and post term limits has remained fairly stagnant. Yet, as Chapter 2 discusses,
party plays an important role. Term limits create an environment favorable to Republican
candidates. After term limits were implemented in Florida, Republican females nearly double
their numbers in office. For example, in 1990, 8 Republican women held office and this
number increases to 18 in 2000. On the other hand, female Democrats were hurt by term
limits. Pre term limits, female Democratic numbers were consistently averaging 20 a year
but post term limits, the number of women in office drops to an average of 15 a year.
Dissertation Chapter Outline
Chapter 1 explores how term limits effect the eligibility pool of candidates running
for office. The main research question is “Do term limits influence the emergence of female
state legislative candidates?” Examining the number of women running for office is important
since descriptive representation cannot happen if there is not a substantial pool of female
candidates willing to run. I argue term limits have a positive effect on the number of females
running for office since term limits change the incentive structure for running. Since term
limits place a limit on time in office, candidates, specifically women, who do not want to be
career politicians may consider running for office when term limits are present. I find, when
compared to states without term limits, females are more likely to run in states with term
limits.
6
Chapter 2 is complimentary to the current literature on term limits and females. I
look at the question of “Do term limits help or hinder female descriptive representation in
state legislatures?” Most studies on the effect term limits have on women post term limits are
limited in scope due to lack of time. I reexamine female descriptive representation and look at
the effect term limits have on women over multiple time periods. I argue professionalization
of the legislature matters and that term limits effect female Republican and Democrats
differently. I find term limits have a positive effect on female Republicans and a negative
effect on female Democrats. Female Republicans are more successful at winning in term
limited states while female Democrats are more successful at winning in non-term limited
states.
Chapter 3 is a combination of chapters 1 and 2. I answer the research questions,
“Do certain district demographics favor female candidates and How do term limits affect
the women-friendliness of a district?”. Building off work looking at district demographics, I
explore the set of characteristics at the state-level that make a district more women-friendly.
I find districts that are more urban, diverse, democratic, educated, wealthy, and have more
women in the workforce are more likely to be areas where females emerge and win. I also
find term limits do not have a statistically significant impact on the women-friendliness of a
district.
7
[1] Ballotpedia. 2016. “Florida 2006 Ballot Measures”. Retrieved on March 6, 2016 from:https://ballotpedia.org/Florida_2006_ballot_measures.
[2] Cannon, Steve. 2015 “Florida Legislators Propose Increasing Term Limits” Retrieved on March6, 2016 from: http://www.miamiherald.com/news/politics-government/state-politics/article45621033.html#storylink=cpy.
[3] Carey, John M., Richard G. Niemi, and Lynda W. Powell.2000. Term limits in state legislatures.University of Michigan Press.
[4] Deslatte, Aaron. 2006. “Term-Limit Extension Removed from Ballot” Retrieved on March 7, 2016from: https://www.aproundtable.org/news.cfm?NEWS_ID=1269&issuecode=term.
[5] Osborn L, Tracy. 2012. How Women Represent Women: Political Parties, Gender and Representationin the State Legislatures. Oxford University Press.
[6] Price, Charles M. 1992. “The Guillotine Comes to California: Term-Limit Politics in the Golden Sate.” inLimiting Legislative Terms meds Gerald Benjamin and Michael J. Malbin. Washington, D.C.: CQ Press.
[7] Swers, Michele. 2002. The Difference Women Make. Chicago: University of Chicago Press.
8
Chapter Bibliography
CHAPTER 2
FEMALE CANDIDATE EMERGENCE
Abstract
This chapter examines the effect of gender differences on emergence rates in state
legislative elections as a function of term limits. Initial research finds a negative relationship
between term limits and female representation. I offer a candidate-level theory and empir-
ical approach to reevaluate how term limits impact female representation overtime. Using
candidate data in all 50 states from 1990-2000, I find that females are more likely to run for
office in states that implement term limits.
Introduction
The lack of females in office has long been the subject of normative concern for
scholars. Theories on descriptive representation claim the makeup of legislative bodies should
reflect the demographics of the public. Politicians better represent their constituents if they
resemble the populace’s gender, race, social and economic status. Scholars theorize the
importance of descriptive representation and its implications for society because females
bring different issues, experiences and viewpoints to the political table (Swers 2002; Swindt-
Bayer and Mishler 2005; Osborn 2012). For example, Schwindt-Bayer and Mishler (2005)
find legislation will be more favorable towards women with a more representative sample
of females in Congress because male politicians rarely offer the same types of legislation
as females, especially those particularly relating to women’s issues. The logic follows that,
increasing the percentage of women in politics leads to more favorable policies for women
(Swers 2002; Schwindt-Bayer and Mishler 2005; Osborn 2012).
In the United States, female descriptive representation has been, and continues to be,
9
much lower than male descriptive representation at all levels of government. Currently, at
the state level, female representation ranges from a low of 12% in Louisiana to a high of 41%
in Colorado. The average percent of females in all 50 state legislatures is only 24.2% (CAWP
2014; NCSL 2014). Not only is the current percent low but the percentage of females in state
legislative branches has become rather stagnant. Since the 1990s there has not been a large
growth in the percent of females in office (Norander and Wilcox 2012). Given that a large
amount of policy-making is delegated to the states, the lack of female representation across
all states is troublesome (Carroll 2013).
Studying female candidates at the state level is important for furthering our overall
understanding of women and politics. More specifically, research providing a greater under-
standing of when and why females run for office, or emerge as candidates, is a critical stage
in the process towards understanding the unequal representation between males and females
at all levels of government.
I intend to add insight into existing institutional theories on the underrepresentation
of women in office by examining the impact of term limits on female candidate emergence
in state legislative general elections (Bernstein 1986; Ferry 1994; Carroll and Jenkins 2001).
Conventional wisdom suggests open seats will help females gain descriptive representation
since females fair as well as men in elections. Term limits provide more open seats which
suggests females have more opportunities to enter into office (Burrell 1994; Darcy, Welch and
Clark 1994; Fox 2000; Seltzer et al. 1997). Yet, studies conducted after the implementation
of term limits find the greater number of open seats actually had negative consequences for
female descriptive representation (Carroll and Jenkins 2001, Bernstein and Chada 2003).
Here, I revisit this research to try and bridge the puzzling findings resulting in a negative
10
impact on overall female numbers in office. I directly test whether or not term limits have a
causal impact on women running for office using a pre-post experimental design. I find term
limits have a positive impact on the number of females running for office. States with term
limits are more likely to have females running in elections.
This paper proceeds as follows. First, I provide a discussion of the extant literature
in regards to female candidate emergence and term limits. Next, I theorize my expectations
on how term limits will impact female candidate emergence. Then, I introduce the data and
methods used to test my hypothesis and conclude with a short discussion of the findings and
implications of this study.
Previous Literature
Generally speaking, the majority of research on women and politics is conducted at
the national-level. Implications from these studies are not directly transferable to state-
level research because female descriptive representation varies across state. As previously
discussed, some states have a legislative branch with fairly equal percentages of men and
women while others are composed of a super-majority of men. There are two areas of research
I will address below; emergence and term limits. Both fields of study have progressed without
taking much consideration into how one may influence the other. Below is a brief discussion
on the current state of each literature.
Emergence
A particular thread of the women and politics literature focuses on how and why
women run for office (Fox and Lawless 2004; Fulton 2004; Carroll and Sanbonmatsu 2013).
In other words, why do some female candidates emerge and run for election while other
potential candidates do not? Before there can be parity, there have to be females who are
11
willing to run for office and ultimately win their campaign. As previously discussed, female
representation at all levels of government is minimal at best and the lack of female emergence
as candidates is one explanation for the gender disparity reflected in government.
There are four main factors scholars identify that inhibit the emergence of female
candidates. One explanation points to institutional factors acting as barriers to emergence.
The second looks at socialization and sociological factors that may influence a females’ chance
at running for office. A third approach examines ambition levels of potential candidates. The
fourth approach examines political factors that impact who runs for office. Each approaches
hinges on the assumption that females face greater challenges and bear greater costs when
seeking political office.
Institutional Barriers
From an institutional perspective, the low rates of women descriptive representation
in the United States are explained by two main theories: the incumbency advantage and
eligibility pool. Incumbency advantage posits that incumbents’ high reelection rates reduce
opposition and leave little opportunity for new candidates to compete (Carroll and Jenkins
2001; Darcy, Welch, and Clark 1994; Jacobson 2000; Burrell 1992). Therefore, since males
already hold the majority of seats, coupled with the fact that winning reelection is generally
easier than winning as a challenger, males will continue to hold a majority of seats. As such,
females will enter office at a glacial pace because the incumbency advantage is so significant.
Furthermore, research finds female candidates at the congressional level to be just as likely
to win elections as men; indicating that females are not at an electoral disadvantage per se,
but the high incumbency rate continues to keep females out of office (Burrell 1994; Darcy,
Welch, and Clark 1994; Fox 2000; Seltzer et al. 1997).
12
A second institutional constraint is the female candidate eligibility pool is smaller
than the male candidate pool. The eligibility pool, or where the majority of candidates
emerge from, consists mainly of positions in law and business.1 Women in Congress tend to
emerge from careers in teaching, social work, and healthcare without a background in politics
rather than the typical ‘political’ fields previously mentioned (Clark 1994; Dolan, Deckman
and Swers 2007; Carroll and Sanbonmatsu, 2013). The assumption is there are more males
in elected office because males occupy a majority of the jobs from which candidates emerge
(Fox and Lawless 2004). Gertzog (2002) notes the female eligibility pool is changing and an
increasing number of females congressional candidates have previous office-holding experience
and backgrounds in law and business. Females emerging at the national level with previous
office-holding experience generally progress from state legislatures. Therefore, an analysis of
state legislatures will further research at multiple levels of government.
Sociological Factors
Second, extant literature suggests there are sociological, cultural norms at play which
keep females from seeking office. First, women are less likely to be socialized to run for
political office (Fox and Lawless 2004). In other words, males are conditioned early on and
are encouraged more often than females to run for office. Political conditioning research also
suggests that women who are engaged in politics are more likely to emerge and run for office
at any level (Welch 1977; Fox and Lawless 2004). Furthermore, women are more likely to be
engaged in politics if issues are salient to them (Campbell and Wolbrecht 2006). Therefore,
an increase in descriptive representation may potentially lead to more women’s issues being
discussed, which will then influence more females to be actively engaged and interested in
1Men also receive more encouragement than females at a younger age to enter fields that are more likely tolead to political careers such as law and business (Fox and Lawless, 2004)
13
politics which may ultimately increase the likelihood of females running for office.
Furthermore, women tend to have a disproportionate share of family responsibilities
(Fox and Lawless 2004). Not only that, women see private, family commitments as conflicting
with public commitments and this has changed little overtime (Lee 1976, Saprio 1982; Burrell
1994; Carroll and Sanbonmatsu 2013). It is important to note these cultural gender norms
that seem to place time constraints on future candidates do not seem to inhibit other areas
of female political involvement; women are just as, if not more, likely to be engaged in local,
community-level politics. Therefore, sociological based time constraint theories cannot be a
sufficient explanation to a lack of female emergence (Beckwith 1986; Bennett and Bennett
1989; Burns Schulman and Verba 2001).
Ambition
Third, research finds ambition levels differ by gender; specifically, women’s ambition
levels are inherently lower than males. The gender differences in political ambition levels
is in part due to the fact that women perceive themselves to be less qualified than men
to run for office (Fox and Lawless 2004; Fulton et al. 2006). The most extensive research
on candidate ambition is ‘The Citizen Political Ambition Study’, conducted by Fox and
Lawless (2004). The study targets males and females who are in positions that tend to lead
to political careers, i.e. potential candidates in the elligibility pool. The survey targeted
ambition levels by asking questions of qualification, desire to run for office, etc. One of the
conclusions of the survey is ambition is the sufficient explanation for low numbers of females
in office because within the eligibility pool of potential candidates, females are not willing
to run for office due to low levels of political ambition and this pattern has not changed
overtime (Fox, Lawless and Feeley 2001; Fox and Lawless 2004).
14
Political Factors
Last, there are political factors that influence whether or not a person will run for
office. Research finds parties have a large and significant impact on whether a female will
run for office. Most female candidates do not consider running for office until they are
approached by a party leader. Without recruitment, the majority of females had never
seriously considered running for office. There are more women in office when women are
actively recruited; whether by party or political organization (Sanbonmatsu 2006; Carroll
and Sanbonmatsu 2013). And while some literature finds women are less likely to be recruited
(Fox and Lawless 2004), an increase in party recruitment of females, should increase the
overall number of females in office.
Term Limits
Term limits are an important institutional attribute that may serve as a bridge be-
tween emergence and descriptive representation by creating more open seats. Term limits
in state legislatures place restrictions on the number of times an official can serve in office.2
States began enacting term limits in the early 1990s as a way to force government turnover.
Many citizens were displeased with the lack of effective governance at the state level and
wanted to stop incumbents from running for office (Reed and Schansberg 1995; Thompsson
and Moncrief 1993).
While states with term limits enacted a different set of limits upon their legislators,
the main premise is that an incumbent cannot continue running for office term after term.
For example, Oklahoma has a twelve-year lifetime limit across both chambers. Therefore,
2Since the 1990s, 21 states have passed legislation for term limits but 6 states have repealed them. Termlimits were generally passed via citizen initiatives and in 4 states (Massachusetts, Oregon, Washington andWyoming) the state supreme courts repealed the term limits and in 2 states (Idaho and Utah) the statelegislature repealed the term limits (NCSL 2014)
15
if a person runs and wins election in the lower chamber and serves for twelve years, they
are unable to hold office in the upper chamber. Other states, such as California, imposes
term limits for each chamber allowing a politician to term out of one chamber (the act of
no longer being able to run for office) and then run in the other chamber until they are
termed out there. There is also another variation in term limits where the limits are only
set consecutively. In other words, in a state like Colorado, a person can hold office for eight
years in the lower chamber an then take an election period off and then run and win their
seat back for another eight years (Farmer, Rausch and Green 2003; NCSL 2014).
Prior to the enactment of term limits, most researchers hypothesized term limits
would be beneficial for females. Term limits act as a way to combat the incumbency ad-
vantage and create open seat elections. These open seats would allow females to have more
opportunities to run for office. Overall, there was wide consensus that term limits should
have positive effects on female descriptive representation (Carroll and Jenkins 2001; Darcy
Welch and Clark 1994; Ferry 1994; Crane 1995; Thompson and Moncrief 1993).
Yet, when term limits began to take effect, researchers found term limits did not seem
to have the intended effects on women in office (Carroll and Jenkins 2001; Bernstein and
Chada 2003). In fact, in 1998, seven states had their first round of termed out politicians
which opened up 215 seats across all seven states. Women before being termed out held 58
of those 215 seats and post-1998 election only 53 women held seats. In reality, term limits
actually hindered women’s overall numbers in state legislatures because there was a net loss
of 5 seats for women (Bernstein and Chada 2003).
In a cross-sectional analysis of the impact term limits had for the year 1998, scholars
find an overall negative impact on females but there was variation by state. Some states,
16
such as Arizona and Arkansas, saw minor gains in the number or females in office but most
suffered, which reflected in a declining number of females in office (Carroll and Jenkins
2001). Yet, what explains these differences across states? Further research on descriptive
representation and term limits has been limited since the initial term limit waves took place.
Given that there is limited research on the long term impact of term limit effects, further
research is needed to resolve the conflicting evidence that would suggest term limits help
females but in practice they seem to be another disadvantage for females.
Emergence and Term Limits Theory
The literature on candidate emergence and term limits offer conflicting explanations
on the expectations of female emergence. On the one hand, the emergence literature finds
there are numerous barriers inhibiting women from running for office such as family respon-
sibilities, incumbency advantage, lack of political recruitment and ambition as well as a lack
of females in the eligibility pool. On the other hand, it is thought open seats created by
term limits are the perfect opportunity for females to gain office. Initial studies did not
find evidence to support the term limit assumptions but scholars posited there may still be
hope that term limits could have positive implications on female descriptive representation
(Carroll and Jenkins 2001; Bernstein and Chada 2003).
Here, I develop a theoretical argument reconciling these two areas of literature and
the puzzle that term limits create open seats, but have not increased the overall number
of females in office. The term limits literature mainly focuses on overall success of female
candidates and largely ignores the impact term limits may have on candidate emergence.
Perhaps women are running for office more frequently in the term limited states, but are not
winning elections. I incorporate expectations about open seats from term limits literature
17
and eligibility pool assumptions from the emergence literature to develop a general argument
regarding female candidate emergence in term limited states. I expect open seats help females
gain office, but the initial impact of term limits negatively impacted female descriptive
representation. Specifically in this chapter, I argue that over time term limits influence
female candidate emergence.
Open Seats
Existing research regarding female descriptive representation is concerned with aggre-
gate levels of female success. As discussed, a critical stage of analysis is candidate emergence.
The only way females will win office and gain parity is if they run. All else being equal,
in terms of electoral outcomes, term limits can only be effective if the number of female
candidates running within the states increases. For example, if there are more open seats
(created by term limits or not) and the same number, or fewer, females running for office,
then open seats will not have a positive impact on the number of females in office. This idea
is especially important in term limited states since the nature of term limits creates more
open seats.
Conventional wisdom suggests more open seats will lead to more women in office,
however, without increased female candidate emergence, there will not be an overall increase
of women in office. Carroll and Jenkins (2001) show that in states with term limits, the
number of female officeholders decreased when term limits were enacted. However, it is
possible this finding may be due to lower levels of female candidate emergence. The lack of
female candidate emergence is apparent in Carroll and Jenkins’ (2001) finding that females
did not run in a large number of primary elections for either party.
The greater number of open seats created by term limits provide potential female
18
candidates with more opportunities to emerge. Since an incumbent cannot continuously
run, the overall likelihood of winning an election will increase for new candidates. Further,
Fulton et al (2004) find women are more likely to run in races they perceive their chances of
winning to be high. Open seats, created by term limits or otherwise, offer a better chance at
winning an election than running against an incumbent. Therefore, states with open seats
created by term limits should see an overall increase in the number of female candidates
emerging because opportunity costs of losing are lower.
Eligibility Pool
The presence of term limits may influence more females to emerge and run as can-
didates in state legislative elections. This is due to the fact that with term limits, politics
is not necessarily a career path. In a 2008 CAWP Recruitment Study by Carroll, Sanbon-
matsu and Walsh, nearly 43% of women in the lower chamber of the state legislatures had
no previous officeholding experience. Further, and more importantly, nearly 73% of females
in the lower house claimed their decision to run for office was not done as a stepping stone
towards higher office. The study also finds that nearly 79% of woman in the lower chamber
reported having an occupation that allows for sufficient time and flexibility to hold office was
important when considering their run for office. These findings suggest that many females
in state legislatures do not plan on being career politicians. And therefore, term limits may
allow more opportunities for potential candidates to run and hold office and still avoid being
career politicians.
Term limits create more opportunities for potential candidates who do not want to
be career politicians since a majority of women report they do not see the state legislature
as a stepping stone. Term limits may add a little extra incentive to run for office since the
19
length of time someone can spend in office is limited. Thus, term limits create a unique
opportunity for non career politicians to serve short term in office to make policy changes3
and then leave after they are termed out.
