I’m having some difficulty recognizing the difference between the two; but from what I can gather serial correlation more refers to correlation in the residuals of a regression analysis, which multicollinearity refers to the the independent variables themselves being highly correlated? Would this be an accurate statement and can someone give me an example? Any help appreciated.
Multicollinearity is, as you say, strong correlation of two independent variables. For example, if you were trying to model home prices, you would likely find that for inputs square footage and number of bedrooms display multicollinearity (generally the more bedrooms, the larger the house).
Serial correlation is correlation of a time-series with itself (lagged one period or more). A time series for weekly rainfall in a very seasonal area would display high, positive serial correlation (if it rained a lot last week, it’ll probably rain a lot this week).
thank you!
You’re welcome.