The LTR pilots applied a participatory and low-cost process to systematically cover a total of 3,513 households with some 15,000 plots in four areas (one of them urban) that were chosen to reflect the country's heterogeneity.
The challenge of this study was lack of baseline data to make a credible assessment of the pilot program. This challenge was addressed by sampling on both sides of the borders of the pilot areas-using high precision satellite images and the cadastral survey-that allows the comparison of outcome variables between households inside (treated) and outside (non-treated) of the borders of the pilot cells. The discontinuity created by administrative boundaries in the introduction of the pilot program is, therefore, exploited as an identification strategy on the assumption that households close to a cell boundary, before the start of the program, were similar in unobservable factors affecting relevant outcomes. The sample was designed to yield numbers of households in each pilot cell equivalent to their share in the total, with a size of 3,554 households with some 6,330 land parcels, intended to be split equally across pilot and their neighboring cells.
The sample was to be distributed equally on both sides of the pilot cell boundary to create a treatment group (within the titled cell) and a control group (those just across the border in nonprogram cells). Parcel index maps created by the program were used to sample within pilot cells. For adjacent (control) cells, researchers used high resolution satellite imagery to visually identify dwellings that could then serve as a sample frame.