Furthermore, term limits may also have a positive impact on female politicians who
want to be career politicians. Term limits provide a structured timeline for how long one can
legally stay in office before they move on to the next level of government. For example, a
female may desire to move out of the lower chamber in the state legislature but no opportu-
nities exist in the upper chamber. States that have term limits in both chambers have more
open seats and, therefore, greater opportunities for new candidates to run for office. Term
limits provide an opportunity for females to gain more experience and exposure; especially
for females who may want to progress into federal office after they have termed out of their
state legislature.
Party recruitment is important for increasing movement from eligibility to female
candidates (Fox and Lawless 2004; Sanbonmatsu, 2006). If party recruitment in term limited
states rises, there should be more female candidates emerging. The 2008 CAWP recruitment
study reports that nearly 55% of women in the lower chamber were actively recruited by
a party leader. Further, women organizations may also impact the likelihood of a female
running for office. Nearly 22% of women in the lower chamber were encouraged to run for
office by women organizations. While, many women organizations are quite successful at the
national level, there is still room for improvement and growth at the state-level (Bernstein
and Chada 2003). The state-level is arguably an important strategic focus for parties and
organizations since females are becoming more progressively ambitious and likely to run for
3Also in the 2008 CAWP recruitment study, 35% of women in the lower chamber report their decision torun was based on concern about one or more specific policy issue
20
national office if they have previous office-holding experience.
The initial studies on term limits reported parties and women organizations did not
actively seek female candidates (Bernstein and Chada 2003). Yet perhaps some of these
organizations and parties did recruit candidates but were not successful in convincing females
to run. Or the case may be parties and organizations did not exactly know the best strategies
for recruiting candidates when term limits first took effect. Given the fact females are more
likely to run when asked, (Fox and Lawless 2004; Sanbonmatsu 2006) it is only a matter of
time before more females will be asked to run for office. The entire candidate pool has to
grow in term limited states, otherwise no one will run. Therefore parties and organizations
must have some strategy at hand to increase candidate recruitment levels (for male and
females alike).
In sum, taking into consideration the constraints and variation in the use of term
limits across states, one should not expect a drastic increase of women in office in states
with term limits. Rather, the number of women in office will likely rise fairly steadily
overtime as more officials are forced to vacate the state legislature. There may be loss in
the short term as research finds (Carroll and Jenkins 2001; Bernstein and Chada 2003), but
this should not be the norm. Open seats will gradually increase as incumbents are termed
out. Furthermore, term limited states will only increase the number of women in office if
the number of women running for office increases. Term limits having a positive impact on
the likelihood a female will run for office is the first stage in the process of increasing the
number of females in office.
Therefore,
H1: Term limits increase the likelihood of female candidate emergence.
21
Data
To examine the hypothesis, I utilize a candidate-level data set. The candidate data
comes from the State Legislative Election Returns: 1967-2010 (ICPSR 2014). The dataset
captures candidate-level state legislative elections and includes election results, term length,
type of legislative election, incumbency status and party identification. The large time span
allows me to better test the impact of term limits on female candidate emergence. I conduct
the analysis across all 50 states at the candidate-level in the general election (discussed more
below) using a quasi natural experiment design. I examine the years from 1990-2010, both
pre and post term limits, across all candidates meeting the standard 5% threshold (Canon
1978) using states without term limits as the control group.
The ICPSR data does not contain the candidate’s gender and, therefore, I have coded
each candidate’s gender. As a robustness check I have also utilized the CAWP (Center for
American Women and Politics) database as a cross-reference. CAWP provides a list of
the total number of females who ran in an election for each state. After coding the ICPSR
dataset to include gender, I looked to the CAWP dataset to examine any discrepancies. When
there were discrepancies between ICPSR and CAWP I then cross-referenced the Secretary
of State’s original election results (when available) to fix or confirm the findings from my
dataset. Further, in a handful of cases where the Secretary of State’s election results were
not readily available and discrepancies existed, I searched newspaper articles for gender
candidate cues.4
While the ICPSR dataset contains a large number of election periods, I will only use
the time period from 1990-2010. Since 1996-2000 is the time frame in which most states first
4In the cases where I could not determine the gender of a candidate I did not include these cases in themodels. Some were missing first names in the original ICPSR study and I could not determine first names.The missing candidates make up <3% of the total observations
22
experience the impact of term limited officials, the time frame allows a roughly equal pre
and post number of election cycles. The analysis for this time period will therefore capture
a considerable amount of elections both before and after term limits took effect in states.
Further, the candidate-level dataset is quite large over a 20 year time frame and produces
roughly 80,000 observations.5
Design
As briefly mentioned, the enactment of term limits provides the unique opportunity
of an intervention point for an experimental design. Further, Mooney (2009) addresses a
major concern when treating term limits as a natural experiments: exogeneity. To model a
causal relationship, there needs to be random assignment of the treatment and control group.
To test for random assignment, Mooney (2009) ran a test of the top state-level explanatory
variables and finds that there is no statistically significant difference between states with
term limits and states without term limits. For example, state-level population, ideology,
and electoral turnout rates do not predict the adoption of term limits. Therefore, term limits
can be treated as a natural experiment because the states with and without term limits have
no statistically significant differences between them. The treatment group for the design
includes all the states that enacted term limits and the control group includes all the states
that have not enacted term limits.
The enactment of term limits happened at different points in time across the states.
Term limits first took effect in a handful of states in the late 1990s and continued to be
enacted at random through the 2000s. As such, there will be multiple intervention points
5Some states allow candidate fusion; candidates can run under multiple party identifications in the sameelection. For these cases, I use the party ID which received the most votes. For example if Jane Smithran as a Republican and received 10,000 votes and as a Democrat with 5,000 votes I drop the Democraticidentification and Jane Smith is coded as a Republican for the election year.
23
that are unique to the state enacting the term limits. Using panel-level data, I am less likely
to violate the assumption that an unobservable/unmeasurable variable has an impact on my
main independent variable, term limits. This is important when determining a causal link
between term limits and the dependent variable. Table 2.1 provides a list of the term limited
states (which have not been repealed). The table provides the year in which the term limits
were enacted and the year in which the first round of legislators were forced to leave due to
term limits .
State Year Enacted Year of Impact
Maine 1993 1996California 1990 1996Colorado 1990 1998Arkansas*ˆ 1992 1998Michigan 1992 1998Florida 1992 2000Ohio 1992 2000South Dakota* 1992 2000Montana 1992 2000Arizona* 1992 2000Missouri 1992 2002Oklahoma 1990 2004Nebraska 2000 2006Louisiana 1995 2007Nevada 1996 2010
Source: NCSL 2014: ‘The Term-Limited States’∗ States with multimember districts.
ˆ Arkansas eliminated multimember districts after their 2000 election
Table 2.1: Term-Limited States
Variables
Dependent
The dependent variable is female candidate emergence, which is based on a dichoto-
mous variable where ‘1’ is a female candidate and ‘0’ is a male candidate. Table 2.2 reports
24
the descriptive statistics on the number of female candidates in the sample. Between 1990-
2010 female candidates made up 26% of the total candidates running for office. Table 2.2
provides further descriptive statistics showing the distribution of the data based on candi-
date gender and term limits. At first glance, females are more prevalent in term limited
states, making up 28% of the candidates versus non-term limited seats where females are
only 25.9% of the candidates.
Variable Number Percent of Sample
Female Candidates 24,537 26.30%Male Candidates 68,380 73.30%Non Term Limit Seats 79,029 85.54%Term Limit Seats 13,888 14.46%
Table 2.2: Descriptive Statistics for Female Emergence and Term Limits
Independent Variable
The main explanatory variable of interest is term limits. States with term limits are
coded ‘1’ and all states without term limits are coded ‘0’. Since, the intervention points
will be scattered due to states implementing term limits in different years, each state with
term limits will be coded to reflect the year in which that state’s term limits take effect.
States with term limits are coded as ‘0’ until they implement term limits. Table 2.3 shows
the distribution for states with and without term limits. Approximately 15% of candidates
run in term limited seats. Whereas non term limited seats make up about 85% of the total
sample (The non-term limited seats include states with enacted term limits but elections
before they were implemented). Again, since this is candidate level-data with a focus on
candidate emergence, these numbers represent the distribution of candidates running in the
term limited and non-term limited races (not the total number of available seats).
25
Variable No Term Limits Term Limits
Female Candidates 20,646 (25.9%) 3,891 (28%)Male Candidates 58,383 (73.9%) 9,997 (71.2%)
Table 2.3: Distribution of Candidates and Term Limits
Control Variables
The model includes several candidate-level and district level controls. The first control
is for open seat where ‘1’ is when a candidate runs in a district with an open seat and ‘0’
is when an incumbent is running within that district. As extant research and my theory
suggests, open seats decrease institutional barriers for women to enter office so I expect open
seats to have a positive impact on likelihood a female will run for office.
Research also finds women are more successful when they compete in multimember
districts (MMD) (Carroll 1994; Darcy, Welch, and Clark 1994; Rule 1990). This is due to the
fact that there are more opportunities to win a seat and it is not a zero-sum game. A handful
of state legislatures use multimember districts in their elections rather than single-member
districts. To measure MMDs, I use the District Type variable which differentiates between
the types of districts. The variable is coded as a dummy for purposes of this study where
‘0’ is a single-member district and ‘1’ is a multimember district.
Incumbents have an advantage in elections, even at the state-level. Incumbents are
more likely to win their elections and is an important variable to include in the model. Here,
incumbent is coded so a ‘0’ is a non incumbent and ‘1’ is an incumbent candidate. As with
all literature, I expect incumbency to have a negative impact on female candidate emergence.
Incumbents inhibit females from winning office and may keep them from running completely
due to the costs associated with running a campaign, especially one where the challeneger
is expected to lose against an incumbent. While congressional literature finds women have
26
an increasing tendency to run as challengers against female incumbents (see Lawless and
Pearson 2008), much of the literature finds incumbents dissuade candidates from emerging.
Another important control variable is party. Research finds female candidates tend
to emerge and hold seats in office more often as Democrats than Republicans. Republicans
are coded as ‘0’ and I expect that Democrats, coded as ‘1’, will have a positive impact on
the number of female candidates.
I also include a control for term length. Given my theory that term limits may attract
females who do not wish to be career politicians, I expect there to be a negative relationship
between the length of a term in office and female emergence. Females should be more likely
to emerge in states with fewer years in a term; the longer a term, the less likely a female will
emerge.
A measure of competitiveness is also important; especially since females are less likely
to run for office if they perceive their chances of winning to be low. As a crude measure of
competitiveness I will include the Number of Candidates in each election. While this is not
a direct measure of whether a campaign is competitive or not because it does not include
the types of candidates running6, Lawless and Pearson (2008) find females tend to be in
elections where they are facing more than one person of opposition. While a majority of
races are unopposed or have only two candidates, capturing opposition, especially if there is
zero opposition, is an important to control.
Lastly, I include a measure of legislative professionalism. I use the Squire index
which is a value between 0 and 1 assigned to states based on their level of professionalization
within the legislature. Values closer to 0 are the least professional and values closest to 1
6i.e whether or not the candidate is a a qualified candidate and is more likely to win than a non-qualitycandidate. Non quality, meaning a candidate has little to no relevant experience in politics. Usually specifiedas previous officeholding experience at some level of government
27
are the most professional (see Squire 1992). I expect professionalization to have a negative
impact on the likelihood a female candidate will emerge. This measure is capturing my
theoretical argument that women are more likely to emerge when they do not want to be
career politicians; therefore, when professionalism increases, female emergence will decrease.
Method
I employ two methods to test my hypothesis that term limits increase female candidate
emergence. The first is a difference in differences method. Difference in differences allows
time invariant covariates to be factored out, while incorporating overall exogenous shocks or
trends in the system. Many state-level variables can be controlled for, such as population,
ideology and turnout. There are, however, other unobservable factors such as candidate
ambition levels or time spent on family responsibilities that cannot be measured. While,
these factors should remain fairly constant overtime they are not easily measured without
extensive survey data.
The difference in differences method is shown below. Equation 1 represents the model
for states with term limits. Term limits serve as an intervention point and the difference
before and after term limits were implemented is provided. The model keeps all variables
constant while the unobservable/time-invariant, ai, factors are already factored out. Equa-
tion 2 represents the control group: states without term limits. Equation 3 represents the
difference between the treatment and control groups: states with term limits and states lack-
ing term limits. The error terms are also differenced so this better captures overall shocks
to the system and can capture any general trends. If there is a difference found in equation
3, then term limits have a causal impact on women emergence.
28
∆yiTL = δ0 + β + ∆uiTL (1)
∗where ∆yiTL denotes the difference in y for states with term limits
∆yiNTL = δ0 + β + ∆uiNTL (2)
∗ where ∆yiNTL denotes the difference in y for states without term limits
∆yiTL − ∆yiNTL = β + (∆uiTL − ∆uiNTL) (3)
*where y is the mean number of women emerging
I also run a logistical regression model to examine the impact term limits have on
the likelihood a female candidate will run for office. The unit of analysis for the logit model
is the candidate-year. Therefore, candidates will appear in the model for every election
cycle they run in as a new observation. The logit model allows me to control for election
specific variables previously discussed (control variables) while also controlling for time.
Where difference in differences examines the difference in emergence between male and female
candidates in states, the logistical regression aims to capture and predict the likelihood a
candidate is a female based on the electoral environment at the individual level; i.e. term
limits, partisanships, district type, etc.
Results
A simple difference in difference test shows there is some difference in emergence rates
between male and female candidates in states with term limits. While the difference is minor,
the results suggest term limits increase the mean number of female candidates running for
office. Table 2.4 provides evidence of this relationship within and across states. States with
term limits see a small increase in the number of females emerging while states without term
limits do not see a major increase.
29
Female Emergence Mean Standard Error
Pre Term Limits (e) 0.243 0.004Post Term Limits (e) 0.267 0.004∆yiTL = 0.024Pre Term Limits (c) .240 0.003Post Term Limits (c) 0.258 0.002∆yiNTL = 0.018∆yiTL − ∆yiNTL = 0.006
Table 2.4: Difference in Differences
The control group, states without term limits, saw an increase in the overall percent
of females running for office post term limits (1996) by 1.8%. While the experimental group,
states with term limits, saw an increase in the overall percentage of females running increase
by 2.4%. Substantively, this shows term limits increase the overall percentage of females
running for office by about .6%. While the percent seems rather small, this .6% increases
equates to roughly 10 more females running for office. For example, between 1994 and 2000,
Ohio (a state with term limits), there was a .06 difference in the mean number of females
running for office. Looking at the total numbers for the state show 41 females ran for office
in 1996 and this number increased to 52 in 2000.
Yet, the likelihood of a female emerging on an individual level is important to un-
derstand as well; the mean provided in the difference in differences is important but the
logit model shows more of the individual-level factors that impact the likelihood a candidate
will be female. Table 2.5, below, shows the results of the model, which also supports the
hypothesis that term limits increase the likelihood a candidate female will emerge. While lo-
gistical regressions are not directly interpretable without predicted probability models, term
limits is in the expected direction and is statistically significant. Furthermore, the open seat
variable, consistent with the literature, finds that women are more likely to emerge in open
30
seat races.
Variable Model 1 Model 2
Term Limit 0.101 (0.033) 0.103 (0.033)Open Seat 0.054 (0.023) 0.053 (0.023)District Type 0.296 (0.034) 0.279 (0.036)Incumbent -0.068 (0.026) -0.069 (0.027)Party ID 0.458 (0.029) 0.459 (0.029)Term Length -0.151 (0.028) -0.154 (0.028)Number Candidates 0.043 (0.020) 0.055 (0.020)Professionalism (Squire) - -0.168 (0.109)Constant -1.064 (0.064) -1.025 (0.070)n =88288
Table 2.5: Likelihood Candidate is a Female
Furthermore, the control variables are all in the expected direction with statistical
significance at the typical .05 level. The model is consistent with literature that suggests
women are more likely to run in multimember districts; here Table 2.5 shows a positive
significant value since the baseline category is a single-member district. The results also
show the negative impact the incumbency advantage has on female candidates. Since the
incumbency advantage is so great I expected a negative relationship between incumbents
and female candidates. Research already finds there are fewer females in office as well as
the candidate pool- therefore, incumbent should have a negative relationship with female
candidate likelihood. Furthermore, the baseline categories for the model are Republicans
and consistent with past literature, a Democratic party identification has a positive and
significant impact on the likelihood a candidate will be female.
For Model 2, professionalism in is in the expected direction, but is not statistically
significant at conventional levels. Of note though is that the model stays consistent and all
other variables are still significant and in the expected direction. Model 2 is a good robustness
check, controlling for differences across legislative chambers, but since professionalism was
31
insignificant, I will use Model 1 for predicted probabilities.
While the direction and impact are important, predicted probabilities for the model
provide substantively interpretable results. Tables 6 and 7 provide predicted probabilities
for when a candidate is Democrat and Tables 8 and 9 provide the predicted probabilities
for Republican candidates. All the predicted probabilities values are set based on the most
likely of cases, rather than an overall mean which oftentimes captures unlikely cases (such as
a 2.5 year term). Therefore, competition is set at 1 (to represent 1 opponent), term length
is set to the average, which is 2 years, open seat is to the average which is 0 and then I vary
incumbency and district type to further analyze the relationship. To further analyze the
impact open seats tables, 7 for Democrats and 9 for Republicans, present the impact open
seats have on female emergence in term limited states.
Table 2.6 shows significant differences between term limited and non term limited
states for Democrats. Candidates are more likely to be female in multimember districts and
all around more likely to emerge in term limited states. The predicted probabilities also
show the impact term limits have on non-incumbents. As theorized, non incumbents in term
limits states have a higher probability of being female than non incumbents in non term
limited states. The impact on non incumbents is important to note since term limits lower
the incumbency advantage. Female non incumbents in term limited states emerge 32% of
the time versus 30% of the time in non term limited states. This advantage is even more
apparent in MMDs where females in term limited states emerge about 39% of the time versus
36% in non term limited states.
32
Female Emergence Incumbent Non Incumbent
Single Member DistrictsNo Term Limits 0.284 (0.007) 0.308 (0.006)Term Limits 0.305 (0.008) 0.320 (0.008)Multimember DistrictsNo Term Limits 0.348 (0.009) 0.364 (0.008)Term Limits 0.371 (0.012) 0.388 (0.011)
All margins are significant at p < .05 Std error provided in parentheses
Table 2.6: Predicted Probabilities for Democrats
Female Emergence Open Seat
Single Member DistrictsNo Term Limits 0.317 (0.009)Term Limits 0.339 (0.011)Multimember DistrictsNo Term Limits 0.384 (0.012)Term Limits 0.408 (0.014)
All margins are significant at p < .05 Std error provided in parentheses
Table 2.7: Predicted Probabilities for Democrats in Open Seats
Table 2.77 finds open seat races in term limited states provide an added boost for
female emergence. In single member districts, under term limited seats, females emerge
nearly 34% of the time versus about 32% of the time in states without term limits. These
values, are statistically different and the conventional 95% confidence bands do not overlap.
