The sample comprises 2995 households across 599 villages. To construct this sample, the following two steps were followed. First, in 2015 the research team worked with the MLHUD and the Uganda Bureau of Statistics (UBOS) to randomly select 150 parishes in rural Western and Eastern Uganda to participate in the impact evaluation. These parishes were selected from a larger frame of rural parishes eligible for the land titling intervention and fully covered by existing detailed based maps, which will be used during the intervention for the demarcation of lands. Second, up to 9 villages per parish were then randomly selected for the survey sample. Interviewers worked with chairmen and key informants in the community to gather community-level data, and to create a listing of households in the village with information on (i) their marital status, (ii) whether the household head was in a polygamous marriage, and (iii) whether the household owned any land. Five households with husband-wife pairs were randomly sampled from this list (four monogamous and one polygamous households). The condition to be included in the sample was that the household had to own land individually and at least one parcel of land had to be without a certificate of title proving ownership. In each household only one female spouse was initially interviewed; in case of polygamous households the husband chose the wife to be interviewed.
