Research procedure
Collect official election returns
- Attend polling stations after polls close
- Wait for counting; record official returns
- Aggregate turnout and candidate votes by village
Illustrative summary of the source-described election-day collection across 150 villages in 30 Benin districts, including 60 treatment and 90 control villages. Figures come from electoral commission return sheets, not individual self-reports.
See source →How
When to use
Measuring the official impact of field interventions during an election requires objective administrative vote data. Deploy field research teams directly to local polling stations on election night to collect official vote returns and turnout totals. This method provides aggregate, un-biased voting metrics across treatment and control locations.
Collect official election day returns directly from polling stations as soon as voting closes.
- Deploy research assistants to target polling stations across all experimental villages on election day.
- Wait at the polling stations until election officials complete vote counting.
- Record official turnout numbers and total votes cast for each candidate from the electoral commission return sheets.
- Aggregate polling station figures to construct village-level measures of electoral turnout and vote share.
- Compile recorded data into a central database to compare treatment and control villages.
Why it works
Obtaining official counts immediately after polling closes eliminates reporting delays and prevents administrative loss of localized vote totals.
Example
In March 2011, research teams in Benin visited polling stations across all 150 villages in 30 communes immediately after polls closed, recording official turnout and candidate vote shares from electoral commission reports.
About the study
Participants
Polling stations serving 150 villages across 30 electoral districts (communes) in Benin.
Study Design
On-site election day administrative return data collection across 60 treatment villages and 90 control villages following the March 2011 presidential election.
Results
Official returns showed treatment increased village-level turnout by 3.301 percentage points overall (p < 0.05) and 4.848 percentage points in opposition communes (p < 0.05), but showed no significant difference in candidate vote share at the village level (-0.522 percentage points overall, p > 0.1).