Further, women in multimember districts also are more likely to emerge in states with term
limits. In MMDs, female emergence is nearly 41% versus 38% of the time in non term
limited states. These predicted probabilities are consistent with past literature on MMDs
and consistent with my theory that term limits in fact do provide an bit of an extra incentive
for women to seek office.
7Predicted probabilities- setting the number of candidates in race to 2 and incumbent to zero. Further, Iexamine this variation by SMDs v MMDs
33
Table 2.8 reports the predicted probabilities for Republicans. Here the predicted
probabilities are comparatively lower than the Democratic ones, which is expected given
Democratic female candidate trends. The literature finds females are much more likely to
emerge and ultimately win office under the Democratic party. In any case, the Republican
model still shows a difference between term limited and non term limited states. The prob-
ability is already low for female emergence, but term limits do seem to provide incentive
for female emergence; especially for non incumbents. Female non incumbents in SMD term
limited states make up 23% of the emerging candidates versus 21% in non term limited
states. Further and consistent with the Democratic predicted probabilities, MMDs offer an
advantage to non incumbents and females emerge about 29% of the time in term limited
states versus 25% of the time in non term limited states.
Female Emergence Incumbent Non Incumbent
Single Member DistrictsNo Term Limits 0.201 (0.006) 0.212 (0.005)Term Limits 0.218 (0.007) 0.230 (0.007)Multimember DistrictsNo Term Limits 0.252 (0.008) 0.255 (0.007)Term Limits 0.272 (0.010) 0.286 (0.010)
All margins are significant at p < .05 Std error provided in parentheses
Table 2.8: Predicted Probabilities for Republicans
Female Emergence Open Seat
Single Member DistrictsNo Term Limits 0.227 (0.008)Term Limits 0.245 (0.009)Multimember DistrictsNo Term Limits 0.283 (0.010)Term Limits 0.304 (0.012)
All margins are significant at p < .05 Std error provided in parentheses
Table 2.9: Predicted Probabilities for Republicans in Open Seats
34
Further, Table 2.9 is also consistent with the previous findings and show Republican
females are more likely to emerge in states with term limits. In non term limited, open seat
races, about 23% of candidates are female while females are about 25% of candidates in
term limited, open seat races. This 2% difference is also present in MMDs where Republican
females are more likely to run in term limited states, 30%, versus 28% in non term limited
states.
While the predicted probabilities indicate term limits raise the likelihood a candidate
will be a female, I also checked the confidence intervals on each of the reported predicted
probabilities to make sure the differences are significant. While all differences were signifi-
cant with 90% confidence bands, the most significant difference was found in single-member
districts with non-incumbents (with 95% confidence bands), term limits had the most sig-
nificant impact for this combination of factors. This supports my theory and sheds a bit
more light on the conventional wisdom of the incumbency advantage and term limits. The
results suggest term limits have an impact at some level. I expect the impact to grow slowly
overtime as more open seats become available and more candidates run in order to fill the
growing number of seats.
Conclusions and Implications
Currently, females make up only 24 percent of state legislators in the United States.
This is concerning given a 2009 CAWP study that indicates the growth of female candidates
at the state legislative level has remained fairly stagnant since 1997. Since about half of
women in higher office begin their political career at lower levels of government, this stag-
nation in female state legislators is an important area of research to study. If the number
of women in state legislatures does not increase and/or gain parity, to generally reflect the
35
population descriptively, the implications for descriptive and substantive representation is
concerning at the state and national level.
This chapter offers a theoretical explanation on how the institutional feature of term
limits may have an impact on the number of female candidates deciding to run for office.
Initial studies on term limits suggest term limits have negative and undesirable effects on the
total numbers of females in office. Using a candidate-level dataset and experimental design
on all 50 states, I find that term limits have some minor, positive impacts on the number
of females running for office. Further, the findings suggest term limits are more likely to
increase female emergence than open seats. Women are more likely to emerge in open seats
in states with term limits. This suggests open seats in term limited states are different,
perhaps in incentive structure as my theory suggests, than open seats in non term limited
seats. More research teasing out the relationship between open seats in term limited states
is needed.
Overall, these findings are important because the initial findings on term limits may
not be a problematic for females as expected. A greater number of female candidates can
lead to a greater number or females in office. The next chapter reexamines the initial puzzle
of whether or not term limits increase female descriptive representation.
There is still great room for additional research, especially with the dataset. The
more data on a candidate available, the better fit the model will be. Better data on women
candidates, such as candidate quality, campaign financing, etc. may add insight as to how
term limits impact the eligibility pool in which women emerge. In the future, I intend to
conduct further research on the types of women running for office and pipeline politics.
Initial studies on term limits find the type and quality of officeholders has not changed much
36
ı.e., while there has been turnover, the age and background of politicians has not changed
(Moncreif, Powell and Storey 2007). Yet, I argue in this chapter the incentive structure for
office changes. Therefore, the type of candidate emerging may be different than who wins
and the implications for how these ‘outsider’ candidates may impact an election is interesting
to explore. Teasing out this relationship with some case studies is a worthwhile step forward
to answering these questions.
Furthermore, implications for national representation of women in office are worth
considering. The impact of term limits and the type of candidate attracted to the legislatures
are likely to have an impact beyond the state level. If term limited states are attracting
candidates who do not want to be career politicians, this may negatively impact female
recruitment success for congressional races. Take for example, the possibility of a candidate
originally seeking office in the state with no intentions of being a career politician and leaving
after their term has expired—the candidate may go back to their previous career, or perhaps
they will seek higher office. If this candidate is female and decides no to run for higher office,
overtime, descriptive representation at the national level may be negatively impacted.
37
[1] Allebaugh, Dalene, and Neil Pinney. 2003. “The Real Costs of Term Limits: Comparative Study ofCompetition and Electoral Costs.” in The test of time: Coping with legislative term limits.
[2] Bledsoe, Timothy and Mary Herring. 1990. “Victims of Circumstance: Women in Pursuit of PoliticalOffice.” The American Political Science Review. 84:213-223.
[3] Burrell, Barbara C. 1994. A Woman’s Place Is in the House: Campaigning for Congress in the FeministEra. Ann Arbor: University of Michigan Press.
[4] Canon, Bradley. 1978. “Factionalism in the South: A test of theory and a revisitation of V. O. Key.”American Journal of Political Science. 22:833-48.
[5] Carey, John M., Richard G. Niemi, and Lynda W. Powell.2000. Term limits in state legislatures.University of Michigan Press.
[6] Carroll, Susan and Kira Sanbonmatsu. 2013. More Women Can Run: Gender and Pathways to theState Legislatures. Oxford University Press.
[7] Carroll, Susan and Krista Jenkins. 2001. “Do Term Limits Help Women Get Elected?” Social ScienceQuarterly. 82/1.
[8] Census Bureau. 2012.
[9] Center for American Women and Politics. 2012. “Women in State Legislatures? Fact Sheet”:www.cawp.rutgers.edu
[10] Crane, Edward H. 1995. “Testimony before the Subcommittee on the Constitution, Committee on theJudiciary, U.S. Senate.”
[11] Darcy, R, Susan Welch and Janet Clark. 1994. Women, Elections, and Representation. 2nd Edition.Lincoln: University of Nebraska Press.
[12] Dolan, Kathleen. 2004. Voting for Women: How the Public Evaluates Women Candidates. Boulder,CO: Westview.
[13] Dolan, Kathleen. 2009. “The Impact of Gender Stereotyped Evaluations on Supoport for WomenCandidates.” Political Behavior. 32:69-88.
[14] Elder, Laurel. 2012. “The Partisan Gap among Women State Legislators.” Journal of Women, Politicsand Policy. 33:65-85.
[15] Ferry, Jonathan. 1994. “Women, Minorities and Term Limits: American’s Path to a RepresentativeCongress”. U.S. Term Limits Foundation Outlook Series.
[16] Fowler, Linda. 1992. A Comment on Competition and Careers. In Gerald Benjamin and Michael J.Lambin. eds., Lmiting Legislative Terms Washington, DC: CQ Press.
[17] Fox, Richard L. 2000. “Gender and Congressional Elections.” In Gender and American Politics: Women,Men, and the Political Process, eds. Sue Tolleson Rinehart and Jyl J. Josephson. Armonk, NY: M.E.Sharpe.
[18] Fox, Richard and Jennifer Lawless. 2004. “Entering the Arena? Gender and the Decision to Run forOffice.” American Journal of Political Science. 48:264-280.
[19] Fulton, Sarah et, al. 2006. “The Sense of a Women: Gender Ambition, and The Decision to Run forCongress”. Political Research Quarterly. 59: 235-248.
38
Chapter Bibliography
[20] Fulton, Sarah. 2012. “Running Backwards and in High Heels: The Gendered Quality Gap andIncumbent Electoral Success.” Political Research Quarterly. 20:1-12.
[21] Gershon, Sarah. 2012. “When Race, Gender, and the Media Intersect: Campaign News Coverage ofMinority Congresswomen” Journal of Women and Politics. 33:105-125.
[22] Gertzog, Irwin. 2002. Women’s Changing Pathways to the US House of Representatives. University ofOklahoma Press.
[23] Huddy, Leonie, and Nayda Terkildsen. 1993. “Gender Stereotypes and the Perception of Male andFemale Candidates.” American Journal of Political Science. 37: 119-47.
[24] Kazee, Thomas.1994. “Who Runs for Congress ?” Washington, DC: Congressional Quarterly.
[25] Kahn, Kim Fridkin. 1992. ”Does Being Male Help? An Investigation of the Effects of Candidate Genderand Campaign Coverage on Evaluations of U.S Senate Candidates” Journal of Politics 54: 497-517.
[26] King, David and Richard E. Matland. 2003. “Sex and the Grand Old Party: An ExperimentalInvestigation of the Effect of Candidate Sex on Support for a Republican Candidate.” AmericanPolitics Research. 31: 595-612.
[27] Klarner, Carl, William Berry, Thomas Carsey, Malcolm Jewell, Richard Niemi, Lynda Powell,and James Snyder. State Legislative Election Returns (1967-2010). ICPSR34297-v1. Ann Ar-bor, MI: Inter-university Consortium for Political and Social Research [distributor], 2013-01-11.doi:10.3886/ICPSR34297.v1
[28] Koch, Jeffrey. 2000. “Do Citizens Apply Gender Stereotypes to Infer Candidates’ Ideological Orienta-tions?” Journal of Politics 62: 414-29.
[29] Koch, Jeffrey. 2002. “Gender Stereotypes and Citizens’ Impressions of House Candidates’ IdeologicalOrientations.” American Journal of Political Science. 46(2), 453-462.
[30] Milyo, Jeffrey and Samantha Schosberg. 2000. “Gender Bias and Selection Bias in House Elections.”Public Choice. 105:41-59.
[31] Moncrief, Gary, Lynda W. Powell, and Tim Storey. 2007. “Composition of Legislatures”. in InstitutionalChange in American Politics: The Case of Term Limits. eds. Karl Kurtz, Bruce Cain, and RichardNiemi. The University of Michigan Press.
[32] Mooney, Christopher. 2009. “Term Limits as a Boon to Legislative Scholarship: A Review.” StatePolitics and Policy Quarterly. 9:204- 228.
[33] National Conference of State Legislatures. 2011 “Legislative Term Limits Overview?http://www.ncsl.org/ .
[34] Osborn L, Tracy. 2012. How Women Represent Women: Political Parties, Gender and Representationin the State Legislatures. Oxford University Press.
[35] Palmer, Barbara and Dennis Simon. 2005. “When Women Run Against Women: The Hidden Influenceof Female Incumbents in Elections to the U.S. House of Representatives, 1956-2002.” Gender andPolitics. 1:39-63.
[36] Palmer, Barbara and Dennis Simon. 2006. Breaking the Political Glass Ceiling: Women and Congres-sional Elections. New York: Routlege.
[37] Sanbonmatsu, Kira. 2006. ”Where women run.” Ann Arbor: University of Michigan Press.
[38] Sanbonmatsu, Kira, Susan J. Carroll, and Debbie Walsh. Center for American Women and Politics(CAWP) Recruitment Studies, 2008. ICPSR35244-v1. Ann Arbor, MI: Inter-university Consortium for
39
Political and Social Research [distributor], 2015-05-13. http://doi.org/10.3886/ICPSR35244.v1.
[39] Sanbonmatsu, Kira and Kathleen Dolan. 2009. “Do Gender Stereotypes Transcend Party?” PoliticalResearch Quarterly 62: 485-494.
[40] Schwindt-Bayer, L. A. and Mishler, W. 2005. “An Integrated Model of Women?s Representation.” TheJournal of Politics. 67/2: 407-428.
[41] Seltzer, Richard A., Jody Newman, and Melissa Voorhees Leighton. 1997. Sex as a Political Variable:Women as Candidates and Voters in U.S. Elections. Boulder: Lynne Rienner.
[42] Swers, Michele. 2002. The Difference Women Make. Chicago: University of Chicago Press.
[43] Werner and Mayer. 2007. Public Election Funding, Competition, and candidate Gender.
40
CHAPTER 3
FEMALE CANDIDATE SUCCESS
Abstract
In this chapter, I offer further insight into the lack of females in state legislative office.
While scholars point to the incumbency advantage as the main reason for a lack of descriptive
representation (see: Carroll and Jenkins 2001; Darcy, Welch, and Clark 1994; Jacobson 2000;
Burrell 1992), the implementation of term limits provides a unique opportunity to study the
success of women in office. I argue, using the state politics and the women and politics
literature, term limits are more beneficial for Republican females than Democratic females.
To test my theory, I look at the interactive effect between professionalization, term limits
and female candidates. To do so, I examine state legislative electoral outcomes in general
elections to the lower house (1990-2010) using a logistical regression. The results indicate
that as professionalism increases, female Democrats are less likely to win in term limited
states than non term limited states. Female Republicans on the other hand are more likely
to win their seat as professionalism decreases in term limited states than non term limited
states.
Introduction
The early 1990s experienced a wave of state-level ballot initiatives targeting term
limits for elected state officials. Citizens were frustrated with their politicians and believed
the implementation of term limits would help alleviate the lack of responsiveness at the
state-level by forcing long term politicians to leave office. Term limits successfully passed in
41
21 states1 via ballot initiative (Farmer, Rausch and Green 2003). Women in politics scholars
hypothesized term limits would be the start of a new era, as they would allow more women to
enter office (Ferry 1994; Crane 1995). Yet, subsequent research finds term limits decreases
the number of women in state legislative office (Carroll and Jenkins 2001; Bernstein and
Chada 2003). This chapter explores the discrepancy in expectations by reexamining theo-
retical and empirical questions surrounding the impact term limits have on female descriptive
representation.
Term limits provide a unique institutional tool employed in a handful of states, which
can provide the path necessary to bring parity amongst males and females in office. Females
are far from being equally represented in state legislatures and only make up about 24% of
the total state legislators (NCSL 2015). Descriptive representation has long been a normative
concern in political science based on principles of representative democracy.
While extant literature addresses many concerns with the lack of female represen-
tation, the research since the ‘Year of the Women’ focuses largely on the national level of
government (Carroll and Sanbonmatsu 2013). Yet recently, focus has shifted back to female
descriptive representation at the state-level (Carroll and Sanbonmatsu 2013). Research at
the state-level is important because state legislatures serve as a step along the political am-
bition ladder for most politicians.2 Furthermore, states play a large role on the livelihood
of citizens and deal with many policy issues. Therefore, having descriptive representation in
state legislatures is important for having women’s issues placed on the agenda (Swers 2002;
Osborn 2012).
1Between 1997-2004 6 states repealed term limits via the State Supreme Court or Legislature. Now, only 15states have term limits.
249% of the 113th Congress were former state legislators (NCSL 2013).
42
In this chapter, I seek to offer further insight regarding the lack of female descriptive
representation. Given the puzzling findings between term limit expectations and female de-
scriptive representation, I present a study analyzing term limits as an intervention point in a
quasi-natural experiment. This design allows a better comparison between female descriptive
representation in states with term limits and states without term limits.
This chapter proceeds as follows: I first discuss the extant literature on the historically
low levels of female descriptive representation and the term limits literature with regard to
female descriptive representation. Next, I offer a theoretical argument regarding term limits,
professionalization, party identification and females. Then, I present my hypotheses and
discuss the research design. Finally, I present the findings and reflect with conclusions and
implications.
Females in Office
Assuming there is a sizable pool of eligible female candidates (as discussed in the
previous chapter), it is reasonable to ask what is the causal mechanism that has resulted
in continued low levels of female descriptive representation. Candidate emergence is an
important stage to analyze, but election outcomes, when females win or lose, are important
too. An increase in descriptive representation rests on the ability of females to win office,
not just run for office. This section will focus on the current research explaining constraints
on female office-seekers
Since the 1990s, the overall percentage of females in office has remained fairly stag-
nant (CAWP 2012). Yet, there is large variation in descriptive representation across the
50 state legislatures. There are some states, like Colorado, which has a legislative branch
with fairly equal percentages of men and women. Yet, there are states, like Louisiana, which
43
are composed of a super-majority of men.3 Below, I offer a discussion on the current state
of women and state politics literature. I discuss the factors that influence the variation of
female descriptive representation across the U.S. states. I begin a discussion of voter bias
and media bias against female candidates, then present the institutional constraint litera-
ture, and lastly discuss findings regarding the impact term limits have on female descriptive
representation.
Voter Bias
Voters may have a bias against female candidates placing women at a disadvantage.
The literature points to two main public perceptions that may influence vote choice: Voter
attitudes towards the acceptance of female candidates, and whether or not voters use stereo-
types about female candidates in their decision calculus.
First, in regards to voter attitudes towards female candidates, is whether or not
females should hold office. Historically, politics was a man’s game and women were not
allowed to vote. Further, the few women who were appointed or elected served as tokens
rather than serious politicians (see Foerstel and Foerstel, 1996). Voter attitudes on female
politicians have changed. At the Congressional level, scholars find women to be just as likely
to win elections as men indicating that when women run, they win (Burrell 1994; Darcy,
Welch, and Clark 1994; Fox 2000; Seltzer et al. 1997). The public has become much more
accepting of the idea that a female can hold political office. Yet, some bias still exists since
one in five voters believe men are emotionally better-suited for politics (Lynch and Dolan
2013; Lynch and Dolan 2014).
Secondly, and arguably the most prevalent biases against female candidates, are the
3Colorado has 42 women in office accounting for 42% of the total legislature while Louisiana currently has17 women in office which accounts for about 12% of the total legislature.
44
stereotypes voters use. Extant research finds the public holds stereotypes about both male
and female candidates in regards to personality traits and issue competency (see Lynch and
Dolan 2014).
