2
$\begingroup$

I am trying to understand how the gender of the user relates to her/his use of language in tweets. For this purpose, I have downloaded all tweets of several hundred users and have obtained positive and negative sentiment scores for each tweet. The data looks as the following:

enter image description here

Should I analyze this as mixed-effects model with tweet_possent (or tweet_negsent) as DVs with twitter_user_gender nested under twitter_user? Or should I be converting this to panel data?

I am not sure if panel data analysis is more appropriate here.

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

Being aggressively practical here: Use the same methodology as prior well-cited works on the subject and/or what your audience is more accustomed too.

Realistically this question boils down to the meta-question: Should we deal with this primarily as econometricians (panel data) or as biostatisticians (longitudinal data)? In Biostatistics, longitudinal data are repeated observations of the same individuals (Twitter users in the use case above), these repeated measurements are what Econometrics terms as panel data. CV.SE has some great threads looking into the distinction between fixed/random effects:

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ It appears that either mixed-effects model or panel data analysis is fine as long as there is a precedence in the target discipline. Hope I didn't misunderstand your answer. $\endgroup$
    – SanMelkote
    Commented Jan 26, 2022 at 13:07
  • 1
    $\begingroup$ You don't misunderstand it. :) $\endgroup$
    – usεr11852
    Commented Jan 26, 2022 at 14:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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