Establish correct database for estimation

[melvinhlwong]

A good match is defined by use using two criteria:

1. Preserve about 80% of treatment cells after CEM matching
2. Standardized mean difference between treatment and control units around 0.2 or less

For this issue we need to

• Filter the WDPA without any matches after "sturges" bins (usually small PA)
• Create R script for manual bin adjustmet
• Filter all WDPAs that are part of a follow up phase of the project (multiple treatment)
• WDPA with less than 30 cells. Each cell has 500 ha. Thus we would drop PAs with less than 150 sqm
• Manually adjust CEM bins for WDPA datasets with 75-100% treament cells consverved to achieve a mean stand. difference balancing below 0.2 for each matching covariate