1
$\begingroup$

I am pretty new to econometrics, and I am thinking about substituting a continuous variable (say assets for a company) with a categorical one containing a number based on which quintile the continuos value was (so 1-5). Obviously doing so I will lose some information however I was hoping that doing so would semplify the interepretation of an interaction variable made of the original continuos variable and of another continuos variable. Especially as I am going to plot a marginsplot (in STATA). I wanted to ear the opinion of someone more experienced and if this is a good/ feasible idea or it will backfire in some way.

Thanks in advance

Improve this question
$\endgroup$
1
  • $\begingroup$ Obviosly the categorical variable will converted in a series of dummies with the i command in STATA $\endgroup$
    – Hustler885
    Nov 18, 2020 at 17:52

1 Answer 1

Reset to default
2
$\begingroup$

You can use the continuous variable "Assets" as stand-alone, and use its categorical incarnation for the interaction term.
While we may be accustomed to use "automatically generated" interaction terms as part of ready-made functional forms like the translog, nothing forbids us to introduce selectively interaction terms in a regression setting, when we have an economic, structural argument as to why this specific interaction term is meaningful and why it belongs to the regression.

Therefore, using two different versions of the same variable is in principle acceptable as long as a solid argument can justify it.
From a technical point of view, it does not create any issues.

Improve this answer
$\endgroup$

Your Answer

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

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