Personality traits deal with how the public views candidates in terms of personal
characteristics. Generally speaking, voters view females as more compassionate and honest
while men are viewed as more experienced and better leaders (Alexander and Andersen 1993;
Burrell 2008; Huddy and Terkildsen 1993; Kahn 1996; King and Matland 2003; Lawless 2004;
Leeper 1991; Paul and Smith 2008; Sapiro 1981). Yet, scholars find voters view experience,
rather than honesty and compassion (i.e the stereotypes associated with men) as better
qualities for an elected official (Huddy and Terkildsen 1993; Lawless 2004; Rosenwasser and
Dean 1989). Therefore, despite the positive personality traits voters associate with females,
these traits are not as valued and may work against female candidates.
Issue competency bias is also present in voters’ decision calculus. Extant literature
finds the public perceives women and men to be better at specific, and different issues
(Alexander and Anderson 1993; Huddy and Terkildsen 1993; Kahn 1992; Koch 2000; Law-
less 2004). Female politicians are better at dealing with issues involving social welfare such
as healthcare, education and poverty. Men are viewed as competent in handling economic
and foreign policy issues (Koch 2000; Sanbonmatsu 2002). Depending on the political envi-
ronment surrounding any given election, these issue competency biases may help, hinder or
have no impact on a female’s chance at winning office.
Media Bias
The media can reinforce and perpetuate stereotypes about female candidates. Fram-
ing, or how the media tries to change the fundamental way a person thinks about an issue,
45
is a tool used by the media when covering female candidates (Fulton 2011). The media
portrays males and females differently and the coverage tends to place males in a favorable
light (Kahn 1992). The media examines women candidates through a feminine lens and
emphasizes trivial facts rather than what the candidate plans to achieve if elected; i.e. the
media often comments on what Hilary Clinton is wearing and if she is having a bad hair
day rather than Clinton’s policy stance (Fulton 2011). When the media portrays women
as feminine, this may cue latent gender stereotypes people hold towards female candidates
(Palmer and Simon 2005).
A prime example of media coverage reinforcing gender stereotypes is the former Mas-
sachusetts governor Jane Swift’s time in office from 2001-2003 (CAWP 2015). While in
office, Swift was in the hospital while pregnant with twins. The media coverage of her hos-
pital visit was harsh and she was often referred to as a “supermom on a power trip”. Swift
was regularly criticized for having young children while holding office and was frequently
asked during press conferences who was watching her children. Such news coverage rein-
forces the idea to voters that women are not suited for the political sphere because their
primary responsibilities lie in the home (Palmer and Simon 2006; McGlen et al 2011).
Institutional Constraints
Moving away from voter and media bias, there are also institutional barriers that
hinder female descriptive representation. Below is a discussion on how the institutional
environment has a negative effect on female descriptive representation.
Incumbency Advantage
Incumbents have high success rates when they run for reelection (Erikson 1971). In-
cumbency advantage posits incumbents high reelection rates reduce opposition and leave
46
little opportunity for non incumbents to compete. Congressional-level research on the in-
cumbency advantage finds incumbents enjoy this advantage due to helping constituents with
casework (Fenno 1977), redistricting and a weakened sense of party identification (Erikson
1972) as well as an ability to deter candidates from running (Cox and Katz 1996; Jacob-
son and Kernell 1983; Gordon, Huber, and Landa 2007). Therefore, an increase in women
in office is inhibited by male incumbents who continue to run and win office at high rates
(Burrell 1992; Darcy, Welch, and Clark 1994; Cox and Katz 1996; Jacobson 2000; Carroll
and Jenkins 2001).
The incumbent reelection rate across all state lower house legislatures is around 50%
and has been fairly stagnant since the 1970s (Ansolabehere and Snyder 2002). While the
state-level incumbency advantage is not as pronounced as the federal government, the ad-
vantage persists. While incumbency almost guarantees an easy win for politicians (males
and females alike), the fact remains that overall, the incumbency advantage serves as an
institutional constraint against females (Burrell 1992, Fox and Lawless 2004). The incum-
bency advantage is particularly harmful against increasing female descriptive representation
since most levels of government are currently composed of men.
Electoral Competition
Another institutional factor perpetuating low levels of female descriptive representa-
tion is the finding that women candidates face more competition than males (Lawless 2004;
Palmer and Simon 2006; Lawless and Pearson 2008). While this research is mainly limited
to the Congressional-level, studies find women face more competition in all types of races;
i.e. as an incumbent, challenger or in open seat races (Lawless and Pearson 2008). In other
words, once in the race, female candidates have to work harder than male candidates because
47
females face a more competitive electoral environment. Women are also less likely to run
unopposed and the number of female versus female races is increasing (Lawless and Pearson
2008). When a female faces a female in a general election only one can win. If the number
of females running is constant, female descriptive representation has a greater chance of
increasing when female candidates run against males, rather than other females.
The previously discussed literatures focuses on how voter, media bias, and the insti-
tutional environment negatively affects female descriptive representation. Next, I discuss the
literature on term limits which tends to highlight the positive effect an institutional change
can have on female candidates. Below I discuss the term limits literature as it relates to
female descriptive representation.
Term Limits
Term limits are an institutional attribute that place restrictions on the number of
times an elected official can run for reelection.4 The expectation amongst scholars was that
term limits would have positive effects on female descriptive representation because they
create a greater number of open seats (Burrell 1992; Ferry 1994; Crane 1995). Essentially,
term limits serve to reduce the incumbency advantage. Congressional research shows open
seats are the best way for women to enter the arena. As such, term limits create greater
opportunities for new candidates to emerge (Carroll and Jenkins 2001; Ferry 1994; Crane
1995). Yet, initial studies on term limits find a negative impact on female descriptive repre-
sentation. While variation existed across the term limited states, there was a general decline
in the number of women in office within the newly term limited states (Carroll and Jenkins
2001; Bernstein and Chada 2003).
4Since the 1990s, 21 states have passed legislation for term limits but 6 states have repealed them.
48
Since term limits are a more contemporary and significant institutional change, schol-
ars have researched their impact on legislatures, legislator behavior as well as candidates5.
Term limits do not seem to translate into more competitive elections (Allebaugh and Pinney
2003). Further, extant research finds term limits have little to no impact on the demo-
graphic composition of legislatures— term limits have not created a new pool of candidates
and therefore, legislators have the same demographic, ideological and partisan character-
istics as they did pre-term limit implementation (Gilmour and Rothstein 1994; Reed and
Schansberg 1996; Carey, Niemi and Powell 1998; Carey et al 2006).
Yet, other empirical studies find the partisan makeup of the term limited state leg-
islatures changed (Meinke and Hasecke 2003). Term limited states tend to become more
Republican post term limits implementation. Theoretically, this is because Republicans are
more likely to take on costs with higher risk. In elections, initial costs to run are high and
there is no guarantee of success or a long term payout (Fowler 1992: Kazee 1994). Further,
opportunity costs to enter politics are different for each party and are dependent on the pro-
fessionalization of the legislature. Republicans tend to have majorities in citizen legislatures
and Democrats do well in professional legislatures (Fiorina 1994; 1996). Meinke and Hasecke
(2003) argue term limits create a “partial reversal of professionalization” (901) which alters
the incentives for legislators and creates an environment favoring Republican candidates.
I propose these studies may be missing a larger pattern. The initial implementation of
term limits may not have an immediate impact on the type of candidates running for office-
whether it be candidates’ quality or gender. A study comparing term limited states to non
term limited states using a series of elections overtime may add insight into a broader effect
5see: Farmer, Rausch, and Green 2003) for a comprehensive study on how term limits effect legislativeperformance in the chamber, the effect of lobbyist as well as leadership.
49
term limits have on candidates and female descriptive representation. Below, I offer a theory
linking the findings in the women and politics literature to the state politics literature. I
look at the incentive structure created by term limits and how this affects female descriptive
representation in state legislative.
Theory on Term Limits and Female Success
The previous chapter discusses emergence an important first step when considering
female descriptive representation. I find term limits increase the likelihood a female will
emerge as a candidate. My findings suggest that despite initial studies on term limits showing
females lost seats in office, there is evidence that more female candidates are emerging in
states with term limits. Therefore, if females continue to run (and win, discussed in the
chapter), descriptive representation should increase in term limited states. Since the number
of females emerging has increased in term limited states, the next step is to examine whether
the increased number of female candidates emerges increases the likelihood a female wins.
The Gradual Impact of Term Limits
The decrease in female descriptive representation in term limited states is rather puz-
zling since open seats create more opportunities for women to gain office. With term limits,
the incumbency advantage is no longer a major institutional constraint against potential
female candidates. Incumbents are forced out and this creates open seats. Yet, post imple-
mentation, females lost seats previously held by women, thus, decreasing the overall number
of females in state legislatures (Bernstein and Chada 2003). Term limits forced females who
had attained office to leave and no new females replaced them (Carroll and Jenkins 2001).
These findings suggest the conventional wisdom surrounding female descriptive representa-
tion and open seats may be incorrect. Or that not enough time has passed and eventually
50
females will become aware of open seats and run for office (whether it be personal choice to
seek office or recruitment efforts by parties, organizations, etc.).
Therefore, a short-term decline may not actually contradict the conventional wisdom.
Since institutional changes tend to have a gradual effect, a single cross-sectional analysis does
not provide enough variation in time to determine whether term limits have a negative or
positive impact on female descriptive representation. Term limits, and the open seats they
create, can still increase female descriptive representation if examined over more than one
election cycle.
Party Identification
Beyond the glacial pace term limits may have on female descriptive representation,
the party of the female candidate is also likely to have an effect on descriptive representation.
Party plays a large roll in understanding female descriptive representation because there is
variation in female candidate success rates across the parties. Examining female Democrats
and Republicans as similar, in terms of paths and incentives to office, assumes there is
no difference in female success rates across parties. Studying the two parties separately is
important for further understanding of different paths to office for female candidates (Carroll
and Zerilli 1993; Dolan and Ford 1998; Sanbonmatsu, 2002). The differences across party
are important since the total number of females in all state legislatures for 2015 varies by
party. The majority of female legislators are Democrats (60%), and 30% of female legislators
are Republicans (NCSL 2015).6
Previous research finds women in term limited states initially lost seats but there
were states where females gained seats. Looking at the party of the female candidates can
6There are 4 females from a third party and 10 females from Nebraska where elections are non-partisan.
51
help explain these mixed findings. If term limits create an incentive structure that is more
conducive to a Republican candidates (Fiorina 1994; 1996), term limits can help increase
female descriptive representation. A loss in the aggregate may be because female candidates
are generally Democrat and fewer female Democrats won after term limits. Yet, term limits
may have caused a greater number of female Republicans to run and win.
Since most females in office are Democrats, term limits should not have a significant
impact on their success. While term limits may have initially decreased the number of women
in office, the overall percent of female Democrats is still fairly high, and stagnant. As of 2016,
about 60% of female legislators are Democrats and about 40% are Republicans (NCSL). The
female Democratic candidate pool seems to be much richer than the Republican pool since
there are more female Democrats than there are female Republicans. Therefore, despite
term limits, a significant number of female Democrats are willing to run (as evidence from
Chapter 1) and replace females leaving due to term limits.
Therefore,
H1: Female Republicans in term limited states will have higher success rates than female
Republicans in non term limited states.
H2: Term limits will have no impact on the success rates of Democratic female candidates.
The variation across states may be explained by the level of professionalization in the
legislature as well as the political environment in which female candidates run. The political
environment in which female candidates run also has an impact on whether party identifi-
cation will help or hinder their chances of success. Below I discuss how professionalization
and term limits affect female candidates within each party.7
7Chapter 3 of the Dissertation discusses more of the political environment aspects; mainly the districtdemographics but also open seats in ‘core’ or ‘safe’ Republican/Democrat districts versus competitive/swingdistricts.
52
The research on whether more professional legislatures have a greater number of
females is mixed, (see Squire 1992; Sanbonmatsu 2002). I examine the interactive effect
term limits and professionalization have on female descriptive representation by considering
how the interaction effects each party differently.
Since more professionalized legislatures8 favor Democrats (Fiorina 1994; 1996), the
impact of term limits on female descriptive representation in a highly professionalized legisla-
ture should be negative, but minimal. The number of females in professionalized legislatures
is likely to be high and compromised of mainly female Democrats rather than female Re-
publicans. Further, a more professionalized legislature with term limits will have fewer
Democratic females post term limits. If term limits create an environment more friendly to
Republican candidates, the number of female Democrats will likely decrease.
Citizen legislatures9 are comprised of more Republicans than Democrats (NCSL
2015). In general, these legislatures have fewer females overall since Republican females
do not have as many seats as Democratic females. The incentive structure for candidates
does not change much with the implementation of term limits since the legislature is already
a part-time job. As such, this environment creates one that is more favorable to female
Republicans and less favorable to female Democrats. There are fewer female Democrats in
these citizen legislatures so term limits are unlikely to have any effect on the descriptive
representation of female Democrats.
Hybrid legislatures fall somewhere in between the highly and less professionalized
legislatures and (NCSL 2015) will likely see the biggest change in female descriptive rep-
8Professional legislatures tend to be full-time, well paid, and have a large number of staff available (NCSL2015; Squire 1992).
9Citizen legislatures are generally part-time, low pay, and have a small staff available to legislators (ibid).
53
resentation. Female politicians in these legislatures may or may not be career politicians.
Therefore, implementing term limits may have a negative effect on the overall number of
females in office— especially for career politicians. The nature of term limits makes the
legislature become a bit less professionalized which will decrease the number of Democrats
(Meinke and Hasecke 2003). The number of female Democrats is likely to decrease while the
number of female Republicans should stay relatively constant.
Therefore, the overall effect of professionalization, and the interactive effect between
term limits and professionalization, is hypothesized as:
H3: As professionalization increases, female Democratic descriptive representation increases.
H3a Female Democratic descriptive representation decreases as professionalization increases
and term limits are implemented.
H4: As professionalization increases, female Republican descriptive representation decreases.
H4a Female Republican descriptive representation increases as professionalization decreases
and term limits are implemented.
Data and Design
As with the emergence chapter, I use a candidate-level dataset of general elections
to test my hypotheses and add state-level variables for controls. I use the State Legislative
Election Returns (SLER): 1967-2010 (ICPSR 2014) to capture candidate-level state legisla-
tive elections. The large time span offers the opportunity to better understand the impact of
term limits on female descriptive representation. Using a quasi-natural experimental design
examining both pre and post term limits across all state legislatures, states without term
limits are the control group.
I coded each candidate’s gender and cross-referenced with CAWP (Center for Ameri-
54
can Women and Politics) and NCSL (National Conference of State Legislatures) to check the
total number of female candidates and female legislators by year. If there were discrepancies
between ICPSR and CAWP, I relied on each State’s Election returns to edit or confirm the
data.
The time period for this analysis is 1990-2010. The data captures 1992, the year of
women, and also a considerable amount of time both pre and post term limits took effect in
states (The first wave of term limits began in 1998). Capturing the year of the women in the
dataset is important since 1992 was a year which greatly increased the number of women in
office nationally. Lastly, my design uses 49 states10 and each district within the states for
this analysis to capture systemwide trends.
Method
To test the hypotheses I run two logistical regressions that predict the likelihood of
a winner in a general election: One for Democrats and one for Republicans. This allows me
to analyze how female descriptive representation within each party is unique. Since these
models are dichotomous, using an estimation tool such as the logit model will allow me
to capture the likelihood a winner is female. For the models, the unit of analysis is the
candidate-year. Each candidate, for each election year, is an observation. I also run and
discuss predicted probabilities to substantively interpret the results.
Variables
Dependent
The dependent variable is candidate success. The variable measures each candidate
individually where ‘0’ is a loss and ‘1’ is when the candidate won. Table 3.1 reports the
10I eliminate Nebraska from the models since the elections are nonpartisan
55
success rate of all the female candidates in the sample. For all states, about 60% of females
run and win their seat. In term limited states, 54% of females won and in non term limited
states, 60% of females won their race. While these are descriptive statistics, the pattern
suggests candidates do not win as often in term limited states (similar for male candidates).
Table 3.2 reports the descriptive statistics for male winners, which shows men only winning
about 55% of the time in term limited states.
Variable Number Percent
Overall Win 14,494 59.07%Overall Loss 10,043 40.93%Term Limit Win 2,107 54.15%Term Limit Loss 1,784 45.85%No Term Limit Win 12,387 60.00%No Term Limit Loss 8,259 40.00%
Table 3.1: Descriptive Statistics for Female Success and Term Limits
Variable Number Percent
Overall Win 35,118 60.15%Overall Loss 23,265 39.85%Term Limit Win 5,596 55.98%Term Limit Loss 4,401 44.02%No Term Limit Win 23,265 60.15%No Term Limit Loss 35,118 39.85%
Table 3.2: Descriptive Statistics for Male Success and Term Limits
Independent Variables
The first explanatory variable is term limits which is coded ‘1’ for states with term
limits and states without term limits are coded ‘0’. States are coded as ‘0’ until the im-
plementation of term limits for that particular state begins. I expect term limits to have a
positive effect on candidate success.
Candidate gender is the second main independent variable. Gender is coded as ‘0’
for males and ‘1’ for females. I also include an interaction between term limits and female
56
candidates to examine how females are specifically affected by term limits. I hypothesized
(H1 & H2) term limits will have a positive effect on female Republicans and no effect on
female Democrats.
Professionalization is third explanatory variable. The measure is from the Squire
index and comes from Vanderbilt University’s State Politics and Judiciary dataset. The
measure is a continuous variable from 0-1 where 0 is the least professionalized and 1 is the
most professionalized. The measure takes into account whether the legislature is full-time
versus part-time, how much legislators are paid, staff size, etc. For more details on the mea-
sure, see Squire 1992. I expect professionalization to have a positive effect on Democrats
and a negative effect on Republicans. I also include an interaction between professionaliza-
tion and female candidate to test hypotheses H3 & H4. I expect professionalization to be
positively linked to female Democratic descriptive representation and negatively linked to
female Republican descriptive representation.
A 3-way interaction between female candidates, term limits, and professionalization
is included in both Republican and Democratic models. I expect term limits in more profes-
sionalized states to have a negative impact on female Democratic descriptive representation.
As professionalization decreases I expect female Republican descriptive representation to
increase.11
Controls
Incumbency is an important control variable to include in the models. Even with the
enactment of term limits there will be incumbents with a high electability rate. Incumbency
is coded as a dichotomous variable where ‘0’ is a non-incumbent and ‘1’ is an incumbent. I
11The relationship between the 3-way interaction will be discussed more thoroughly through post estimationpredicted probabilities
57
expect incumbency to have a positive impact on the success rates of candidates.
I also include a control for open seats where ‘0’ is a seat with an incumbent running
and ‘1’ is an open seat. Term limits create open seats but it is important to control for open
seats in non term limited states. I expect open seats to have a positive impact on candidate
success rates.
To control for the electoral competitiveness of a seat I use two measures. The first is
unopposed coded as ‘1’ if the candidate ran unopposed, ‘0’ if otherwise. The second measure
is the margin of victory for the winning candidate. A higher margin of victory indicates the
race was less competitive.
Lastly, I include a control for the district type which is a variable based on single
member versus multimember districts12 The variable is coded as ‘0’ for single member dis-
tricts and ‘1’ for multimember districts. I expect this variable to be positive, indicating
multimember districts are beneficial for candidate success rates.
