Rural development cooperative soil-testing fee waivers and fertilizer-blend adoption among smallholder women in Rwanda: a regression-discontinuity analysis of plot-size eligibility thresholds

Rural development cooperatives in Rwanda frequently waive soil-testing fees for members whose holdings fall below a plot-size threshold, on the assumption that removing the fee will induce adoption of fertiliser blends calibrated to local soil conditions. The design treats the threshold as the assignment variable, specifies the local average treatment effect at the cutoff, and distinguishes the fee-waiver channel from the soil-information channel that the waiver also unlocks. Using the DHS Program indicator 'Knowledge of any method of contraception among married women' (FP_KMET_W_ANY) as a benchmark for the scale and dispersion of a comparable survey-based indicator across 38 African countries and 162 country-survey observations between 1986 and 2025, the paper calibrates the precision that a plot-size discontinuity would need in order to detect plausible adoption effects. The benchmark has a mean of 91.776 and a standard deviation of 11.342, with country values ranging from 57.80 in Chad to 99.90 in Eswatini. Plot-size thresholds are a defensible identification strategy only when the running variable is measured without heaping, when the waiver is the sole treatment at the cutoff, and when the fertiliser-blend recommendation is not itself conditional on the test. It proposes a specification that separates these channels and states the conditions under which the design would recover a credible local effect.