Independent Count Variables

[Elkhan Richard Sadik-Zada]

What are the methodological concerns with regards to using Poisson distributed count variables as an independent variable in the regression analysis?

[Carlo Lazzaro]

Elkhan:
without further details/examples/excerpts, the best advice is to take a look at the following references, authored by the deeply missed Joe Hilbe:
- https://www.stata.com/bookstore/modeling-count-data/;
- https://www.stata.com/bookstore/nega...al-regression/.

[Marcos Almeida]

If I understood correctly, your concerns encompass the count predictors under the OLS regression analysis. In short, you may use such a predictor "directly" (as a continuous variable) and check the estimations as well as postestimations (leverage, dfbeta, etc). You may logtransform the predictor as well. Please keep in mind that the "normal" expected pattern of distribution mainly relates to the residuals.

[Carlo Lazzaro]

Marcos is correct and I've misread Elkhan's query.

[Elkhan Richard Sadik-Zada]

Many thanks