Niger NECS Impact Evaluation Baseline Report
- The NECS-only and control groups look similar on a broad array of village characteristics, but we found some small differences in household-level characteristics.
- We found statistically significant differences in baseline child-level enrollment and absenteeism between study groups, but we saw no differences in child learning based on test scores in French, the local language, and mathematics.
- We explored potential reasons for differences observed in the data and concluded that they are likely the result of chance.
Mathematica is conducting a randomized evaluation of the Niger Education and Community Strengthening (NECS) project, funded by the Millennium Challenge Corporation (MCC) and the U.S. Agency for International Development. The NECS project involves a variety of activities designed to improve learning outcomes, engage the community, and encourage families to enroll and keep their children in school. All of these activities place a special emphasis on girls. NECS is being implemented in villages that took part in IMAGINE (IMprove the educAtion of Girls In NigEr) as well as in non-IMAGINE villages. (The IMAGINE project was also funded by MCC.)
This report documents the baseline measures of village, household, and child characteristics in NECS-only and control villages as well as baseline child-level outcomes (enrollment, attendance, and test scores). We found equivalence between the NECS-only and control study groups on most measures. We did not find any differences between the two study groups in baseline village characteristics, but we did find some small differences in baseline household-level characteristics.
Households in the control villages appeared to be better off than households in NECS-only villages. We also found statistically significant differences in baseline child-level enrollment and absenteeism between the study groups, but we did not find differences in learning based on test scores in French, local languages, and mathematics.
We explored potential reasons for the small differences we observed in the data, including early intervention effects, lack of adherence to random assignment, and chance. Our findings suggest that the differences are likely due to chance, and these differences will be accounted for in the final evaluation.