Results
Table 3.3 reports the results for the four Democratic Models with roust standard
errors clustered around each candidate.13 Model 1 includes no interaction terms and Models
2-4 include interaction terms. Model 1 serves as a test for the extant literature on how term
limits, and professionalization affect the success of Democratic candidates. Model 1 finds
term limits have a significant and negative effect on Democratic candidate success. In term
limited states, Democrats are less likely to win; this is consistent with literature finding term
12States with MMDs: Alaska, Arizona*, Arkansas, Georgia, Idaho*, Indiana, Maryland, New Hampshire,New Jersey*, North Carolina, North Dakota*, South Dakota*, Vermont, Washington*, West Virginia,Wyoming. Note: states with stars are two-member districts (NCSL 2016) while other states may haveall or only some MMDs.
13I also ran each model clustered around the state legislative district and the results were consistent.
58
limits favor a more Republican legislature. Model 1 also supports the state politics literature
findings that professionalized legislatures favor Democrats. The direction on the Squire
measure is positive and significant indicating as professionalization increases, Democratic
candidate success increases. Further, Model 1 also shows being a female positively and
significantly increases the likelihood of success.
Variable Model 1 Model 2 Model 3 Model 4
Term Limit -0.183* (0.046) -0.178* (0.054) -0.184* (0.046) -0.145 (0.093)Female Candidate 0.111* (0.038) 0.114* (0.044) 0.045 (0.070) -0.002 (0.081)Professionalism (Squire) 0.439* (0.122) 0.440* (0.122) 0.346* (0.144) 0.387* (0.167)Term*Female – -0.014 (0.090) – 0.124 (0.162)Female*Professional – – 0.298 (0.259) 0.531 (0.313)Term Limit*Professional – – – -0.137 (0.313)Term*Female*Professional – – – -0.637 (0.548)Incumbent 4.940* (0.055) 4.940* (0.055) 4.940* (0.55) 4.942* (0.055)Open Seat 2.596* (0.050) 2.596* (0.050) 2.596* (0.050) 2.599* (0.050)Unopposed 5.0517* (0.722) 5.052* (0.722) 5.049* (0.722) 5.049* (0.722)Margin of Victory 0.022* (0.001) 0.022* (0.001) 0.022* (0.001) 0.022* (0.001)District Type 0.058 (0.089) 0.058 (0.089) 0.058 (0.089) 0.062 (0.089)Constant -2.981* (0.092) -2.982* (0.092) -2.961 (0.094) -2.978* (0.097)
n =36855R2=0.54
Controls for election years included in the model but suppressed from the table
Table 3.3: Logistical Regression of Democratic Candidate Success
Model 2 tests hypothesis 2 which states term limits will have no effect on female
Democrats. Hypothesis 2 is supported since the interactive effect between term limits and
female candidates is not statistically significant. Term limits are negative and statistically
significant, and female is positive and significant, but there is not an interactive relationship.
Further, the model is consistent with Model 1. Professionalism has a significant and positive
relationship on Democratic success.
Models 3 and 4 test hypothesis 3. Hypothesis 3 expects a positive relationship be-
tween professionalization and female candidates. Model 3 does not fully support hypothesis
59
3. While the coefficient is in the expected direction, the significance falls just outside the
conventional .05 level. Hypothesis 3a deals with the negative interactive effect term limits
have on female candidates and professionalization. As I hypothesize, the interaction is in
the expected direction and just on the cusp of significance. The interactive relationship sup-
ports my theoretical argument that term limits change the incentive structure in a way that
does not favor Democrats— especially female Democrats. When term limits are present,
professionalization decreases and the likelihood a Democratic female will win the election
decreases.
Since logistical regression outputs are not directly interpretable, I run predicted prob-
abilities for models to further interpret the results. I run the predicted probabilities based
on Model 4 to more closely examine how the relationship between term limits and female
Democrats changes with level of professionalization. Further, I set margin of victory to the
mean, which is 42.71%; district type to single member districts, for seats that are open and
not unopposed. I vary the level of professionalization to reflect a citizen legislature (set
to .100 on Squire), a hybrid (.300) and professional (.600). Table 3.4 and Figure 1 below
presents the results for these predicted probabilities.
Female Winner Citizen Legislature Hybrid Professionalized
No Term Limits 0.619 (0.011) 0.639 (0.010) 0.669 (0.014)Term Limits 0.575 (0.014) 0.596 (0.013) 0.627 (0.015)
All margins are significant at p < .05 Std error provided in parentheses
Table 3.4: Predicted Probabilities for Female Democrats
60
Figure 3.1. Female Democrat Success
Table 3.4 provides further evidence to support hypothesis 3 and 3a. As professional-
ization increases, the probability a female will win the seat increases. The predicted proba-
bilities also indicate that at the same level of professionalization, women in the term limited
states are less successful at winning their seat in office. In a citizen legislature, a female is
predicted to win the seat 62% of the time versus 58% of the time for a female in a term
limited state. The difference is also present in hybrid legislatures. Females in states with
no term limits have a 64% predicted probability and females in term limited state only have
a 59% predicted probability to win the seat. Lastly, in professionalized legislatures, female
Democrats are also at a disadvantage in term limited states. In state without term limits the
predicted probability a female will win the election is about 70% compared to states with
term limits where the probability is at 63%.
Figure 3.1 graphically shows the effect term limits and professionalization have on the
probably a female Democrat will win. The table provides confidence intervals to further show
61
the statistically significant differences between term limits and non term limits. Further,
the figure also shows the differences across levels of professionalization are also significant.
Female Democrats in citizen legislatures have the lowest probability of winning and this
probability is statistically different from the hybrid as well as the most professional.
Overall, Table 3.4 and Figure 3.1 indicate that highly professionalized legislatures in
states without term limits are where female Democrats are the most successful. Females are
least likely to win their seat in states with citizen legislatures and term limits. Further, the
predicted probabilities show female Democrats are more likely to win in states without term
limits. A female candidate’s likelihood of success increases as professionalization increases,
but term limits decrease their likelihood of winning. The margin between term limited and
non term limited states is about 5% for each level of professionalization. The margin between
a citizen legislature, no term limits, and a professionalized, with term limits, is only about
1% point. The small margin further indicates that term limits have a negative effect on
female candidates and this relationship is an interactive one based in part on the level of
professionalization.
Moving to the Republican candidates, Table 3.5 below also includes 4 models based
on the same explanatory and control variables as the Democratic candidate models. The
standard errors are robust standard errors clustered around the candidate. Model 1 is the
simple model that does not include any interactions. Term limits do not have a statistically
significant impact on the likelihood a Republican will win. While the direction is positive
indicating term limits seem to help the likelihood for a Republican win, the level of statistical
significance falls just outside the traditional levels of significance. Further, being a female
candidate does not have a statistically significant effect on the likelihood the success. This
62
finding seems to suggest being a female Republican neither increases or decreases the likeli-
hood of success. Since professionalization is significant and in a negative direction, Model 1
supports the extant literature finding Republicans are likely to win seats in legislatures with
lower levels of professionalization.
Variable Model 1 Model 2 Model 3 Model 4
Term Limit 0.063 (0.047) 0.136* (0.052) 0.064 (0.047) 0.005 (0.089)Female Candidate -0.003 (0.043) 0.069 (0.048) 0.039 (0.079) 0.022 (0.090)Professionalism (Squire) -0.495* (0.122) -0.491* (0.122) -0.455* (0.137) -0.637* (0.164)Term*Female – -0.347* (0.104) – -0.101 (0.164)Female*Professional – – -0.190 (0.287) 0.219 (0.337)Term Limit*Professional – – – 0.557 (0.296)Term*Female*Professional – – – -1.014 (0.606)Incumbent 5.012* (0.055) 5.014* (0.055) 5.012* (0.055) 5.015* (0.055)Open Seat 2.407* (0.042) 2.408* (0.043) 2.407* (0.043) 2.406* (0.043)Unopposed 7.504* (1.010) 7.504* (1.001) 7.504* (1.001) 7.501* (1.010)Margin of Victory -0.008* (0.001) -0.008* (0.001) -0.008* (0.001) -0.008* (0.001)District Type 0.046 (0.085) 0.043 (0.085) 0.045 (0.085) 0.040 (0.085)Constant -2.375* (0.088) -2.392* (0.088) -2.384* (0.089) -2.355* (0.093)
n =34436R2=0.54
Robust standard errors clustered around legislative districtControls for election years included in the model but suppressed from the table
Table 3.5: Logistical Regression of Republican Candidate Success
Model 2 provides the results for hypothesis 1 which speculates female Republicans
in term limited states will have a higher likelihood of winning. The negative and statis-
tically significant coefficient on the interaction indicates the hypothesis is not supported.
Republican females in term limited states are less likely to win.
Models 3 and 4 relate to hypotheses 4 and 4a dealing with professionalization and
likelihood of winning. Hypothesis 4 states that as professionalization increases, female suc-
cess will decrease. Model 3 shows the interactive relationship between female candidates
63
and professionalization. The direction is negative and in the expected direction, but the
relationship is not significant, and the interaction is just outside the cusp of significance.
Model 4 presents the 3-way interaction between term limits, female candidates and profes-
sionalization. Hypothesis 4a states that term limits, and lower levels of professionalization
will help female candidates win and is supported by the model. The negative coefficient
(indicating a lower level of professionalization, which is favorable to Republicans) increases
the likelihood a female will win office. The 3-way interaction is in the expected direction
and is statistically significant at a .10 level.
To interpret the results beyond direction and statistical significance, Table 3.6 and
Figure 3.2 provide the predicted probabilities for Republican females varying by level of
professionalization. As with the Democratic predicted probabilities, the the results are for
open, single member districts in opposed seats and the margin of victory is set to the mean
at 40% for Republicans.
Figure 3.2. Female Republican Success
64
Female Winner Citizen Legislature Hybrid Professionalized
No Term Limits 0.501 (0.011) 0.477 (0.012) 0.440 (0.016)Term Limits 0.517 (0.015) 0.493 (0.014) 0.455 (0.017)
All margins are significant at p < .05 Std error provided in parentheses
Table 3.6: Predicted Probabilities for Female Republicans
Table 3.6 and Figure 3.2 above show further support for hypothesis 4a. In general, as
professionalization in the legislature increases, the predicted probability for female success
decreases. Further, term limits have a positive effect on the predicted probabilities for female
candidates success. Overall, females have a higher predicted probability for success in states
with term limits.
In citizen legislatures, females are predicted to win about 50% of the time in non term
limited states versus about 52% of the time in term limited states. This pattern continues
for hybrid legislatures where females are predicted to win about 48% of the time in non term
limited states versus 50% of the time in term limited states. Lastly, female Republicans have
a 44% predicted probability of winning in highly professionalized, non term limited states
and about a 46% probability in term limited states.
Figure 3.2 includes 95% confidence intervals for each level of professionalization.
Model 4 is significant at the .10 level so there is some overlap in the intervals. Overall,
across levels of professionalization, the probability of a female winning is statistically sig-
nificant. Females are less likely to win in the highly professionalized legislatures and their
probability of success is significantly lower when compared to both hybrid and citizen leg-
islatures. The difference term limits make within legislatures is only significant at the .10
level. The difference between term limits and no term limits, as discussed in Table 3.6’s
results, is only about 2% for each level of professionalization.
65
Lastly, across both Democratic and Republican models, all control variables were in
the expected direction. Incumbency, open seat and unopposed races were all positive and
highly significant. Margin of victory is negative and also statistically significant. For both
the Democratic and Republican models, the only control variable not statistically signifiant
is the district type variable which controls for the multimember districts.14
Conclusions and Implications
This chapter discusses the next step in the process of female descriptive representation
in the states; election outcomes. The first chapter finds term limits have a positive effect
on the likelihood a female candidate will emerge, and this chapter presents results on the
likelihood a female will win. Using state politics literature, which looks at incentive structures
for candidates in term limited states, and women in politics literature, which finds when
women run they win, I develop a theory suggesting term limits will increase female descriptive
representation. Further, I examine this by party and find term limits increase the likelihood
a female Republican will win office but decrease the likelihood a female Democrat will win.
Overall, the results compliment congressional literature that finds when women run,
they win. This seems to be the case at the state-level since women are likely to win their
seat 50% of the time, and even more likely if the female is a Democrat. Female Democrats
are predicted to win at high rates (term limits or no term limits, their predicted success
rates range from 58%-67%) and female Republicans are also predicted to win in the range of
about 44%-52%. Further, term limits increase the likelihood a female Republican will win,
14The Appendix includes a logistical regression for Republicans and Democrats in the same model. Toavoid a 4-way interaction, the dependent variable is female winner and party identification replaces femalecandidate. The results are consistent and the new three way interaction between party, term limits andprofessionalization is positive but not statistically significant. These findings are further indication thatmodels separating Republican and Democratic females are useful since Republican females tend to be washedout by Democratic female effects.
66
by about 2%, suggesting term limits can increase descriptive representation of women.
Term limits have a negative effect on female Democrats but their likelihood of winning
is still high. The high predicted probabilities of success implies that if female Democrats
continue to run, they will continue to win. Since Democratic females hold more seats than
Republican females, term limits should increase the descriptive representation overall. Im-
plications on the long term effect of term limits on descriptive representation should continue
to be examined since parties may adapt new strategies in order to adapt. For example, if
Republican females do well in term limited states, a new strategy may be to actively recruit
more female Republicans to office.
Future research should look to explain the large difference in success rates between fe-
male Republicans and female Democrats. Explanations on the likelihood of female candidate
success differences may be due to strategic candidate entry, campaign finance differences, or
even candidate quality. A more nuanced look at some of these campaign-level/ candidate
specific explanatory factors may be useful.
The next chapter will examine a new set of explanatory variables to determine if
location effects the likelihood a female will emerge and win office. I theorize on how district
demographics such as urbanization, education levels, etc. play a role in the emergence as
well as descriptive representation of female candidates.
67
[1] Allebaugh, Dalene, and Neil Pinney. 2003. ”The Real Costs of Term Limits: Comparative Study ofCompetition and Electoral Costs.” in The test of time: Coping with legislative term limits.
[2] Ansolabehere, Stephen, and James M. Snyder Jr. 2002. “The incumbency advantage in US elections:An analysis of state and federal offices, 1942-2000.” Election Law Journal 1.3: 315-338.
[3] Bledsoe, Timothy and Mary Herring. 1990. “Victims of Circumstance: Women in Pursuit of PoliticalOffice.” The American Political Science Review. 84:213-223.
[4] Burrell, Barbara C. 1994. A Woman’s Place Is in the House: Campaigning for Congress in the FeministEra. Ann Arbor: University of Michigan Press.
[5] Carey, John M., Richard G. Niemi, and Lynda W. Powell.2000. Term limits in state legislatures.University of Michigan Press.
[6] Carroll, Susan J., and Linda MG Zerilli. 1993. “Feminist challenges to political science.” Politicalscience: The state of the discipline II: 55-76.
[7] Carroll, Susan and Krista Jenkins. 2001. “Do Term Limits Help Women Get Elected?” Social ScienceQuarterly. 82/1.
[8] Census Bureau. 2012.
[9] Center for American Women and Politics. 2012. “Women in State Legislatures? Fact Sheet”:www.cawp.rutgers.edu .
[10] Cox and Katz. 1996. “Why did the incumbency advantage in U.S. House elections grow?” AJPS 40:478-497.
[11] Crane, Edward H. 1995. “Testimony before the Subcommittee on the Constitution, Committee on theJudiciary, U.S. Senate.”
[12] Darcy, R, Susan Welch and Janet Clark. 1994. Women, Elections, and Representation. 2nd Edition.Lincoln: University of Nebraska Press.
[13] Dolan, Kathleen. 2004. Voting for Women: How the Public Evaluates Women Candidates. Boulder,CO: Westview.
[14] Dolan, Kathleen, and Lynne E. Ford. 1998. “Are All Women State Legislators Alike?.” Women andelective office: Past, present, and future: 73-86.
[15] Dolan, Kathleen. 2009. “The Impact of Gender Stereotyped Evaluations on Support for WomenCandidates.” Political Behavior. 32:69-88.
[16] Elder, Laurel. 2012. “The Partisan Gap among Women State Legislators.” Journal of Women, Politicsand Policy. 33:65-85.
[17] Erikson, Robert. 1971. “The Advantage of Incumbency in Congressional Elections”.Polity. 3: 395-405.
[18] Farmer, R., Rausch, J.D. and Green, J.C., 2003. The test of time: Coping with legislative term limits.Lexington Books.
[19] Ferry, Jonathan. 1994. “Women, Minorities and Term Limits: American’s Path to a RepresentativeCongress”. U.S. Term Limits Foundation Outlook Series.
[20] Fiorina. 1977. “The Case of the Vanishing marginals: The bureaucracy did it” American Political
68
Chapter Bibliography
Science Review. 71:177-81.
[21] Fowler, Linda.1992. A Comment on Competition and Careers. In Gerald Benjamin and Michael J.Lambin. eds., Limiting Legislative Terms Washington, DC: CQ Press.
[22] Foerstel, Karen, and Herbert N. Foerstel. 1996. Climbing the hill: Gender conflict in Congress.Greenwood Publishing Group.
[23] Fox, Richard L. 2000. “Gender and Congressional Elections.” In Gender and American Politics:Women, Men, and the Political Process, eds. Sue Tolleson Rinehart and Jyl J. Josephson. Armonk,NY: M.E. Sharpe.
[24] Fox, Richard and Jennifer Lawless. 2004. “Entering the Arena? Gender and the Decision to Run forOffice.” American Journal of Political Science. 48:264-280.
[25] Fulton, Sarah et, al. 2006. “The Sense of a Women: Gender Ambition, and The Decision to Run forCongress”. Political Research Quarterly. 59: 235-248.
[26] Fulton, Sarah. 2012. “Running Backwards and in High Heels: The Gendered Quality Gap andIncumbent Electoral Success.” Political Research Quarterly. 20:1-12.
[27] Gershon, Sarah. 2012. “When Race, Gender, and the Media Intersect: Campaign News Coverage ofMinority Congresswomen” Journal of Women and Politics. 33:105-125.
[28] Gordon, Sanford C., Gregory A. Huber, and Dimitri Landa. 2007. “Challenger entry and voterlearning.” American Political Science Review 2: 303-320.
[29] Huddy, Leonie, and Nayda Terkildsen. 1993. “Gender Stereotypes and the Perception of Male andFemale Candidates.” American Journal of Political Science. 37: 119-47.
[30] Jacobson and Kernell. 1983. “Strategy and choice in Congressional elections.”
[31] Kazee, Thomas.1994. “Who Runs for Congress ?” Washington, DC: Congressional Quarterly.
[32] Kahn, Kim Fridkin. 1992. ”Does Being Male Help? An Investigation of the Effects of Candidate Genderand Campaign Coverage on Evaluations of U.S Senate Candidates” Journal of Politics 54: 497-517.
[33] King, David and Richard E. Matland. 2003. “Sex and the Grand Old Party: An ExperimentalInvestigation of the Effect of Candidate Sex on Support for a Republican Candidate.” AmericanPolitics Research. 31: 595-612.
