Title Can Common Knowledge Improve Common Goods? A Field Experiment in an African Democracy
Post date 03/08/2015
C1 Background and Explanation of Rationale

How can we better understand the contexts under which information about politician performance affects citizens' electoral behavior? Research so far has offered inconclusive evidence about the effects of legislator information on voter behavior; in turn, scholars are now trying to understand the conditions under which information might have an effect. We address this question with a field experiment in Benin, in which we propose to treat citizens with information about legislator performance (in the run-up to the country's April 2015 legislative elections) while also studying two moderating factors that have thus far received little attention in the literature: (1) the expectations that citizens hold about their legislators' activities and performance, and (2) whether performance information is delivered privately or in a space allowing for coordination across voters. We locate our study in Benin where voters know little about how well national deputies perform formal legislative activities and instead vote largely based on the provision of clientelistic constituency services.

C2 What are the hypotheses to be tested?

"H1a: Positive information will increase support for the incumbent.
H1b: Negative information will decrease support for the incumbent.

H2: A civics message providing explicit arguments about the value of legislator performance for voter welfare (increasing the salience of legislator performance) will amplify the effect of information on voter behavior.

H3: The provision of public information will have a greater impact on voter behavior than the provision of private information.

H4: Public provision will have the greatest impact in combination with the civics message.

H5: The civics message increases the importance the voter attaches to legislator performance information when making voting decisions.

H6: The civics message will have a greater effect among individuals who have lower prior beliefs about the \emph{value} of legislator performance.

H7: Increasing the dosage of treatment in a commune increases the impact of information.

H8: Public provision of information about legislator performance has a larger effect on voter turnout than private provision.

H9: Public provision increases the voter's perception that others will attach importance to legislator performance information when making voting decisions.

H10: Public provision increases the voter's ability to coordinate by improving expectations of how others will vote.

H11: Public provision should have a stronger effect in rural relative to urban areas, because incentives to vote as a bloc are stronger in rural areas.

H12: Where voters believe their vote can be monitored or observed, public provision will have an even greater marginal impact than private provision.

H13: Private provision will have a stronger impact where the ethnic homogeneity of the village is greater.

H14: Public provision will have a relatively stronger impact where the density of recognized opinion leaders among randomly selected intervention participants is greater.

H15: Information about legislator performance increases voter turnout.

H16:The impact of information will be increasing in the differential between prior beliefs about incumbent legislative performance and the actual performance information given.

H17: The impact of information will be increasing in the differential between the type of performance information given (good/bad) and the voter's prior preference for the incumbent.

H18a: Negative information will have a weaker effect when the voter and incumbent have a pre-existing clientelist linkage.

H18b: Positive information will have a stronger effect when the voter and incumbent have a pre-existing clientelist linkage.

H19a: Negative information will have a weaker effect when the voter and incumbent are coethnics.

H19b: Positive information will have a stronger effect when the voter and incumbent are coethnics.

H19c: \label{itm:ethnic3} These effects will be bigger among individuals with a stronger attachment to their ethnic group.

H20: The impact of information will be higher in information-poor contexts where voters lack other sources of information about incumbent performance.

H21: Because information may already be high in very competitive districts and not matter in very uncompetitive districts, the relationship between the impact of information and political competitiveness will have an inverse-U shape.

H22: The effect of information will be greater the higher the incumbent is on their party's list (because the candidate is more viable).

H23: Information will have a greater marginal effect on women than men.

H24: The marginal impact of information will be increasing in education level.

H25:Poor performing incumbents will react to the intervention by challenging the credibility of the information provided.

H26: Since the information will be available more publicly in the public provision condition, we should observe more backlash against the intervention in the common knowledge group.

H27: Incumbents in treated areas should invest more heavily in vote buying in an effort to counteract the negative impact of performance information.

H28:Poor performing incumbents in electorally competitive areas should be most likely to invest more heavily in vote buying."

C3 How will these hypotheses be tested? *

We will test the hypotheses using regression analysis. We will run all specifications first without control variables other than block fixed effects. To improve precision of estimates, we will also run specifications with appropriate controls. We will test all hypotheses on heterogenous effects with interaction analysis.

C4 Country
C5 Scale (# of Units) not provided by authors
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number not provided by authors
C9 Date of IRB Approval not provided by authors
C10 Will the intervention be implemented by the researcher or a third party? Centre d’Etude et de Promotion de la Democratie (CEPRODE)
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? not provided by authors
C13 JEL Classification(s) not provided by authors