3.1 Large-N Tests
3.1.4 Methods and Estimation Techniques
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
35 developed by the Center for Systemic Peace (CSP). The Polity Score examines the
concurrent qualities of democratic and autocratic authority in government institutions.
Lastly, existing literature argues that the overall level of economic development in the country has positive effects on women’s political opportunities (Duflo 2012; Matland 1998; Stockemer 2015). In short, their argument is that development leads to a
weakening of traditional values, greater urbanization, greater participation of the labor force and education for women and changes in attitude in the perceptions of appropriate roles for women. To control the level of economic development of the countries, I control for GDP per capita, which basically is the gross domestic product divided by the midyear population. This variable is operationalized as the logged value of the GDP per capita for the year prior to the election year. The data is from the World Bank’s World
Development Indicators.
3.1.4 Methods and Estimation Techniques
In this research, I employ “mixed methods” which combines quantitative and qualitative approaches into a new methodology. As a method, it focuses on collecting, analyzing and combining quantitative and qualitative data into a single study (Johnson and Onwuegbuzie 2004). The reason why I employed this methodology is that the utilization of qualitative and quantitative approaches gives a better comprehension of the research problem than any approach by itself. In addition, mixed methods research provides an opportunity to compensate for the inherent weaknesses and unavoidable biases of the methods, as well as to take advantage of the inherent strengths of the methods (Greene 2007; Harwell 2011; Johnson and Turner 2003).
For the quantitative approach, I use two estimation techniques. Firstly, I employ Ordinary Least Squares regression (OLS), often called lineal regression, in order to estimate women’s representation in the lower chambers in Latin America. I have chosen this technique because I usetime-series–cross-section (TSCS) data (i.e., my dataset includes multiple observations of the same country over time). These observations may
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
36 not be absolutely independent and, therefore, the analysis has risks of serial correlation and heteroskedasticity (Beck 2001). Hence, to take care of these threats, I employ robust standard errors grouped by country.
‧
Table 2: List of Variables.Variables Labels Measurement Sources
Dependent Variable Women’s Representation in Lower Chambers
Percentage of seats held by women in the unicameral national
parliaments or lower chambers of the national legislatures.
For this variable, the data is from two sources: The World Bank’s Gender Statistics Data Base and the Statistical Archive of Women in National Parliaments from The
Inter-Parliamentary Union (IPU). For Argentina's 1989, 1991, 1993, 1995 Elections, the data is from a case study elaborated by Elisa María Carrio published on the International Institute for Democracy and Electoral Assistance in 2002. For Bolivia's 1993 Election, the data is from a case study about Bolivia's Legislative Power developed by Naciones Unidas para el Desarrollo (PNUD, United Nations Development Programme).
In the case of Chile's 1989, 1993, 1997, 2001 Elections, the data was collected from a candidacy report conducted by Candidaturas Chile (www.candidaturaschile.cl); and for Guatemala's 1994 Election, the data is from a case study elaborated by ATENEA, a project lead by UN Women, UNDP, and International IDEA, which seeks to measure the real exercise of women's political rights in Latin America and the Caribbean. For Mexico's 1991 Election, the data is from the Instituto Nacional Electoral (INE, National Electoral Institute).
Independent Variables
Gender Quota Laws
A dummy variable coded as “0”
when there are no regulations of quota for women and “1”
otherwise
The data is from the Gender Quota Database provided by the International Institute for Democracy and Electoral Assistance (International IDEA) and from the Gender Equality
Observatory for Latin America and the Caribbean developed by the United Nations Economic Commission for Latin America and the Caribbean (UNECLAC).
Actual Gender Quota (%) The percentage of seats required by the Gender Quota Laws.
The United Nations Economic Commission for Latin America and the Caribbean (ECLAC). Available on internet at:
https://oig.cepal.org/en/laws/3/country Degree of Egalitarianism
of Political Institutions An interval from low to high (0-1).
To operationalize this variable, I use the Egalitarian Component Index developed by the Varieties of Democracy (V-Dem) Institute.
‧
The women’s percentage in secondary education age that is matriculated in secondary school.
The World Bank’s Gender Statistics Data Base.
Female Labor Force
Participation The percentage of the labor force
comprised of women. The World Bank’s Gender Statistics Data Base.
Overall Education Level
of the Countries An interval from low to high (0-1).
This variable is operationalized by the Human Development Index (HDI), developed by the United Nations Development Programme.
President’s Ideology
A dummy variable coded “1” for a leftist government and “0”
otherwise.
The Database of Political Institutions developed by the Inter-American Development Bank.
Log of Average District Magnitude
Logged average district magnitude size.
The data for this variable is from the dataset developed by Laura Wills Otero. Available online at
https://uniandes.academia.edu/LauraWillsOtero.
Closed-Party List System
A dummy variable coded as “1”
when the country has a closed-list system, and “0” otherwise.
The Gender Equality Observatory for Latin America and the Caribbean developed by the United Nations Economic Commission for Latin America and the Caribbean (UNECLAC).
Degree of Religiosity
An interval variable that is the average of the available databased on the ordinal answer (0-4) for the AmericasBarometer survey question how often do you attend the religious services.
The data for this variable comes from survey data available from 2004 to 2014 from the AmericasBarometer developed by the Latin American Public Opinion Project (LAPOP).
Polity Score
The “Polity Score” evaluates the type of regime in a country with a range from -10 (hereditary monarchy) to +10 (consolidated democracy). This variable is lagged by one year.
The Polity IV dataset developed by the Center for Systemic Peace (CSP).
Logged GDP per capita (current US$)
GDP per capita is gross domestic product divided by midyear population. This variable is logged and lagged by one year.
The World Bank's World Development Indicators.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
39 Second, I use pooled Time-Series Cross-Sectional (TSCS) analyses along with panel-corrected standard errors. This technique is especially useful because my data is available only for a short time (two decades or three) for the 18 Latin American countries. Also, TSCS analyses gives me the possibility of capturing variation of time and space simultaneously (Fortin-Rittberger 2014).
In this research, I employ two different series of models. First, I use one set of models that considers the variables’ independent effect. Regarding the second series of models, I incorporate all the variables, including the interaction term of gender quota laws and the egalitarianism of political institutions. Because there are two time-invariant variables in my models, which are closed-list system and religiosity, it will be inappropriate to estimate country fixed-effect models.