A Comment on 'Gender Bias in Parental Attitude: An Experimental Approach' by Begum, Grossman and Islam (2018)

Begum et al. (2018) study gender bias in parental attitudes using an experimental approach in rural Bangladesh. Households are reported as randomly assigned to treatment conditions in a lab-in-the-field allocation task. We show that the group assignment was inherited from Islam (2019), a previous non-randomized experiment conducted in the same region. The lack of randomization contradicts the design descriptions provided by the authors in Begum et al. (2018) and elsewhere, and raise concerns about the validity of comparisons across treatment groups. It also points to serious shortcomings in the reporting and transparency of the study design-issues that mirror those that led to the retraction of Islam (2019) from the European Economic Review.