Sanctuary Control Group Protocol
RCTClinical Trial
In field experiments evaluating public services, active outreach to treatment units risks polluting control groups or creating artificial service deprivation if control units are strictly denied baseline access.
Picture this
Imagine testing a new school tutoring program. To evaluate it cleanly, teachers actively advertise and offer tutoring to classroom A, while taking zero proactive steps toward classroom B. However, if a student from classroom B walks in and asks for help, they are not turned away.
What the evidence says
Uninterrupted linear trends in control firm contacts confirmed counterfactual stability, while treatment firms experienced a 488% jump in counselor visits and a 152% increase in phone calls without distorting control baseline behavior.
- Who was studied
- N = 7,438 unique firms across 129 local employment agencies in France.
- How
- Stratified Randomized Controlled Trial with a 6.5-month "sanctuary" period where control firms received no proactive counselor contact but retained non-refusal access upon request.
What to do
Implement sanctuary control protocols that halt proactive marketing while preserving standard, request-based service access to maintain clean control counterfactuals in field RCTs.
From the source
"The control group was 'sanctuarized' until March 31, 2015: No proactive action was to be taken towards these firms. Importantly, control firms were not refused services if they requested it."
Are_Active_Labor_Market_Policies_Directed_at_Firms_Effective_Evidence.pdf