close× Call Us
close×
Title Voter Registration in Kenya: A Field Experiment to Overcome Information and/or Operational Barriers to Registration
Post date 03/07/2017
C1 Background and Explanation of Rationale

What are the barriers to voter registration in Africa? Can simple educational interventions about the process of voter registration improve registration rates? Or are operational interventions required? While a wide body of political science literature focuses on possibilities for information to improve political participation, few studies explore the material barriers to participation. This is surprising, given that in Africa practical, spatial, and/or economic constraints loom large. One case in point is voter mobilization. A core tenet of democracy assistance and governance aid in the developing world is that civic education improves political participation, and a growing literature confirms this belief. But are these gains large, and if so, relative to what? We suggest that a missing element in this broad field of research is the study of practical, material constraints.

C2 What are the hypotheses to be tested?

Our treatments are 1) "local registration opportunity", where polling station locations will be visited by election commission officials with portable voter registration equipment (R), and two additional information treatments: 2) face-to-face canvassing (I_f), and 3) mass, impersonal text-message invitations in the informational intervention (I_m).

The first set of hypotheses tests the effectiveness of each intervention.

H1 Localization (R) to polling station i, increases registration rates for i.
H2 Canvassing (I_f) in polling station i, increases registration rates for i.
H3 A text message blast (I_m) to polling station i, increases registration rates for i.

The second set of hypotheses explicitly test the conjecture that the absence of opportunity for Kenyans to register is a larger barrier than the absence of information to voter registration. Furthermore, previous literature suggests that direct, face-to-face appeals (e.g., door-to-door canvassing) generate larger effects than impersonal appeals. We test this in Kenya directly.

H4 Localization (R) to polling station i increases registration rates for i more than canvassing (I_f).
H5 Localization (R) to polling station i increases registration rates for i more than a text message blast (I_m).
H6 Canvassing (I_f) in polling station i increases registration rates for $i$ more than a text message blast (I_m).

In the next set of hypotheses, we test whether the interventions combined are more effective than alone.

H7 Localization and canvassing together (RI_f) at polling station i increases registration rates for i more than localization alone (RC).
H8 Localization and a text message blast together (RI_m) to polling station i increases registration rates for i more than localization alone (RC).
H9 Localization and canvassing together (RI_f) to polling station i increases registration rates for i more than localization and a text message blast together (RI_m).

We also set out to explore the effect for different groups of the population:

H10 The intervention (R) increases registration rates more for blocks farther from constituency office.
H11 The intervention (R) increases registration rates more for blocks in poorer areas than for blocks in richer ones.
H12 The intervention (R, I_f, or I_m) increases registration rates more for women than for men.
H13 The intervention (R, I_f, or I_m) increases registration rates more for voters 25 or younger than for older voters.
H14 The intervention (I_m) increases registration rates more in areas that were treated more heavily.
H15 The intervention (I_f) increases registration rates differentially based on the type of canvassers.
H16 The interventions (R, I_f, or I_m) increases registration rates differentially based on the day of intervention.
H17 The intervention (R, I_f, or I_m) increases registration rates more in counties characterized by opposition support.

Finally, we test whether the interventions might have effects on polling station areas not directly targeted by the IEBC.

H18 The intervention (R, I_f, or I_m) to polling station i increases registration rates for j (spatial effect).

C3 How will these hypotheses be tested? *

To examine the impact of information-based and operational interventions to improve voter registration in Kenya, we use a 2x3 factorial design, which crosses two interventions. The first involves an informational intervention: providing citizens with basic information about voter registration via canvassing or SMS. The second involves an operational intervention designed to reduce the cost of voter registration: bringing voter registration officers and technology to the village, removing the need for citizens to travel to a distant central office to add their name to the register. This study targets 1,674 polling stations across seven counties in Kenya.

C4 Country Kenya
C5 Scale (# of Units) 1,674 polling stations
C6 Was a power analysis conducted prior to data collection? No
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number #039-2016
C9 Date of IRB Approval 12 October 2016
C10 Will the intervention be implemented by the researcher or a third party? IEBC
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? Yes
C13 JEL Classification(s) C93, D72