Intent-to-Treat (ITT) Reduced-Form Effect Estimation
RCTClinical Trial
Estimating the impact of an assigned intervention without adjusting for non-compliance measures the policy impact of initial assignment rather than the direct effect of receiving the actual treatment.
Picture this
If a doctor assigns patients to a specific health routine but some patients do not follow it, checking how much healthier the whole assigned group became measures the effectiveness of giving out the advice, rather than the physical impact of doing the routine itself.
What the evidence says
The intent-to-treat reduced-form estimate yielded a coefficient of ~0.280 to 0.326, which reflected the true treatment impact (~0.955) diluted by an overall compliance rate of approximately 29.3%.
- Who was studied
- N = 27,255 randomized students in grades 4–8 across 6 urban school districts.
- How
- Ordinary Least Squares regression of student end-of-year achievement directly on the predicted effectiveness of the randomly assigned teacher ($\hat{\tau}_{jAssigned}^{S}$).
What to do
Calculate the intent-to-treat reduced-form regression of outcomes on assigned treatment variables to establish the baseline causal impact of program assignment prior to instrumental variable adjustments.
From the source
"In instrumental variables estimation, this is commonly referred to as the 'reduced form' or 'intent-to-treat' model."
Have We Identified Effective Teachers? Validating Measures of Effective Teaching Using Random Assignment