I was wondering. What would be the effect of the presence of autocorrelation on R² or R²_adjusted? Especially in a dataset used in multiple linear regression?
edit: So the correlation of each variable with itself in time.
Asked
Active
Viewed 92 times
3
-
A somewhat related question: ["$R^2$ and adjusted $R^2$ in presence of overlapping observations"](https://stats.stackexchange.com/questions/432173). – Richard Hardy Jul 08 '20 at 19:20
-
Similar https://stats.stackexchange.com/questions/160421/serially-correlated-regressors. Although not a good answer provided. – Cagdas Ozgenc Jul 09 '20 at 10:03
-
Yet another vaguely related post: https://stats.stackexchange.com/questions/331139/showing-that-r-squared-might-not-be-useful-in-time-series-data – Christoph Hanck Jul 09 '20 at 11:47
-
1Can you clarify the effect *relative to what* you are interested in? – Christoph Hanck Jul 09 '20 at 11:48