[34] Koch, Jeffrey. 2000. “Do Citizens Apply Gender Stereotypes to Infer Candidates’ Ideological Orienta-tions?” Journal of Politics 62: 414-29.
[35] Koch, Jeffrey. 2002. “Gender Stereotypes and Citizens’ Impressions of House Candidates’ IdeologicalOrientations.” American Journal of Political Science. 46(2), 453-462.
[36] Lindquist, S. A. 2007. “State Politics and the Judiciary Codebook. Predictability and the Rule of Law:Overruling Decisions in State Supreme Courts.” Nashville, TN: National Science Foundation GrantSES #0550618.
[37] McDermott, Monika L. 1997. “Voting cues in low-information elections: Candidate gender as a socialinformation variable in contemporary United States elections.” American Journal of Political Science:270-283.
[38] Meinke, Scott R., and Edward B. Hasecke. 2003. “Term limits, professionalization, and partisan controlin US state legislatures.” Journal of Politics 65: 898-908.
[39] Milyo, Jeffrey and Samantha Schosberg. 2000. “Gender Bias and Selection Bias in House Elections.”
69
Public Choice. 105:41-59.
[40] Mooney, Christopher. 2009. “Term Limits as a Boon to Legislative Scholarship: A Review.” StatePolitics and Policy Quarterly. 9:204- 228.
[41] National Conference of State Legislatures. 2011 “Legislative Term Limits Overview?http://www.ncsl.org/ .
[42] Osborn L, Tracy. 2012. How Women Represent Women: Political Parties, Gender and Representationin the State Legislatures. Oxford University Press.
[43] Palmer, Barbara and Dennis Simon. 2005. “When Women Run Against Women: The Hidden Influenceof Female Incumbents in Elections to the U.S. House of Representatives, 1956-2002.” Gender andPolitics. 1:39-63.
[44] Palmer, Barbara and Dennis Simon. 2006. Breaking the Political Glass Ceiling: Women and Congres-sional Elections. New York: Routlege.
[45] Reed, W. Robert, and D. Eric Schansberg. 1996. ”An analysis of the impact of congressional termlimits on turnover and party balance.” In Legislative Term Limits: Public Choice Perspectives, pp.129-143. Springer Netherlands.
[46] Rosenwasser, Shirley Miller, and Norma G. Dean. 1989. “Gender role and political office.” Psychologyof Women Quarterly 13: 77-85.
[47] Sanbonmatsu, Kira. 2002. “Political parties and the recruitment of women to state legislatures.”Journal of Politics 64: 791-809.
[48] Sanbonmatsu, Kira. 2006. Where women run. Ann Arbor: University of Michigan Press.
[49] Sanbonmatsu, Kira and Kathleen Dolan. 2009. “Do Gender Stereotypes Transcend Party?” PoliticalResearch Quarterly 62: 485-494.
[50] Schwindt-Bayer, L. A. and Mishler, W. 2005. “An Integrated Model of Women?s Representation.” TheJournal of Politics. 67/2: 407-428.
[51] Seltzer, Richard A., Jody Newman, and Melissa Voorhees Leighton. 1997. Sex as a Political Variable:Women as Candidates and Voters in U.S. Elections. Boulder: Lynne Rienner.
[52] Swers, Michele. 2002. The Difference Women Make. Chicago: University of Chicago Press.
[53] Werner and Mayer. 2007. Public Election Funding, Competition, and candidate Gender.
70
CHAPTER 4
WOMEN-FRIENDLY DISTRICTS
Abstract
This chapter explores the possibility of women-friendly districts in U.S. state legis-
latures. Examining the lower chamber in each state, I model the theory after Palmer and
Simon’s (2006) seminal work on congressional women-friendly districts. Further, I theorize
on how each state’s electoral variations have the possibility of helping districts become more
women friendly. Using logistical regression, I find women-friendly districts are present in the
states and as districts become more women-friendly, females are more likely to run and win
office.
Introduction
The first and second chapter examine the positive effect term limits have on female
candidates. Specifically, I examine whether term limits change the incentive structure for
candidates to seek and win office. The findings suggest, contrary to initial studies (Carroll
and Jenkins 2001; Bernstein and Chad 2003), that term limits are beneficial to females. Term
limits increase the likelihood of female candidate emergence and success in a term limited
states. While the two previous chapters focused mainly on the candidate-level electoral
context of the race, this chapter includes district demographics as part of picture.
As previously discussed in the first two chapters, the number of females in state
legislatures has become rather stagnant since the early 1990s. Yet, there is exists a puzzle
in that there are difference across the states– women are gaining seats overtime in some
states but losing seats in other states. For example, as of 2015, Vermont’s legislature is 41%
females. The percent of females has increased by about 10% over the last ten years, which is
71
counter to the overall pattern of stagnation seen across all 50 states (CAWP 2015). Further,
women hold the most seats in western states that include Colorado, Washington, Oregon,
Montana, Arizona and Nevada (CAWP 2015). These patterns across the states should be
explored to determine whether or not this stagnation in descriptive representation is likely
to continue.
At the congressional level, Palmer and Simon (2006) created an index which uses
district demographic variables to predict the ‘women-friendliness’ of a district. The idea is
that certain districts are historically better for women since women in these districts tend
to run and win at higher rates than women in other districts. Palmer and Simon (2006) find
that when a congressional district has a high average income, a more educated population,
is geographically small and in a racially diverse and urban area, female candidates are much
more likely to win office. This chapter seeks to expand our understanding of what constitutes
a ‘women-friendly’ electoral environment by examining state legislative elections. I explore
whether district variation in demographics, as well as state legislative variation(in the form of
professionalization and term limits) also help explain where female candidates are emerging
and winning.
Further, I will determine whether term limited states are more or less women-friendly.
Since results from chapter one and two indicate that women are more likely to run and win
in term limited states, I propose there may be a positive interactive relationship between
term limits and women-friendly districts. Term limited districts with high levels of women-
friendliness, will be more likely to have female candidates emerge and win than non term
limited districts.
This chapter proceeds as follows, I will discuss the extant literature on demographic
72
characteristics effecting female descriptive representation in state legislatures. Next, I of-
fer a theory that incorporates how state legislative characteristics such as term limits and
professionalization condition the impact the friendliness of a district has on emergence and
descriptive representation. Then, I discuss the design and method I use to test my theory.
Then I run a series of logistical regressions with predicted probabilities to substantively in-
terpret how female candidates are effected by the women-friendliness of a district. Lastly, I
conclude with a brief discussion of the results and implications this chapter has on future
work.
Literature
Fenno’s (1978) observational study emphasizes the important effect of district demo-
graphic characteristics on a representative’s behavior. Each representative recognizes their
district has a unique set of demographic characteristics and that these differences matter for
reelection. The representatives emphasize geography, diversity, jobs, and incomes; Most of
which Palmer and Simon (2006) use in predicting women-friendly districts and will also be
discussed in this chapter.1
Below, I discuss four of the main explanations in the extant literature for why can-
didates tend to emerge and win office in recognizable patterns. Utilizing both state, and
congressional level literature, the focus is how the geography, party, diversity, and socioe-
conomic make-up of state legislative districts influence candidates and their election into
office.
1Rule’s (1981) cross-national analysis focuses on state legislatures and is an early study pointing to contextualdistrict factors that influence females and where they run for office.
73
Political Culture
Political culture is an underlying explanation for regional similarities and differences
across the states (Elazar 1966). And while there is large debate over the measures and with-
standing of distinct political cultures, Elazar’s (1966) three state political culture classifica-
tion is still the leading typology. The three political cultures are Traditionalistic, Moralistic,
and Individualistic. In the context of women in office, states with the same political culture
have similar patterns of women gaining and maintaining office overtime.
Traditionalistic states are typically southern states where social connectedness mat-
ters and politics is for the elite. Further, party politics is not particularly strong and the
driving force behind recruitment; rather, families are seen as the important recruitment
ground for future candidates (Elazar 1966; 1984). Women in these states have seen the most
growth in female legislators overtime (Norrander and Wilcox 2014). Women gaining seats in
the more traditionalistic states such as Texas and Alabama is most likely due to the fact that
their numbers were so few to begin with, there is a lot of ground to be covered.2 For example,
Texas and Alabama’s legislatures were 16% and 5% in 1993, respectively. Both were lower
than the national average of around 20% at the time. By 2015, the percents increased to
20% and 14%, respectively. Again, both states are still below the overall average of 24%
females legislatures, but are increasing overtime (Norrander and Wilcox 2014; CAWP 2015).
States that have a Moralistic political culture are usually northern states, where
politics is for the good of society and politicians are public servants (Elazar 1966). Women
have done well gaining office in Moralistic states (Norrander and Wilcox 2014). For example,
Vermont and Colorado are both moralistic and regularly rank in the top 10 states for their
2Explanations for few females in office, especially in the South sometimes center around the idea that femaleswere late to be included in the political system. This is generally measured by voter participation (Cassel1979)
74
high numbers of women in office. While these states have higher numbers of women in office,
they are much more stagnant. Colorado is also a term limited state and after an initial loss
in female descriptive representation, women now make up an all time high of 42% of the
seats in the state. Yet, some Moralistic states have decreased their overall number of women
in office. In 1993 Washington (not term limited) had 39% females in office, but in 2015, this
number decreases to 34% (Norrander and Wilcox 2014; CAWP 2015).
The last type of political culture is the Individualistic culture. Individualistic states
usually have traditional political party organizations, and politics is generally a game of
personal gains (Elazar 1966). Traditional party organizations in these states do not favor
females since there is a reluctance to recruit and endorse female candidates (Sanbonmatsu
2002; Norrander and Wilcox 2014). One example of a state characterized as individualistic
is Ohio. The percent of women in the legislature was particularly low (11-13% throughout
the late 1980s, early 1990s) until the mid 1990s when the percent hovered around 24. After
the implementation of term limits in 2000, the percent of females increased, decreased and
has now steadily increased since 2008 with women making up 25% of the legislature in 2015
(CAWP 2015). In general, individualistic states have also seen increases in the number of
women in office overtime.
Party
Extant literature finds party identification serves as a main cue for voters (Campbell
et al 1960; Flanigan and Zingale 1994, 2002; Lau and Redlawsk 2001; Lau and Sears 1986;
Rahn 1993). A majority of the time, Republican voters vote Republican and Democrats vote
Democrat. Yet, research finds voters view the parties through a gendered lens and there is an
intersection between the two (Anderson, Lewis and Baird 2011). Therefore, party alone is not
75
able to predict the women-friendliness of a district (Palmer and Simon 2006). Overall, voter
stereotypes of a candidate’s sex is rather complex (Anderson, Lewis and Baird 2011) and
therefore, party is an important part of the picture for predicting women-friendly districts.
Demographics
At the congressional level, research varies as to whether or not the racial diversity
of a district impacts the number of women in office. Burrell (1984) finds racial diversity
in districts have no impact on women candidates. Other studies find women do better in
districts with a more racially diverse population (Welch 1985; Palmer Simon 2006). Lastly,
research also finds women of color do well in majority-minority districts (Bedolla, Tate and
Wong 2014).
Further, urbanization and population density also effect female descriptive represen-
tation. Democratic candidates tend to win elections in densely populated, urban areas and
Republican candidates fair better in rural districts that are larger in geographic scope (Lang,
Sanchez, Berube 2008; Palmer and Simon 2006). In particular, female House members tend
to represent more urban districts (Burrell 1992; Darcy and Schramm 1977; Diamond 1977;
Rule 1981; Welch 1985; Welch et al. 1985; but see Kirkpatrick 1974; Werner 1966). Women
fair better in urban areas due to an ease in recruitment efforts because there are more poten-
tial female candidates in the area (Darcy and Schramm 1977) as well as a larger number of
seats available (Rule 1981). Further, women are perceived as being ‘better’ fit to deal social
welfare issues (Rule 1981; Koch 2000; Dolan 2009; Fulton 2011) and these issues tend to be
of great concern in urban areas, giving females an edge (Rule 1981).
76
Socioeconomic Factors
Socioeconomic factors within districts are shown to predict electoral outcomes. Amongst
them are average income, education levels within the district, marriage statistics as well as
occupations (blue-collar versus white-collar). Historically, Republicans were rather success-
ful in districts with high incomes because people with lower incomes were much more likely to
support Democratic candidates. Recent literature finds this pattern is not so clear anymore
because wealthy districts are also known to support Democratic candidates and vice versa
(Frank 2004; Gelman 2009). In regards to women in office and wealth of district, research
finds women in Congress are more likely to be elected from wealthier districts (Burrell 1994;
Nechemias 1987; Rule 1981; Welch 1985; Palmer and Simon 2006).
Education is also an important district characterization. Overtime, the education
levels of Americans has increased (also leading to a decrease in blue-collar workers). Specif-
ically, the number of of high school drop outs has decreased while the number of college
educated adults has increased (Census 2015). Female candidates are more likely to emerge
and win in areas with higher levels of education. Studies suggest this is the case because more
educated populations do not hold traditional views on women’s roles (Welch and Sigelman
1982; Arceneaux 2001; Palmer and Simon 2006). Research also finds women who run for
office are highly educated (Fox and Lawless 2004) Further, income and education are highly,
and positively correlated— since women are more likely to win office in wealthier districts,
they are also more likely to emerge and gain office in more educated districts (Palmer and
Simon 2006).
77
Electoral Variations
A point of departure from Palmer and Simon’s (2006) study on women-friendly dis-
tricts is to include a discussion on how the electoral makeup of the state’s legislature impacts
the friendliness of a district.
Literature on the impact the professionalization of the legislature has on women’s
numbers in office (largely discussed in chapter 2) has mixed findings. Squire (1992) finds the
professionalization of the legislature has little to no impact on women’s numbers in office but
Sanbonmatsu (2002) finds mixed evidence for whether professionalization impacts women’s
numbers in office. Further, my analysis in chapters one and two suggest as professionalization
increases, women’s numbers decrease. Since the findings seem to be mixed, including a
discussion on how professionalization may impact female candidates and teasing out the
relationship in further analysis is important.
Lastly, this dissertation stresses the importance of term limits on the likelihood of
female candidate emergence and success. Conventional wisdom suspects term limits will
increase female descriptive representation because females do well winning open seats. Yet,
initial studies found term limits negatively impact female numbers in office (Carroll and
Jenkins 2001; Bernstein and Chadha 2004). The theory and analysis in the previous two
chapters provides long term evidence supporting the conventional wisdom that term limits
increase female candidate emergence as well as overall numbers in office. Further, looking at
the term limited states as a whole, of the 15 states, only two (Maine and Arizona) experienced
decreases levels of women in office from 1993 to 2011 (Norrander and Wilcox 2014). The
differences in the overall percentage lost for both these states is rather small. Maine had a
3% difference and Arizona had a 1% loss in female descriptive representation.
78
Women-Friendly State Legislative Districts Theory
Palmer and Simon (2006) find women-friendly districts are distinct from party-friendly
districts. I theorize demographic characteristics making for a women-friendly district at the
state-level are similar to the congressional-level demographics. Since politicians tend to
follow a pipeline into office and gradually work their way up through the system, similar
women-friendly demographics across levels of government is a reasonable claim. Therefore,
if districts are women friendly at the congressional level, the women who were more easily
elected to Congress likely came from state legislatures with similar electability features. Be-
low I form a theory which combines demographic characteristics with institutional features
within a district and how they are likely to positively effect the women-friendliness of a
district.
District Demographics
The partisan makeup of a district is likely to effect the women-friendliness of the
district. Female Democrats make up the majority of female officeholders and therefore it
is likely these women are being elected from districts leaning more Democratic. A district
leaning Democratic is likely to increase the women-friendliness of a district but it is not
the only factor correlated with increasing whether a female candidate runs/wins. Women
friendly and party friendly are different since female Republicans can benefit from voter
stereotypes and run and win in districts favoring the Democratic Party.
Further, not all open seats (created by term limits or not) are created equally. In other
words, some open seats may be safe Democratic seats while others may be safe Republican
seats and still others seats may be competitive seats where the party in charge alternates
frequently (despite the makeup of the legislature and how ‘Democratic versus Republican’
79
it is). If a Republican seats opens, a female might not run since females are more likely to
run as Democrats. Further, if a Democratic safe seat opens, the chances of a female running
and winning may be higher since there are more female Democrats. Examining women-
friendliness in a way that captures safe versus competitive seats is important. Women across
both parties should be more successful in safe (or core) districts with high levels of women-
friendliness.
Palmer and Simon (2006) find a racially diverse district is more likely to be women-
friendly. A racially diverse district is more likely to lean Democratic in elections, especially
given that a majority of African American voters are Democrats. Given the previous chap-
ter’s discussion on voter ideological stereotypes, a liberal leaning district may help women,
especially those running as a Democrat. Furthermore, extant research finds female (and
black) legislators tend to represent black interests (see: Thomas and Welch 1991; Swers
1998; Canon 1999; Carroll 2001; Tate 2003) Therefore, minority voters may assume women
are better at representing their interests and will be more likely to elect women.
More urban areas are likely to be more women-friendly. Urban, metropolitan areas
are likely to have a larger concern for social welfare issues. Women are viewed as better suited
to handle issues of social policy which can give them an advantage in elections. Therefore,
an area that is likely to place emphasis on social policy concerns, such as urban areas, may
be more willing to elect a female over a male.
Lastly, education, income levels, and women in the workforce in a district are likely
to be important contributing factors to the women-friendliness of a district. Education levels
and income are positively correlated since the likelihood of making more money increases as
education levels increase. Districts with high levels of education, women in the workforce,
80
and/or income will be more likely to be women friendly. Educated populations have more
women working and hold less traditional views of women, which help dissipate stereotypes
voters may hold over whether women are qualified to run for office. Further, women in these
areas are likely to be educated, which may foster an environment in which females are active,
and aware of politics–making women more likely to be recruited in these areas and willing
to run for office.
Institutional Variation: Professionalization and Term Limits
The level of professionalization within a state may also condition the impact of dis-
trict’s level of women-friendliness. I suspect, as professionalization in a legislature increases,
the women friendliness in a district will decrease. This is largely due to research suggesting
women seem to be less likely to run and hold office in more professionalized legislatures.
Evidence from chapter one suggests as professionalism increases, women are less likely to
emerge. Also, chapter two finds higher levels of professionalization is beneficial for Demo-
cratic females but not Republican females.
I expect the implementation of term limits to have a positive effect on the women-
friendliness of a district. Term limits leads to higher turnover within legislatures. Since the
incumbency advantage is lessened by the implementation of term limits, females will have
more opportunities to run and win elections in districts with termed-out legislators. Running
and winning against an established incumbent is difficult; term limits decrease the electoral
advantage incumbents normally enjoy term after term.
Further, results from chapters one and two stand as further evidence that term limits
are increasing the number of females emerging and winning. An environment which creates
the opportunity for more females to run and win is one that is ‘women-friendly’ and should
81
therefore have a positive impact on the level of friendliness seen within each district.
