How to deal with outlier for panel data?

[Arief RRH]

Hello everyone, do you know the steps how to deal with outlier in panel data in STATA, I mean how the observation can categorized as outlier with the difference time e.g year, outlier is identically only for cross section data, cmiiw guys

[Nick Cox]

An outlier in my experience usually calls for a model with appropriate data generation process. There isn't a single, simple way either to identify what is an outlier or how to "deal" with it. More discussion at (for example) https://stats.stackexchange.com/ques...iers-with-mean (a thread that is much broader than its title).

[Carlo Lazzaro]

Arife:
unless the so called outliers are a blatant example of mistaken data entry, they are simply a matter of fact, and should be kept in as the data generating process justifies them. An example that springs to my mind is the positively skewed gamma distribution that is usually used to modelizing the distribution of healthcare resources for a given disease: a small fraction of patients may be mistakenly considered outliers because they require much more visist/tests/whatever than the rest of the patients with the very same diseases.

[Arief RRH]

Thanks Nick & Carlo