Hello everyone,
I need to evaluate the impact of the Labour-Intensive Public Works programme on social cohesion using DiD. I chose a set of control variables and an outcome variable. The thing is to choose the right control and treatment groups for DiD estimation. The initial settings are as follows:
"Public Works Programmes (PWPs) are transfer programmes that require participants to work on public projects for wages that are deliberately set below the market rate. Participants eligible for the LIPW are first selected via non-categorical community-based targeting without clear and standardized procedures and criteria. Second, the poor choose (self-select in) projects in public sites.
The survey was conducted in four purposely selected village catchments where the implementation of the LIPW had started in early 2018. The first wave was administered in February 2017. The second round took place two years later in 2019. The households participating in the survey were randomly selected in every catchment. The decision which households become eligible to participate in the PWP was the outcome of the regular targeting process, thus, not randomly assigned. The 2019 survey made evident that in one catchment there are no LIPW participants, whereas in the other three catchments both LIPW participants and non-participants are present."
My question is should I exclude the catchment without any LIPW participants from my data sample and use the rest three catchments for DiD analysis? This means that I use participants from these three catchments as treatment group and non-participants from the same three catchments as control group. Or is it incorrect to exclude this fourth catchment from the analysis?
Thank you!
I need to evaluate the impact of the Labour-Intensive Public Works programme on social cohesion using DiD. I chose a set of control variables and an outcome variable. The thing is to choose the right control and treatment groups for DiD estimation. The initial settings are as follows:
"Public Works Programmes (PWPs) are transfer programmes that require participants to work on public projects for wages that are deliberately set below the market rate. Participants eligible for the LIPW are first selected via non-categorical community-based targeting without clear and standardized procedures and criteria. Second, the poor choose (self-select in) projects in public sites.
The survey was conducted in four purposely selected village catchments where the implementation of the LIPW had started in early 2018. The first wave was administered in February 2017. The second round took place two years later in 2019. The households participating in the survey were randomly selected in every catchment. The decision which households become eligible to participate in the PWP was the outcome of the regular targeting process, thus, not randomly assigned. The 2019 survey made evident that in one catchment there are no LIPW participants, whereas in the other three catchments both LIPW participants and non-participants are present."
My question is should I exclude the catchment without any LIPW participants from my data sample and use the rest three catchments for DiD analysis? This means that I use participants from these three catchments as treatment group and non-participants from the same three catchments as control group. Or is it incorrect to exclude this fourth catchment from the analysis?
Thank you!