Most importantly, is the conditional relationship term limits have on women-friendly
districts. Since the findings in the previous chapters suggest term limits increase female
emergence and success, I expect when term limits are interacted with women-friendly dis-
tricts, these districts provide females an even greater advantage. For example, in comparing
districts that are not women-friendly, having term limits increases the likelihood a female
will emerge and win than they otherwise would have in a non term limited district. Further, I
expect this relationship in the most women-friendly districts as well. Females in term limited
states will be much more likely to emerge and win than in the non-term limited states.
This interaction is important since not all vacant term limited seats are women-
friendly. Given the research finding women lost seats in the initial implementation of term
limits, it is important to examine the nature of the vacant seats. There are some geographic
areas as previously discussed that women are historically left out. Term limits conditioned on
the women-friendliness of a district is a unique way to examine the relationship between term
limits and female emergence and success. While term limits may provide a small increase
in the likelihood a female candidate will emerge and win, the likelihood of winning a term
limited seat will increase more as the women-friendliness of a district increases. Essentially,
where a female candidate runs matters and running in districts that are more women-friendly
(especially when term limited) increases the likelihood of success.
Therefore:
H1: As a state legislative districts becomes more women-friendly, the likelihood a female
candidate will emerge increases.
H2: As a state legislative districts becomes more women-friendly, the likelihood a female
82
candidate will win increases.
H3: Female candidate emergence increases as the women-friendliness of a district increases,
and term limits are implemented.
H4: Female candidate success increases as the women-friendliness of a district increases, and
term limits are implemented.
Data and Design
The unit of analysis is the candidate-year and I examine general elections in single
member districts. To measure the effect women-friendly districts have on female candidates,
I use an original dataset that combines candidate-level data with state legislative district
data. The candidate-level data comes largely from the State Legislative Election Returns
(SLER): 1967-2010 (ICPSR 2014). For this analysis, I use the time period from 2000-20103
and coded each candidate’s gender into the SLER dataset.4 The state legislative district
demographics data comes from the U.S. Census Bureau. Using the 2000 Sample Data on
State-Legislative Districts (Lower House), I compiled urbanization, diversity, income, women
in the workforce, and education into a dataset which was then merged with the SLER data.
Method
To test the hypotheses, I first ran a series of means tests to determine the difference
across parties as well as across gender (See Appendix). Palmer and Simon (2006) note
women-friendly districts are unique from party-friendly districts, and the difference in means
tests determines which explanatory variables are unique to gender. For presentation of the
3I use this single time period of 2000-2010 to control for redistricting.
4As a robustness check, I reference CAWP (Center for American Women and Politics) and NCSL (NationalConference of State Legislatures) databases to check total women candidates and women in office by year.
83
effect women-friendly districts have on female emergence and success, I compile a women-
friendly district variable for each party and run two logistical regression models. I follow-up
with predicted probabilities for a more substantive interpretation of the effect women-friendly
districts have on female emergence and success.
Variables
Dependent
The dependent variable for the first model is female candidate emergence. This is a
dichotomous variables which measures the sex of a candidate running in the general election.
Males are coded ‘0’ and females are coded ‘1’. The second model’s dependent variable is
female winners. This variable is also dichotomous where a candidate who is female and won
the general election is coded ‘1’ and male winners are coded ‘0’.
Independent Variables
The first explanatory variable is the women-friendly district variable. I use six
district-level variables to compile the ‘women-friendly district’ variable: percent democratic
vote share, percent black, urbanization, average income, education (percent holding college
degree), and percent women in the workforce. The women-friendly district variable is a count
variable ranging from 0-7 with 0 being the least friendly and 7 being the most friendly.5 The
variable is a difference between the national average and district average. For example, if
the national urban average in core Democratic districts is 50% and an individual district is
over 50%, then the district is more women-friendly in regards to urbanization than other
districts. This process is done for each of the six variables and then summed to create the
5All districts outside of the South are considered more women-friendly and are also a part of the women-friendly district count. For example, all state districts not in the South will be at least 1 on the women-friendlydistrict variable.
84
0-7 count variable.
The variable is compiled separately for each party after determining core Democratic
and core Republican districts. There are six election periods for the time span of this data6.
A core district is one in which the party won at least five of the six election periods7 in
the study (2000-2010). A district is also considered a core district if the term length of
the legislature is four years8 (versus two) and the party won at least 2 of the three election
periods.
The next set of districts are the swing districts. Swing districts are districts that are
not core Democratic or core Republican districts. A district is a swing district if each party
won the seat anywhere from 2-4 times. For example, in a two year term legislature, a swing
district could be one where Democrats won the seat 2 times and a Republican won the seat
the other 4 elections.
Table 4.1 below shows the distribution of core Democratic, core Republican seat, and
swing seats. Core Democratic seats make up about 34% of the sample and core Republican
seats make up about 30% of the sample. The majority of seats are in swing districts and make
up about 36% of the sample. Since the data is candidate-level, the distribution represents
the total number of candidates running in the core districts, rather than the number of core
districts. This also explains the larger number of candidates in swing districts; theoretically
these seats are likely to be more competitive and less likely to have no opposition. Core
Democratic and Republican districts are likely to be less competitive– having a larger number
of unopposed races.
6The election periods are: 2000, 2002, 2004, 2006, 2008, 2010
7For states where terms are two years.
8The election periods for these states are generally 2000, 2004, 2008.
85
District Frequency Percent
Core Democratic 17,242 33.85%Core Republican 15,416 30.26%Swing District 18,286 35.89%
Table 4.1: State Legislative District Descriptives
Term limits is the second main independent variable and is coded ‘1’ for states with
term limits and ‘0’ for state without term limits. States will be coded as ‘0’ until their
implementation of term limits for that particular state begins. Therefore, there is not set
year for when the term limited states will be coded to ‘1’ and the value is dependent upon
implementation.
The last independent variable is an interaction between women-friendly districts and
term limits. I expect a positive interactive relationship between term limits and women-
friendly districts. Term limits increase the likelihood a female will emerge and win a seat
across each level of of the women-friendly district variable.
Controls
The first three controls are in both emergence and success models. This first control
in both models is incumbent. It is a dichotomous variable where ‘0’ is a non-incumbent and
‘1’ is an incumbent. The second control is open seat. This is a dichotomous variable coded ‘0’
for a seat where an incumbent is running and ‘1’ if the seat is vacant and has no incumbent
running. Lastly, professionalization is a measure is from the Squire index and comes from
Vanderbilt University’s State Politics and Judiciary dataset. The measure is a 0-1 interval
where 0 is the least professionalized and 1 is the most professionalized. The measure takes
into account whether the legislature is full-time versus part-time, how much legislators are
paid, staff size, etc. For more details on the measure see Squire 1992. As literature, and
86
my theory states, I expect highly professionalized states to have a negative impact on the
overall women-friendliness of districts.
The next set of control variables are also in both emergence and success model.
They are Elazar’s (1966) three classifications for a state’s political culture: Individualistic,
Moralistic and Traditionalistic. To measure these, I created two dummy variables. The first
is Individualistic and the second is Traditionalistic. Therefore, Moralistic is the baseline
category.
The last two controls are in the success models which capture to electoral environment
surrounding the race. The first is unopposed which is a dichotomous variable. This variable
is coded as ‘0’ for races in which at least two candidates run for office and ‘1’ where a
candidate runs in an uncontested seat. The second is margin of victory which is an interval
variable capturing the percent difference in vote share between the winning candidate and
the losing candidate(s).
Results
Table 4.2 provides the logistical regression analysis for female candidate emergence.
The table has a column for each party model: core and swing districts. The women-friendly
district variable is positive and signifiant for each model. This means that as a district
becomes more women-friendly (i.e.; more democratic, diverse, urban, educated, wealthier,
women working, non Southern state), a female candidate is more likely to emerge. The
political culture variables are also statistically significant and in the expected direction for
the core districts; political culture does not have a statistically significant impact on female
emergence in the swing districts. Further, the last set of state-level variables, term limits and
87
professionalization, do not have a statistically significant effect on female emergence.9 The
models indicate that women-friendly districts are present in state-legislative districts and
have a positive effect on the likelihood of a female candidate emerging. To further interpret
the effect women-friendly district have on female emergence, I provide predicted probabilities
below.
9In the simple models with no interaction, women friendly district variable is significant in all models butterm limits are on the cusp of significance.
88
Variable Core Republican Core Democrat Swing Republican Swing Democrat
Women Friendly District 0.148* (0.028) 0.157* (0.028) 0.063* (0.017) 0.080* (0.019)Individualistic -0.191* (0.075) -0.337* (0.075) -0.087 (0.077) -0.075 (0.077)Traditionalistic -0.221* (0.082) -0.170* (0.089) -0.160 (0.078) -0.123 (0.079)Term Limit 0.157 (0.166) -0.084 (0.150) -0.169 (0.156) -0.062 (0.137)Professionalism (Squire) -0.052 (0.257) -0.141 (0.232) -0.423 (0.276) -0.354 (0.276)Term Limit*WFD -0.052 (0.047) 0.052 (0.048) 0.049 (0.043) 0.020 (0.043)Incumbent -0.381* (0.065) 0.391* (0.061) 0.105* (0.046) 0.091* (0.046)Open Seat -0.202* (0.053) 0.352* (0.054) 0.068 (0.051) 0.062 (0.051)Constant -1.500* (0.118) -1.464* (0.124) -1.009* (0.093) -1.038* (0.090)
n =15020 n =16895 n=17919 n=17919
Robust standard errors clustered around legislative districtControls for election years included in the model but suppressed from the table
(See Appendix for full model with year dummies)
Table 4.2: Logistical Regression of Female Candidate Emergence
89
Figure 4.1. Core Democratic Female Emergence
.2.3
.4.5
Prob
abilit
y C
andi
date
is F
emal
e
0 2 4 6Democratic Women Friendly District
Female Democratic Emergence
Figure 4.1 shows the predicted probabilities for female candidate emergence in open
seat races in term limited states10. The women-friendly district variable ranges from 0-7 and
women are more likely to emerge in districts that are more women-friendly. In the least
women-friendly district, 0, a female democratic candidate is predicted to emerge around
24% of the time versus a women-friendly district, at 6, a female is likely to emerge about
46% of the time in these core Democratic districts.
10The political culture for all the predicted probabilities is set to Moralistic and other variables are setto their mean. Consistent with hypothesis 2, the Moralistic political culture is the most women-friendly.Females in Moralistic states are predicted to emerge about 3% more than in Individualist or Traditionalisticstates.
90
Figure 4.2. Core Republican Female Emergence
.15
.2.2
5.3
.35
.4Pr
obab
ility
Can
dida
te is
Fem
ale
0 2 4 6Republican Women Friendly Districts
Republican Female Emergence
Predicted probabilities for female Republican emergence are in Figure 4.2. As with
the Democratic women-friendly districts, the probability a female Republican will emerge
increases as the district becomes more women-friendly. Republican females are predicted to
emerge around 34% of the time in most women-friendly core Republican districts versus only
18% of the time in the least women-friendly districts.
Women-friendly districts are also present across the swing districts. As Figure 4.3
and 4.4 indicate, female emergence increases, for Democrats and Republicans, as a district
becomes more women-friendly. Women emerge around 30-33% of the time in districts that
have 4 to 6 ‘women-friendly’ features. On the low end of the women-friendly scale, females
emerge around 24-26% of the time. Overall, districts that have more women-friendly features
(more urban, diverse, higher education, higher income, more women in the workforce, and
higher democratic vote share) have more females emerge as candidates. The next set of
models and figures examine the relationship between women-friendly districts and female
91
success.
Figure 4.3. Swing Democratic Female Emergence
.2.2
5.3
.35
.4Pr
obab
ility
Can
dida
te is
Fem
ale
0 2 4 6Democratic Women Friendly Districts
Swing Districts: Democratic
Figure 4.4. Swing Republican Female Emergence
.2.2
5.3
.35
Prob
abilit
y C
andi
date
is F
emal
e
0 2 4 6Republican Women Friendly Districts
Swing Districts: Republican
92
Variable Core Republican Core Democrat Swing Republican Swing Democrat
Women Friendly District 0.173* (0.046) 0.140* (0.039) 0.043 (0.027) 0.034 (0.029)Individualistic 0.001 (0.106) -0.214* (0.089) -0.145 (0.106) -0.140 (0.107)Traditionalistic -0.137 (0.132) -0.253* (0.113) -0.310* (0.123) -0.315* (0.124)Term Limit 0.033 (0.251) -0.058 (0.200) -0.227 (0.222) -0.193 (0.192)Professionalism (Squire) 0.496 (0.354) -0.211 (0.264) -0.365 (0.384) -0.272 (0.380)Term Limit*WFD -0.018 (0.069) 0.022 (0.061) 0.064 (0.058) 0.060 (0.060)Incumbent 6.358* (0.233) 6.541* (0.190) 3.320* (0.133) 3.319* (0.133)Open Seat 3.097* (0.186) 3.349* (0.145) 1.605* (0.116) 1.604* (0.116)Unopposed 4.163* (0.650) 3.724* (0.338) 2.935* (0.487) 2.936* (0.486)Margin of Victory 0.023* (0.003) 0.012* (0.002) 0.016* (0.002) 0.016* (0.002)Constant -6.272* (0.341) -5.458* (0.233) -3.438* (0.221) -3.388* (0.212)
n =7599 n =9463 n=5842 n=5842
Robust standard errors clustered around legislative districtControls for election years included in the model but suppressed from the table
(See Appendix for full model with year dummies)
Table 4.3: Logistical Regression of Female Success
93
Table 4.3 is the results for the likelihood a female will win. Consistent with the
emergence model and hypotheses, females are more likely to win as the women-friendliness
of a district increases. In the core Democratic and Republican models, the women-friendly
district variable is positive and signifiant. For the swing districts, the women-friendly district
variable is in the expected direction and on the cusp of significance11. Also of note, the
interaction between term limits and women-friendly districts is in the expected direction for
each of the models but not statistically significant at conventional levels. The interactive
effect between term limits and women-friendly districts is on the cusp of significance in the
swing district models implying term limits have a greater, positive effect for female winners
in the swing districts. Further, the political culture variables are in the expected direction12;
moralistic is the base line and traditionalistic and individualistic have a negative effect on
the likelihood a female will win a seat.
As with the emergence models, I run predicted probabilities to substantively interpret
the effect women-friendly districts, and term limits have on female candidate success. The
predicted probabilities are for term limited, Moralistic states with all other variables set to
their mean.13 Figures 4.5 and 4.6 show how an increase in a district’s women-friendliness also
increases the probability of female candidate success. Figure 4.5 shows Democratic females
win about 20% of the time in districts that are least women-friendly versus about 38-41% of
the time in the most women-friendly districts.
11In the simple model with no interaction between women-friendly district and term limits, the women-friendly district variable is significant and in the expected direction across all model
12with the exception of individualistic in the Republican model.
13Female candidates are more successful in Moralistic states. They are predicted to win about 3% more ateach level of the women-friendly variable in comparison to the Individualists and Traditionalistic states.
94
Figure 4.5. Core Democratic Female Success
.1.2
.3.4
.5Pr
obab
ility
a Fe
mal
e W
ins
0 2 4 6Democratic Women-Friendly District
Democratic Female Success
Figure 4.6. Core Republican Female Success
.1.1
5.2
.25
.3Pr
obab
ility
a Fe
mal
e W
ins
0 2 4 6Republican Women-Friendly District
Republican Female Success
Figure 4.6 above shows female Republican success increases as a district becomes
more women-friendly. At the lower end of the women-friendly scale, the probability of a
female Republican winning a seat is low–around 11%. The probability of success is nearly
95
tripled in the most women-friendly districts. The probability of female success rates are
around 25-29% in districts with 6-7 women-friendly characteristics.14
Conclusions and Implications
The chapter finds Palmer and Simon’s (2006) women-friendly districts are also present
at the state-level. The same demographic characteristics are present in each of the models;
women run and win more seats in areas that are more diverse, more urban, more democratic,
more educated, have higher levels of income, more women in the workforce and are outside
the South. Further, the political culture of a state matters. Women are more likely to run and
win in Moralistic political culture states than they are in Individualistic or Traditionalistic
states. Contrary to my hypothesis, term limits do not seem to have a positive interactive
effect with women-friendly districts to increase female emergence and success. The borderline
significance findings may be due to the majority of females running and winning in mid-range
women-friendly districts (i.e. with 3-5 features). With fewer women running and winning in
the least and most friendly districts, this may be washing out some effects term limits might
have on women-friendliness.
Results from the chapter show that even in the most women-friendly districts, female
candidates win less than 50% of the races in these districts. The likelihood of emerging
and winning increases as women-friendliness increases, but even the most women-friendly
districts do not guarantee a female winner; Especially if a female does not run. Overall, the
results for descriptive representation at the state-level are promising since women are more
likely to run and win in the women-friendly districts. Yet, what about the districts that are
14I do not include predicted probabilities for female success rates in swing districts since women-friendlydistricts are not statistically significant in the model. Success rates for female Democrats and Republicansin the swing districts range from 20-25% but the confidence intervals cross at all levels of the women-friendlydistrict variable
96
not women-friendly? The implications for descriptive representation of women in districts
that do not produce female candidates or winners is something to consider.
In the future, inclusion of states with multi-member districts is a great opportunity to
further examine women-friendly districts in the states. Further, the current women-friendly
district measure weighs all points of the women-friendly district variables the same. Women
may be more successful in highly urbanized areas versus highly diverse areas. For example, in
some districts, urbanization might matter for female emergence more than diversity within
the district. Perhaps, this is what is happening in the swing districts where the women-
friendly district variable was not statistically significant. The seats in the swing districts
are much more competitive and perhaps a women-friendly district variable looks different
in these districts; i.e being in the South might matter more and may hurt women more in
swing districts than the core districts.
97
[1] Bedolla, Lisa, Katherine Tate and Janelle Wong. 2014. “Inedible Effects: The Impact of Women ofColor in the U.S. Congress.” in Women and Elective Office: Past, Present, and Future” eds. Thomas,Sue and Clyde Wilcox. New York, NY: Oxford University Press.
[2] Burrell, Barbara C. 1994. A Woman’s Place Is in the House: Campaigning for Congress in the FeministEra. Ann Arbor: University of Michigan Press.
[3] Campbell, Angus, et al. 1960. The American Voter. Chicago: University of Chicago Press.
[4] Carroll, Susan and Krista Jenkins. 2001. “Do Term Limits Help Women Get Elected?” Social ScienceQuarterly. 82/1.
[5] Census Bureau. 2012.
[6] Center for American Women and Politics. 2012. “Women in State Legislatures? Fact Sheet”:www.cawp.rutgers.edu.
[7] Darcy, R, Susan Welch and Janet Clark. 1994. Women, Elections, and Representation. 2nd Edition.Lincoln: University of Nebraska Press.
[8] Darcy, R. and Sarah Slavin Schramm. 1977. “When Women Run against Men,? Public OpinionQuarterly 41: 1-12.
[9] Dolan, Kathleen. 2004. Voting for Women: How the Public Evaluates Women Candidates. Boulder,CO: Westview.
[10] Dolan, Kathleen.2009. “The Impact of Gender Stereotyped Evaluations on Supoport for WomenCandidates.” Political Behavior. 32:69-88.
