Share
Explore BrainMass

Basic question about datasets. Not a

I'm working with a cross-section of the NLSY97 (using the program Stata). I'm a new econometrics grad student. I'm trying to make sure I understand how these datasets work but I'm almost embarrassed to ask this question: the observations are linked somewhere in the dataset right?

What I mean is, its not just statistics on the answers to individual questions, like it appears in the codebook. The program must know all the answer of a particular oberservation number (that is, a particular respondent), right?

Otherwise, if you wanted to remove respondents who answered a particular way on one question ("question A" lets call it) from another question ("question B") you wouldn't be able to do it would you? Do you see what I'm asking?

I'm confused because it seems like this information must exist somewhere in the dataset, based on the type of commands I'm able to give the program. But I don't see that represented anywhere. So, theoretically, could I pick a respondent (observation number) and see all the answers for that one particular person?

Solution Preview

Question 1: "The observations are linked somewhere in the dataset right?"
Answer 1: The observations can be linked via the unique identifier for each respondent. In the NLSY97 supplement, it is called UID.xx. This variable should be in the dataset for each round and each line.

Question 2: Otherwise, if you wanted to remove respondents who answered a particular way on one question ("question A" lets call it) from another question ("question B") you wouldn't be able to do it would you? Do you see what I'm asking?

Answer 2: You can. The problem is that it will require some effort to ...

$2.19