Question: In a recent study, Banerjee et al. (2015) provide evidence from a randomized controlled trial on how a multifaceted program can help to improve the lives of the very poor. The file banerjee data.csv includes the data for one country analyzed by Banerjee - Ethiopia. The variable id is a unique identifier number for each household. The variable treatment indicates whether the household was provided an income support program (1) or not (0). Because this was a randomized controlled trial, the treatment status of the household was randomly assigned by Banerjee et al. (2015) as part of the experiment. The variables foodsecurity bsl and foodsecurity end are indices that measures five components of food security.5 The food security indices are numeric, and are constructed so that a score of 0 indicates a household is at the mean food security level for all the households included in the sample, a positive score indicates the household is more food secure than the mean household in the sample, and vice-versa for a negative score. The suffix _bsl refers to the measure of food security before the program was implemented, and the suffix _end refers to the same measure 3 years after the program was started.

1. Before the program was initiated, was there any difference in mean food security between households that were eventually to receive the treatment compared to those that did not? Is the difference statistically significant (i.e., are you at least 95% sure there's a difference)? Why do you think this is?
2. Is the program associated with a change in mean food security for treated households relative to untreated households (to determine the change in security, subtract the baseline value from the endline value)? How big was it? How confident are you that the program is associated with a change in food security?
3. Does this provide evidence that the program caused a change in food security for treated households?

Price: $2.99

Solution: The downloadable solution consists of 4 pages
Deliverable: Word Document