The sample for the study was composed of 250 villages–each with an average population of roughly 1,000 people–selected from ten districts spanning northern, northeastern, eastern, central, and western Afghanistan (southern areas were excluded due to security concerns). Half of the villages was randomly assigned to hold district elections and the other half to hold at-large elections. Under district elections, the village was split into geographically-defined districts and each villager could only vote for a single candidate residing in the same district.
The authors conducted their study in three constituencies selected to include communities that were electorally competitive or strongholds for either of the major parties. They videotaped and edited the debates to show 1,991 respondents different portions of the debates on smartphones. The treatment entailed showing respondents different combinations of segments. In one treatment condition (N=391), participants watched only the personal background segment. In another treatment condition (N=409), participants only viewed the policy segment.
This study was implemented during two local elections for county legislature during the 2010 general election in the U.S. Registered voters who had participated in the last 3 party primaries were randomly assigned into three groups—one to receive a positive message about the candidate, one to receive a negative message about the opponent, and one as control (which received no messaging). The messages were delivered by mail, and also provided the individual with the option of donating to the campaign.
Three different data sources were gathered and analyzed: a citizen survey with an embedded conjoint experiment, biographical information on real candidates in one Malawian district, and focus group discussions with real candidates for local office. The primary component of the research was a survey experiment of 604 citizens from Malawi’s Kasungu district. The survey, fielded in support with a local research firm, asked each respondent to evaluate six different hypothetical candidate profiles, resulting in 3,579 total profiles evaluated.
The authors recruited an online national sample (N=1,035) of Black and Latino respondents in May 2013. Respondents were then randomly assigned to read one of three version of a news article about charges of ethics violations against U.S. Congressman Charlie Rangel. In the control condition, respondents were given no information about Rangel’s racial background.
We conducted six Randomized Controlled Trials (RCTs) across five countries to answer this question. The types of information on incumbent behavior provided to voters include legislative performance (Benin), municipal spending irregularities (Brazil), quality of public services (Burkina Faso), municipal government malfeasance (Mexico), candidate quality via debates (Uganda 1), and budget irregularities (Uganda 2). A planned seventh study on incumbent criminality in India did not take place due to implementation challenges.