Need Clearer Explanation of Confounding, Multicollinearity, and Masking Effect

[muhammadalmaskhan]

Hi PyMC Team,

I noticed that the explanations for confounding variables, multicollinearity, and masking effects are quite brief. These issues arise when transitioning from simple to multiple linear models, and clearer explanations with practical PyMC examples would be very helpful.

Would it be possible to improve this section availible on page 125 onwards, of the book Bayesian Analysis with Python
Second Edition? Thanks!

[aloctavodia]

Do you want that to be explained in the pymc documentation?

One book that discuss those topics is statistical rethinking. The books uses R, but here you can find the code ported in PyMC https://www.pymc.io/projects/examples/en/latest/gallery.html#statistical-rethinking-lectures

[muhammadalmaskhan]

Hi @aloctavodia ,
Thank you for your response. The book is indeed excellent; however, some topics require clearer explanations. At times, certain concepts can be confusing. For instance, on the page mentioned above, the simulation of X1 as a confounding variable is somewhat unclear.
Thanks in advance

[aloctavodia]

Thanks for the feedback. Still unclear where do you want that clarification to be addressed.

This is the github repository for the pymc package, issues/discussions should be related to pymc features or bugs.

If you want that clarification to be added to the documentation of pymc, then better to open an issue in the pymc-examples repository https://github.com/pymc-devs/pymc-examples.

If this is just a question better to ask it here https://discourse.pymc.io

If this is a request for a clarification in a future edition of BAP then open an issue in the book's repository https://github.com/aloctavodia/BAP3