[11] Elazar, Daniel. 1966. American Federalism: A View from the States. New York: Crowell.
[12] Fenno, Richard. 1978. Home Style: House Members in Their Districts. Boston: Little Brown.
[13] Flanigan, William and Nancy Zingale. 1994. Political Behavior of the American Electorate, 8th edition.Washington, DC: CQ Press.
[14] Flanigan, William and Nancy Zingale. 2002. Political Behavior of the American Electorate, 10th editionWashington, DC: CQ Press.
[15] Fox, Richard L. 2000. “Gender and Congressional Elections.” In Gender and American Politics: Women,Men, and the Political Process, eds. Sue Tolleson Rinehart and Jyl J. Josephson. Armonk, NY: M.E.Sharpe.
[16] Fox, Richard and Jennifer Lawless. 2004. “Entering the Arena? Gender and the Decision to Run forOffice.” American Journal of Political Science. 48:264-280.
[17] Frank, Thomas. 2004. What’s the Matter with Kansas? New York: Metropolitan Books.
[18] Fulton, Sarah et, al. 2006. “The Sense of a Women: Gender Ambition, and The Decision to Run forCongress”. Political Research Quarterly. 59: 235-248.
[19] Fulton, Sarah. 2012. “Running Backwards and in High Heels: The Gendered Quality Gap andIncumbent Electoral Success.” Political Research Quarterly. 20:1-12.
[20] Gelman, Andrew. 2009. Red State, Blue State, Rich State, Poor State. Princeton: Princeton UniversityPress.
98
Chapter Bibliography
[21] Koch, Jeffrey. 2000. “Do Citizens Apply Gender Stereotypes to Infer Candidates’ Ideological Orienta-tions?” Journal of Politics 62: 414-29.
[22] Koch, Jeffrey. 2002. “Gender Stereotypes and Citizens’ Impressions of House Candidates’ IdeologicalOrientations.” American Journal of Political Science. 46(2), 453-462.
[23] Lau, Richard and David Redlawsk. 2001. “Advantages and Disadvantages of Cognitive Heuristics inPolitical Decision Making”. American Journal of Political Science 45: 951-71.
[24] Matland, R.E., Studlar, D.T., 2004.“Determinants of Legislative Turnover: a Cross-National Analysis.”British Journal of Political Science. 34: 87-108.
[25] National Conference of State Legislatures. 2011 “Legislative Term Limits Overview?”http://www.ncsl.org/.
[26] Norrander, Barbara and Clyde Wilcox. 2014. “Trend in the Geography of Women in the U.S. StateLegislatures.” in Women and Elective Office: Past, Present, and Future” eds. Thomas, Sue and ClydeWilcox. New York, NY: Oxford University Press.
[27] Palmer, Barbara and Dennis Simon. 2005. “When Women Run Against Women: The Hidden Influenceof Female Incumbents in Elections to the U.S. House of Representatives, 1956-2002.” Gender andPolitics. 1:39-63.
[28] Palmer, Barbara and Dennis Simon. 2006. Breaking the Political Glass Ceiling: Women and Congres-sional Elections. New York: Routlege.
[29] Rahn, Wendy. 1993. “The Role of Partisan Stereotypes in Information Processing about PoliticalCandidates”. American Journal of Political Science 37: 472-96.
[30] Rule, Wilma. 1981. “Why Women Don’t Run: The Critical and Contextual Factors in Women’sLegislative Recruitment”. Western Political Quarterly 34:60-77.
[31] Sanbonmatsu, Kira.2006. “Where women run.” Ann Arbor: University of Michigan Press.
[32] Sanbonmatsu, Kira and Kathleen Dolan. 2009. “Do Gender Stereotypes Transcend Party?” PoliticalResearch Quarterly 62: 485-494.
[33] Schreiber, Ronnie. 2014. “Conservative Women Run for Office.” in Women and Elective Office: Past,Present, and Future” eds. Thomas, Sue and Clyde Wilcox. New York, NY: Oxford University Press.
[34] Swers, Michele. 1998. “Are Congresswomen More Likely to Vote for Women’s Issue Bills than TheirMale Colleagues?” Legislative Studies Quarterly 23(3): 435-448.
[35] Swers, Michele. 2002. The Difference Women Make. Chicago: University of Chicago Press.
[36] Thomas, Sue and Clyde Wilcox. 2014. Women and Elective Office: Past, Present, and Future” NewYork, NY: Oxford University Press.
[37] Werner, Emmy. 1966. “Women in Congress: 1917-1964.” Western Political Quarterly 19: 16-30.
99
Appendix
Table 1 below is a difference of means test for female winners. The difference of means
test is run for both parties to determine if the set of variables that make up the women-
friendly district variable are significantly different with each party. In other words, the
difference of means test is a robustness check to confirm women-friendly districts are different
from a party-friendly district. Table 1 confirms that across both parties the differences are
significant. The only variable not significant is the percent black in a district (diversity) for
Republicans.
100
Table A.1 : Party-Friendly versus Women-FriendlyVariable Male Republican Female Republican Male Democrat Female Democrat
Party 25.63 27.34* 75.92 74.48*Diversity 4.48 4.33 14.03 16.46*Urban 0.64 0.69* 0.75 0.81*Income 45278.53 46579.92* 38392.49 40126.96*Women in Workforce 59.77 60.32* 57.58 59.76*Education 16.3 17.63* 13.79 16.1*South 0.31 0.25* 0.32 0.24*
101
Table 2 shows the results for the female emergence logit model using district demo-
graphics. The model controls for election year (not reported in table) and clusters around
the state legislative district. Table 2 supports the results in the difference of means test in
Table 1 above and finds each variable is in the expected direction. State districts that are
more diverse, more urban, and have higher levels of education will be more friendly towards
women. Each of these variables are in the expected direction15.
Further, districts with a higher levels of Democratic vote share are more women-
friendly. The party variable captures this and a positive direction indicates a higher demo-
cratic vote share. MMDs and term limits also increase women-friendliness and professional-
ization will decrease women-friendliness. Each of these variables are in the expected direction
and statistically significant. Lastly, the political culture variables are also in the expected
direction and statistically significant. Females are less likely to emerge in individualistic and
traditionalistic political cultures, in comparison to the moralistic political culture, which is
the baseline in the model.
15Income and education were highly correlated at about .60. I then ran a model with education and womenin the workforce which were also highly correlated at .51. Therefore, in the final models I use education anddrop income and women in the workforce to reduce multicollinearity
102
Table A.2: Likelihood Candidate is a Female
Variable Coefficient Standard Error
Party 0.004* 0.001Urbanization 0.002* 0.001Black 0.009* 0.001Individualistic -0.296* 0.044Traditionalistic -0.372* 0.050Education 0.022* 0.003Term Limits 0.083* 0.038District Type 0.166* 0.046Professionalization -0.250* 0.147Constant -1.657* 0.069n =46324
The next analysis is on the likelihood of a female winner is in Table 3 below. As
with the emergence model, females are winning seats that are more urban and educated.
Further, females are winning seats in more diverse districts and also districts with a higher
democratic vote share. Each of these variables are in the expected positive direction and are
statistically significant at the traditional .05 level.
Further, women are more likely to win seats in MMDs and less likely to win seats
as the professionalization of the legislature increases. The model also shows women are less
likely to win seats in term limited states. Taken in consideration with the emergence model,
the analyses show women are more likely to run in these term limited states but are less
likely to win the seat.
103
Table A.3: Likelihood Winner is a Female
Variable Coefficient Standard Error
Party 0.011* 0.001Urbanization 0.002* 0.001Black 0.013* 0.002Individualistic -0.133* 0.051Traditionalistic -0.080 0.058Education 0.020* 0.003Term Limits -0.224* 0.045District Type 0.114* 0.054Professionalization -0.550* 0.174Constant -1.968* 0.102n =26253
The next set of tables below are the tables presented in the paper with the election
years included.
Table A.4: Core Democrat: Female Emergence
Variable Coefficient Standard Error
dem women friendly avg .1565631 .0282458individualistic pc -.3365629 .0748279traditionalistic pc -.170424 .0890139term limit -.0836265 .1504956termlimit wfdem .0516483 .0480877legprof squire -.1411437 .2323934incumbent .3911542 .0605985open seat .3518542 .0538863elect 2000 -.0912832 .0557684elect 2002 -.1284345 .0507623elect 2004 -.0258399 .0485041elect 2006 -.0543958 .0451579elect 2008 -.0474389 .0394987cons -1.463704 .1237331
104
Table A.5: Swing Democrat: Female Emergence
Variable Coefficient Standard Error
dem women friendly avg .0797918 .0190978individualistic pc -.0745123 .077107traditionalistic pc -.1226604 .0787729term limit -.0623158 .137266termlimit wfdem .0201905 .0431716legprof squire -.3544942 .2757757incumbent .091498 .0461193open seat .0620944 .0511741elect 2000 -.1714078 .0588112elect 2002 -.2322726 .0553721elect 2004 -.1707884 .0533979elect 2006 -.0349468 .0462753elect 2008 -.0363515 .04327cons -1.038344 .0899591
Table A.6: Core Republican: Female Emergence
Variable Coefficient Standard Error
rep women friendly avg .147747 .0276422individualistic pc -.1913911 .075356traditionalistic pc -.2212877 .0816072term limit .1568898 .1659367termlimit wfrep -.0342328 .0474659legprof squire -.0520752 .257339incumbent -.3812159 .0654788open seat -.2016688 .0533414elect 2000 .0468717 .0663321elect 2002 .0179355 .0624462elect 2004 -.0479129 .0606049elect 2006 -.0027185 .0564105elect 2008 -.0220036 .0507171cons -1.349211 .126791
105
Table A.7: Swing Republican: Female Emergence
Variable Coefficient Standard Error
rep women friendly avg .0638575 .0169588individualistic pc -.0872312 .0771295traditionalistic pc -.1600073 .0781087term limit -.1691452 .1557498termlimit wfrep .049498 .043242legprof squire -.4233907 .2758529incumbent .1051049 .0460237open seat .0678409 .0511774elect 2000 -.1611315 .0582281elect 2002 -.2277087 .0550821elect 2004 -.172 .0534261elect 2006 -.0342929 .0461984elect 2008 -.0350461 .0432426cons -1.008653 .0925727
Table A.8: Core Republican: Female Success
Variable Coefficient Standard Error
female cand -.5002398 .0701227rep women friendly avg .0273552 .0095952individualistic pc .0148918 .0166657traditionalistic pc .049343 .0221703term limit -.0095145 .0502801termlimit wfrep -.0034838 .0132543legprof squire -.0010355 .0601022incumbent 6.215262 .1654569open seat 3.059142 .0842767unopposed 5.475893 .678396margin victory .0008582 .0005891elect 2000 .0533524 .0255042elect 2002 -.0147778 .0268848elect 2004 -.002193 .0233656elect 2006 .0179845 .0255359elect 2008 .0235038 .0240681cons -3.138546 .101559
106
Table A.9: Core Democrat: Female Success
Variable Coefficient Standard Error
female cand .3450051 .0663762dem women friendly avg -.0405944 .0094148individualistic pc .0317028 .0231128traditionalistic pc -.0453398 .0336459term limit -.079386 .0423397termlimit wfdem .0185845 .0139668legprof squire .0088317 .0645849incumbent 6.255734 .1457224open seat 3.074011 .0757197unopposed 4.058447 .3333078margin victory .0037032 .0006704elect 2000 .0039934 .0265013elect 2002 -.0445716 .0249698elect 2004 -.0038756 .0240392elect 2006 -.1049973 .0285525elect 2008 -.0538214 .0272173cons -3.232047 .0834231
107
Table A.10: Swing Republican: Female Success
Variable Coefficient Standard Error
female cand .0573801 .0567672rep women friendly avg -.0105447 .0044935individualistic pc -.0153122 .0145414traditionalistic pc -.0549274 .030834term limit -.1849904 .0438697termlimit wfrep .0330612 .0110045legprof squire .0991332 .0541343incumbent 3.199984 .0941815open seat 1.519908 .0498495unopposed 4.009929 .4750092margin victory .0013476 .0004078elect 2000 -.0219494 .0199121elect 2002 -.0250504 .0201461elect 2004 .02788 .0186179elect 2006 -.0932389 .026431elect 2008 .0291471 .0182408cons -1.588223 .0564312
108
Table A.11: Swing Democrat: Female Success
Variable Coefficient Standard Error
female cand .0582481 .0567276dem women friendly avg -.0188258 .0050216individualistic pc -.0181286 .0146655traditionalistic pc -.0667596 .032422term limit -.1850051 .0414143termlimit wfdem .0377782 .0130036legprof squire .1139771 .0543251incumbent 3.200017 .0941845open seat 1.521311 .0498454unopposed 4.013425 .4744298margin victory .0014195 .0004185elect 2000 -.0208039 .0199954elect 2002 -.024844 .0201764elect 2004 .0280286 .0186432elect 2006 -.0931715 .0263905elect 2008 .0292412 .0182321cons -1.569373 .0560399
109
CHAPTER 5
CONCLUSION
The puzzle which inspired this dissertation was two-fold. This first puzzle is that
women make up an equal portion of the population but are underrepresented in government.
I was particularly interested in the underrepresentation of women in state legislatures. The
number of women has increased overtime since the 1970s but became stagnant in the late
1990s. To help understand why there is stagnation but also variation across the states, I
looked to institutional features—mainly term limits. Further, I wanted to examine literature
that may explain why women run and win office (rather than literature that keeps women
out of office). Term limits were this unique institutional change that happened in the 1990s
(around the time of stagnation) to help try and resolve the puzzle of why female descriptive
representation is so low.
The second puzzle involves term limits and their effect on legislators. Research finds
women are successful in open seat races. Since term limits create a large number of open
seats, scholars theorized that term limits would have a positive effect on female descriptive
representation. Yet as discussed throughout the dissertation, initial studies on term limits
found that there was a negative effect on female descriptive representation. Women who
were in office left due to term limits and no women replaced them. Exploring the possible
reasons why the conventional wisdom and initial research were at odds was an important
motivator for the dissertation.
In general, in the dissertation I find term limits have a positive effect on female
candidates. Term limits create more open seats and allow new candidates to seek office. The
emergence chapter finds females are much more likely to emerge as candidates after term
110
limits are implemented. This is consistent with congressional literature which finds women
are more likely to run in open seat races since their chances of winning are higher when they
are not a challenger to an incumbent.
The female success chapter finds term limits positively effect female descriptive repre-
sentation for Republican females and negatively for Democratic females. Female Republicans
are about 2% more likely to win in term limited legislatures. Female Democrats are about
5% less likely to win in term limited states. Overall though, female Democrats, despite term
limits, are very successful and are still likely to win their race over 50% of the time. Female
Republicans on the other hand are helped by term limits but their success rates are not as
high as female Democrats. This implies that female descriptive representation has the poten-
tial to increase overtime as more female Republicans gain office. Lastly, the women-friendly
district chapter is a first look at how state legislative districts (in the lower house) can have
a certain set of demographics that make them more likely to elect female candidates. I find
support for women-friendly districts for both Republican and Democratic parties. Females
are more likely to emerge and win seats as districts become more women-friendly. The key
demographics making up a women-friendly district are: high levels of income, education,
women in the workforce, more democratic, more urban, and diverse. The analysis in the
chapter does not find and interactive effect between term limits and women-friendly districts.
In other words, a district does not become more women-friendly when in term limited states.
Implications
Term limits affect female descriptive representation in parties differently. Overall, I
find term limits increase the likelihood a female candidate (both Democrat and Republican)
will emerge. Yet, I find term limits increase female Republican success but decreases female
111
Democratic success. The implications of such for Democrats, and female Democrats are
important to consider. A closer look at party-friendly and district-friendly seats post term
limits may provide insight. For example, if term limits do cause a “partial reversal in
professionalization” (Meinke and Hasecke 2003), favoring Republicans, are core Democratic
districts becoming swing districts, or Republican districts? If term limits do cause safe
Democratic districts to become less safe, this hurts the Democratic party, which in turn
hurts female Democrats. Overall, if term limits are hurting the Democratic party, and
females, this may decrease female descriptive representation overtime given the majority of
successful female candidates are Democrats.
On the other hand, something to consider is if Republicans, specifically female Re-
publicans, are taking advantage of term limits. If Democratic females are emerging, but
winning less in term limited states understanding the reason why they are losing is critical.
Are Democrats recruiting candidates they do not think can win or are Republican candi-
dates post term limits better suited (qualified, financially or otherwise) to win? Oftentimes,
researchers discuss the possibility of sacrificial lambs (See Canon 1992) Essentially, it is a
candidate that will likely not win the election but are run anyway. Generally sacrificial lambs
run against incumbents, rather than in open seats created by term limits. Term limits may
be increasing the number of sacrificial lambs in a race.
One last implication, and not directly pursued in this work, is whether or not term
limits positively effect substantive representation of women. I conclude term limits have the
potential to increase descriptive representation. Specifically since female Republicans are
more successful at winning post term limits. Policy from legislatures that are comprised of a
majority of Republicans will likely to substantially different than legislatures with equal num-
112
bers of Democrats/Republicans. Further, policies that favor women tend to come from the
Democratic party since the party typically supports the expansion of social welfare programs
(amongst other policies)— especially those which support women and children. Therefore,
an interesting implication to consider is the type of policies coming from legislatures with
more female Republicans in them.
Future Research
In terms of future research, there is still a vast amount of work that can be done on
females in the states. I intend to complete a few research projects that will build upon the
current chapters. In terms of candidate emergence, future research will examine more of
the parties within the states and see how they have adapted strategies post term limits. I
speculate in the emergence chapter that parties (and female recruitment agencies) will adapt
new recruitment strategies in term limited states. In order to test my theory, I would like to
conduct some surveys of state legislators and party organizations in the term limited states.
The surveys will try and understand if the new legislators have different motivations for
running and specifically whether term limits were an incentive to run for office. Further, I
am interested in knowing whether or not some state parties seek out female candidates for
office post term limits.
In terms of female descriptive representation, future research will look into what
women do once termed out of office. After term limits, are these women going to the state
senate or Congress? Or, in states with consecutive term limits are women going back to
the lower house after the designated time away? These questions are important to look into
because depending on where women end up after term limits has the potential to increase
(or decrease) female descriptive representation at other levels of government. Further, where
113
women go after term limits may also add insight into the original motivations to run for
offIce. For example, if term limits are desirable to women because it is a temporary political
position, women should not seek office after being termed out.
Lastly, work on women-friendly districts in the states is fairly new. Furthering the
nuances of the women-friendly district variable, as previously discussed, will be a valuable
addition to the field. Linking the state women-friendly district variable to the congressional
women-friendly districts is one future project I plan to investigate. Exploring the similarities
and differences between the women-friendly districts, especially the swing districts, may be
useful for increasing female descriptive representation. For example, if there a areas where
districts are women-friendly in the lower house, upper house, and Congress these may be good
areas for women’s groups or parties to actively recruit women. This is especially important
if the district is is a state with term limits. Overall, an active eligibility pool of females
who will run and win races in term limited states is important for increasing descriptive
representation.
114