Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Scholars studying incumbency advantages have focused on identifying and measuring variables that explain and enhance the good chances that incumbents enjoy (Hogan, 2000,2004; Erikson 1971; Mayhew 1974; King 1991;). However, until now, little research has considered the incumbent advantage outside the United States case. Moreover, almost no efforts have been made to investigate the effect that different electoral rules have on the matter. Here, I move beyond the American case and try to explain incumbency advantages in legislative elections under PR closed-list and open-list systems. Building on the premise that in preferential systems (PR-OL) electoral competition tends to be candidate-centric and in PR-CL it tends to be party-centric, I show that incumbency advantages are stronger in cases where voters are entitled to rank their preferences. Colombia offers an interesting scenario to test the previous premise given that in this context parties can choose whether to present open or closed candidate lists. Using electoral data from the 2006 legislative election, I compare the incumbents' advantage in open and closed lists, while controlling for other confounding factors such as length term, term limits and constituency services.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: While previous research using cross-national studies has concluded that closed list electoral systems are more favorable to women than those that allow candidates’ individual vote totals to determine who wins legislative seats, an analysis of approximately 1700 districts from Brazil’s 2000 municipal elections shows conclusively that female candidates can compete successfully within PR systems using open lists provided that they can get on the ballot. The electoral studies literature mostly has emphasized four means by which democratic societies can make their electoral systems more favorable to the election of women: adopting PR rather than single member district electoral systems, increasing the number of seats per district, implementing gender-based quotas for party nominations, and—in the case of systems that already have gender quota legislation—strengthening the enforcement of nomination quotas. The present study uncovers two additional mechanisms that increase female candidacy rates: the presence of additional political parties and raising the number of candidates nominated per party. Controlling for district-level variations in socioeconomic development, the electoral strength of Brazil’s dominant leftist party, district magnitude, and party magnitude, OLS regression models indicate that that an increase in either the number of parties contesting an election or in the number of candidates nominated by each party in a district leads to improved ballot access for women even in settings where gender quotas are not being rigorously met. Boosting the overall number of parties, however, is ultimately less beneficial to women than raising the number of candidacies per party (and requiring that the parties actually fill their nominations). While both circumstances lead to higher female candidacy rates, increases in the number of parties tend to result in only the strongest candidates from each party being elected, a situation that generally works against the ability of women candidates to win seats in systems that use OLPR.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: This paper examines how voters attach responsibility to individual legislative candidates in open list PR systems in response to contextual changes in the economy.

Publication Type: Paper/Poster Proposal
Abstract: The National Cancer Institute’s (NCI) Health Information National Trends Survey (HINTS) gathers information from a nationally representative sample (with over-samples of ethnic-minority populations) about cancer communication practices, information preferences, risk behaviors, attitudes, and cancer knowledge. HINTS data may guide NCI’s information planning efforts and development of effective health communication strategies. It is paramount that valid and reliable measurement of cancer knowledge be demonstrated. Our analysis points to difficulties in current measurement of knowledge of general cancer risk in the HINTS. We contrast estimates of knowledge of specific risk factors obtained using open-listing and targeted questioning for population subgroups defined by race, language of interview, sex, age, education, and cancer experience. For example, all respondents were asked: “Can you think of anything people can do to reduce their chances of getting cancer? Anything else?” 64% of non-Hispanic Whites (NHW) mentioned not smoking or quitting smoking, compared with 50% of non-Hispanic Blacks (NHB) and 49% of Hispanics. A random-half of respondents were also asked: “I’m going to read you some things that may affect a person’s chances of getting cancer. Do you think that smoking increases a person’s chances of getting cancer a lot, a little or not at all or do you have no opinion? 94% of NHW responded a lot or a little, compared with 91% of NHB and 94% of Hispanics. Additionally, estimates deriving from targeted questioning may be affected by priming. Among NHW who did not mention smoking in response to the open-listing question, 86% responded a lot or a little to the targeted question, compared with 98% of those who did mention smoking. For NHB, the contrast is 86% versus 96%; for Hispanics the contrast is 92% versus 95%. We present additional examples and discuss implications for development of items to measure knowledge of cancer risks.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: This research investigates the prevalence of corruption among Brazilian legislators by building a model of the incentives and opportunities that make them susceptible to such illicit practices. I argue that the risk of corruption involvement by an elected legislator is primarily a function of the legislator’s career trajectory, geographic patterns of electoral campaigns, contagious effects of corruption, and regional conditions that make deal-making with local bosses an effective electoral strategy. I use data for the deputies of the 52nd Congress to test my argument.