
Paired t-test
I am conducting a study whereby I am trying to determine whether or not participation in a microcredit program by women in Cajamarca, Peru reduces the chances that these women are living in poverty. To determine the likelihood of whether or not an individual woman is in poverty, the women completed PPI questionnaires before entering the program and subsequently after completing one loan cycle.
In order to determine the probability that participation in the program reduced the likelihood of a woman to live in poverty (vs. the null hypothesis), I intend on running a paired t-test on the results - as I have data on each woman before and after the intervention. Do you feel that the poverty likelihoods as provided by the PPI poverty table (in this case I used the USAID comparison) is an appropriate test statistic to be used in a t-test?