Good morning to everyone,
I have a sample of workers from several Italian companies, and I would like to make it representative of all national workers. I am currently standaridzying some variables (such as job classification, gender, contract type, etc.) in order to append my DB on the national one run a probit model that would provide the Mills ratio.
I have seen this approach commonly used with the probability of employment. Do you think it would be feasible to use the probability of being included in my sample as the dichotomous outcome instead? However, this would mean that no information in the national database would have that dummy variable equal to 1.
Alternatively, do you have any suggestions for constructing weights?
Thank you in advance for your help!
I have a sample of workers from several Italian companies, and I would like to make it representative of all national workers. I am currently standaridzying some variables (such as job classification, gender, contract type, etc.) in order to append my DB on the national one run a probit model that would provide the Mills ratio.
I have seen this approach commonly used with the probability of employment. Do you think it would be feasible to use the probability of being included in my sample as the dichotomous outcome instead? However, this would mean that no information in the national database would have that dummy variable equal to 1.
Alternatively, do you have any suggestions for constructing weights?
Thank you in advance for your help!
