Page 1 of 1

some suggestions

Posted: Thu May 30, 2019 2:42 pm
by Nurzhan
1) More detailed conceptual framework on how migration and remittances may affect the educational outcomes of those left-behind can be useful for further identification of mechanism of influence. Especially on the potential interaction effect of migration and remittances.

2) If data allows it can be useful to use additional criteria for education performance (other than the school enrollment), say exam scores etc. Because some empirical papers argue that absence of parents because of the migration may have psychological negative effect with consequences on poor education performance.

3) If data is longitudinal why not to use the panel data fixed effect ?

4) For PSM estimations alternative matching algorithms (other than the NN) would serve as robustness analysis. Also on PSM
Rosenbaum bound analysis can be suggested.