0
$\begingroup$

I am examining a petition involving all UK constituencies. In this dataset 2 of the 632 constituencies have not participated in the petition - in terms of data quality how does this affect my examination?

I am examing which parts of the UK tend to vote for left/right ideals.

$\endgroup$
0
$\begingroup$

When I read this, I guess you are up to a causal model?

When you use methods like linear regression or logistic regression, $n=630$ is a healthy sample size. So no problem with that. What could be a problem is that you have censored data (you don't know about the remining two constituencies).

IMO the best way to deal with this problem is to cleary state that you have missing data for this constituencies, so that you only can make claims for the remaining ones. It is a common problem that there is missing data. If it is really only about two constituencies and if you look at "left/right" tendence, the missing data is not a big problem as long as the missing observations are not highly important (e.g. super large) and there is no reason to believe that the missing observations come from an entirely different data generating process (e.g. extremely different to all other observations).

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.