Data Analysis part 4

Part 4 of the Data Analysis course. The first 20 min is non-essential material on (philosophical) types of knowledge and comments on what can be learned (or not) about each type. The second half is on Simpson's paradox, and design of experiments to avoid drawing mistaken conclusions due to confounding variables.

[@郝嘉伟]

Thanks for the video and tips. I would like to ask for the uncovered slides and terms (i.e., Nuisance parameters), do we also need to understand them ourselves (for our report) or just read them for our reference? Thank you for your help in advance.

[@FaranWhyde]

You won't need to write about Nuisance Parameters. Well, the assignment (up on Moodle) does involve computing an estimate of two parameters, using chapter 4-5 of the notes. In a situation where there were two parameters, and you only wanted to estimate the first, but you needed to estimate the second one at the same time to do so, then the second one would be a nuisance parameter. Not a hugely important term, but in case you encounter it in a textbook you need to know it is a technical term.