I have a sample with wage data of which I know it is not representative of the data. To be specific, there are many high-paying wages of pastors but I know that they made up at most like 0.5% of the workforce but like 25% of my sample. Is it consistent to first sample from the pastoral wages to get only a few observations and then use these datapoints with the original dataset (without all the pastoral wages) to conduct the bootstrap? Or is there a different way consistent with bootstrap assumptions to correct the sample, such that it is more likely representative of the population?
Asked
Active
Viewed 31 times
