Timeline for Multicollinearity when controlling for a variable
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 31 at 12:50 | vote | accept | user20501139 | ||
| Jul 31 at 10:27 | vote | accept | user20501139 | ||
| Jul 31 at 10:34 | |||||
| Jul 30 at 16:58 | comment | added | gung - Reinstate Monica | @user20501139, a "red herring" is an irrelevant distraction. It's an expression in English. What I mean is that choosing between running a correlation or a regression is a waste of time, because they are the same thing presented differently. Unfortunately, the "isolated effect" of something is not specific enough to be meaningful. You could fit a model w/o volume, or with both & just interpret pattern. | |
| Jul 30 at 13:34 | comment | added | user20501139 | Thank you, this is helpful! I didn't quite understand what you meant with red herring, did you mean the observations under point 4? This probably also happens if I run a regression with pattern metric and volume, it's just that I noticed it when running a correlation. If that happens with a regression as well, would you also say it's normal behaviour and not a sign that something is off? And did you mean that a correlation with a residualized pattern would answer a different question than including volume into a regression would? What I want to answer is the isolated effect of the pattern | |
| Jul 30 at 13:11 | history | answered | gung - Reinstate Monica | CC BY-SA 4.0 |