When we do correlation & regression as a topic, we know correlation is talking about the relationship between two variables, but what exactly is “regression”?
So this is a bit of English. To regress something is to reduce, and what do you think we are regressing here? For most regression questions, we start off with a bunch of sample data points and somehow, into the GC it goes and out comes a regression line. The GC did regress (reduce) something, the errors between the data points, one can consider it as the distance between the fitted values and collected data values. So the regression line, in layman terms, is really just the best fit line that students always attempt to draw during their science SPA haha. Just that we got an even more accurate one, I describe to students that this is the “best of the best” best fit line. 🙂
[…] independent of translation and scaling? #2 5. Why can’t we extrapolate regression lines? 6. What is a regression line? 7. Simulation for Hypothesis Testing 8. Type I vs Type II Errors 9. A little about Central Limit […]