BMO Recruitment Intent Survey Stratification
RCTClinical Trial
Randomizing firm-level interventions across broad employer populations without accounting for baseline hiring intentions introduces noise and reduces statistical precision in measuring policy impacts.
Picture this
Think of testing a new fertilizer across farmland. If researchers sort plots based on whether farmers plan to plant crops this season before applying the fertilizer, they can accurately measure growth without confusing unplanted fields with failed crops.
What the evidence says
Stratifying assignment using annual recruitment survey data ensured balanced treatment-control groups across firm size tiers (73% under 26 employees) and industry sectors (42% services, 25% commerce, 28% manufacturing/construction).
- Who was studied
- N = 7,438 establishments drawn from approximately 400,000 French businesses surveyed annually in the BMO (Besoin en Main d'Oeuvre) platform.
- How
- Stratified Randomized Controlled Trial using baseline BMO survey responses on 2014 hiring intentions, local agency attachment, and workforce size tiers to stratify assignment.
What to do
Incorporate annual firm recruitment intention survey data as a core stratification variable when designing randomized labor market field experiments.
From the source
"The sample was stratified by indicators for the agency to which the firm was administratively attached, if the firm had intended to recruit in 2014 and by the number of employees on the firm's payroll..."
Are_Active_Labor_Market_Policies_Directed_at_Firms_Effective_Evidence.pdf