Multiple linear regression models a continuous outcome as a linear function of several predictors, with the same diagnostics as simple linear regression.
Mechanically identical to simple linear regression; the menu split is purely a UX convenience for picking multiple predictor columns at once.
The key extra diagnostic in the multivariate case is the variance-inflation factor (VIF) for each predictor — it tells you how much each coefficient's SE is inflated by collinearity with the other predictors. VIF > 5 ⇒ cautious; VIF > 10 ⇒ serious problem, consider dropping one of the collinear predictors or combining them.
Numeric response + ≥ 2 predictors.
Same as linear_regression + a VIF column in the coefficient table.
Each β_j is the partial effect of X_j holding all other predictors fixed — the same as a partial correlation re-scaled. This 'holding other predictors fixed' interpretation only makes sense when the other predictors plausibly *could* be fixed in practice.
If two predictors are highly collinear, neither β can be precisely estimated even though their combined contribution to R² is large. Report this honestly — don't pretend one of them is significant when the data can't separate the two.