Counterfactual analysis techniques for rigorous impact evaluation of extension technology dissemination

Abstract

How do you measure the impact of something that didn't happen? That question sits at the heart of extension impact evaluation. This research compared five counterfactual analysis techniques propensity score matching (PSM), difference-in-differences (DID), regression discontinuity design (RDD), instrumental variables (IV), and synthetic control applied to evaluate technology dissemination programs across three Iranian provinces. Secondary data from 1,247 farming households collected between 2019 and 2023 were analyzed. DID showed the strongest performance controlling temporal bias (0.89), while RDD was most effective against selection bias (0.91). No single method dominated all categories. Programs evaluated using at least two complementary methods produced impact estimates with 34.7% lower standard errors than single-method evaluations. The findings argue for methodological triangulation as the standard approach.