Hi, I am working with the Health and Retirement Study and going to use the mean of observations across waves as the study sample(collapsed sample). But I also want to restrict the age of respondents to 50-65 in my final study sample but am confusing the strategy to achieve this goal. The reason I want to use the collapsed sample is that I want to make my study sample only one observation per person in order to match it with the genetic sample. I have two ideas:

1. drop those whose age<50 or age>65 from the panel sample, and then collapse this panel sample to my final study sample. This is like a rolling sample, people will be excluded from the sample if he is over 65.
2. Do nothing in the panel sample, but drop observations based on the birth year of the respondent in the collapsed sample.

Does anyone have any suggestions for me? Many studies have the step of restricting age in the study sample (for example, the one study attached below), but I am not sure what I should do.

Thanks in advance! Array