Resilience Index and Measurement Analysis 2018

The sample includes 705 households in total, of which 395 are refugee households (202 in the Kyegegwa District and 193 in the Kamwenge District) and 310 are part of the host communities (158 in the Kyegegwa District and 152 in the Kamwenge District). The sample size is determined based on power of design, which, in addition to taking into account the size of the population, uses the estimated impact we expect to see based on previous exercises, the standard deviations on the main indicator of interest (FAO-RCI), as well as measures to reduce possible estimation errors. The sample size has been calculated in order to be able to detect a minimum impact of 10 percent, with a 95 percent level of confidence. Other assumptions include (i) a null expectation in differences between the two populations (refugee/host) at the beginning, but possibly seeing one over time; (ii) a correlation between the resilience capacity between the two groups of 0.4 and 0.5; (iii) differences in the standard deviations between the two groups, i.e. more homogeneous characteristics amongst the host communities than the refugees; and (iv) intra-cluster correlation. This estimation was done at both settlement and district level clusters. The 20 percent oversampling was added at settlement level, to account for the expected fluid nature of refugee status. The sample is representative at district and settlement levels.