Please fill out this short user survey of only 3 questions in order to help us improve the site. We appreciate your feedback!
Information on polls, turnout and partisanship in South Africa's first competitive election
Last registered on December 08, 2016


Trial Information
General Information
Information on polls, turnout and partisanship in South Africa's first competitive election
Initial registration date
December 08, 2016
Last updated
December 08, 2016 9:44 AM EST
Primary Investigator
University of Oxford
Other Primary Investigator(s)
Additional Trial Information
Start date
End date
Secondary IDs
People are predicted to turn out more in competitive elections where their vote is more pivotal (Riker and Ordeshook 1968) and behave according to theory in lab experiments (Levine and Palfrey 2007; Duffy and Tavits 2008). It is difficult to test this finding in real world settings. There is higher turnout in close elections (Blais and Dobryznska 1998; Nevitte 2000; Franklin 2004; Ashworth et al 2005), but close elections have more media coverage and party spending. Experiments could give information on electoral outcomes, but the control group usually know an election is competitive (Enos and Fowler 2010 provide one exception in a runoff election). Effects of competitiveness on partisanship are ambiguous and differing effects are found in lab work. There may be underdog effects, where people swing behind a party which is behind, or bandwagon effects, where they swing behind a winning party (Levine and Palfrey 2007).

I test the effects on turnout, partisanship and political engagement of being exposed to information, delivered face-to-face, about the likely outcome of the election. The study randomly assigns 2,023 participants from around Soweto in Johannesburg to one of four groups -- control, basic information and two treatments. The basic information message provides information about voting processes at municipal elections and the date of the election. The treatment messages provide information about likely electoral outcomes from the first publicly available pre-election polls conducted by Ipsos-Mori in South Africa.
External Link(s)
Registration Citation
Orkin, Kate. 2016. "Information on polls, turnout and partisanship in South Africa's first competitive election." AEA RCT Registry. December 08. https://doi.org/10.1257/rct.1662-1.0.
Former Citation
Orkin, Kate. 2016. "Information on polls, turnout and partisanship in South Africa's first competitive election." AEA RCT Registry. December 08. http://www.socialscienceregistry.org/trials/1662/history/12324.
Sponsors & Partners

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
Experimental Details
Control group: this group receive no intervention.

Information group: this group receive information regarding the importance, process and implications of voting in the municipal election.

Treatment group: this group will receive the same information as the Information Group as well as a message about the competitiveness of the upcoming municipal election. The message will make use of statistics collected by Ipsos-Mori in a series of pre-election polls and published on eTV news.
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
The outcome variables of interest are: voter turnout; party voted for, partisanship and voter engagement.
Primary Outcomes (explanation)
Voter turnout: (1) One measure uses self-reported turnout (both must indicate that the respondent did vote); (2) A second measure makes use of more objective information - whether or not the respondent has a stamp in his/her ID book (if the respondent claims that he/she voted).
Party voted for: this is captured by self-reported vote, administered using a secret ballot provided to the respondent.
Partisanship: this is a combination of ANES questions on party identification and strength of support.
Voter engagement: this is captured with a combination of variables capturing the respondent's extent of political knowledge and his/her participation in election-related activities (e.g. participation in rallies).
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
Baseline survey: Over twelve days (including ten weekdays and two Saturdays) prior to the 2016 municipal election in South Africa, enumerators will recruit respondents to survey. Surveys will take place on tablets. Respondents will be randomly assigned to one of four groups (two treatment groups, information and control).

Intervention delivery: Information-group participants will receive a face-to-face message during the baseline survey regarding the municipal election voting process. Treatment-group participants will also receive this information as well as one of two messages regarding the prospective competitiveness of the upcoming municipal election.

Endline survey: After the elections respondents will be resurveyed.

Experimental Design Details
Randomization Method
Randomization done during the survey through the application of an algorithm in surveyCTO.
Randomization Unit
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
The experiment is conducted in 12 locations on Soweto but randomisation is at individual level
Sample size: planned number of observations
2,023 registered voters
Sample size (or number of clusters) by treatment arms
2,023 South Africans: 648 control, 682 information, 693 treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We initially calculated power based on data from research in similar areas in 2015 and from the Afrobarometer and South African Social Attitudes Survey. We then reran calculations based on the baseline and endline data. For varying attrition levels (5-10%), the MDEs are: Control vs Information Message: 0.1358-0.1618 Information Message vs Treatment 1: 0.1529-0.1821 Information Message vs Treatment 2: 0.1561-0.1604 Control vs Treatment 1: 0.1544-0.1851 Control vs Treatment 2: 0.1588-0.1891
IRB Name
University of Cape Town
IRB Approval Date
IRB Approval Number
Post Trial Information
Study Withdrawal
Is the intervention completed?
Intervention Completion Date
July 26, 2016, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
November 10, 2016, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Data Publication
Data Publication
Is public data available?
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)