Research design
Delay control delivery until after endline
Randomly assign market and village clusters
- Immediate treatment
- Deliver after baseline; complete evaluation surveys.
- Postponed control
- Complete endline first; then deliver the same intervention.
Illustrative summary of the crossover timetable. After evaluation, 46 control markets and 74 control villages received the same materials and presentations within the same calendar year.
See source →How
When to use
In randomized controlled trials testing beneficial social or information interventions, withholding effective programs from control communities permanently creates ethical concerns. Program directors use this method to deliver the proven intervention to control clusters after all evaluation data collection ends.
Roll out proven experimental interventions to postponed control clusters immediately after completing final evaluation surveys.
- Assign sample market and village clusters randomly to either an immediate treatment group or a postponed control group.
- Deliver the experimental intervention exclusively to treatment clusters during the active evaluation period.
- Complete all baseline and endline survey rounds across both treatment and control clusters.
- Verify that data collection for the main evaluation trial is fully completed.
- Deploy intervention teams to all control markets and villages using the exact same methods, scripts, posters, and pamphlets as the original treatment.
- Execute the crossover intervention within the same calendar year to ensure control communities receive program benefits prior to the next agricultural season.
Example
After completing endline impact surveys across Morogoro Region, Tanzania, researchers rolled out the information intervention across all 46 control markets and 74 control villages. Field teams distributed identical shop posters, customer pamphlets, and village extension presentations before the subsequent agricultural season.
About the study
Participants
Postponed control markets (N = 46) and control villages (N = 74) in Morogoro Region, Tanzania.
Study Design
Randomized controlled trial crossover rollout executed immediately after endline survey completion using identical market posters, pamphlets, and extension presentations.
Results
All control markets and villages received full information intervention coverage within the same calendar year, eliminating ethical concerns regarding unequal access to verified market data.