explain the Assumptions of regression with some examples L.I.N.E Linearity The relationship between X and Y is linear Solution: Nonlinear modelling using e.g. squared IVs. Independence of Errors Residuals are statistically independent (no clustering) Solution: Multilevel modelling Normality of Error Residuals are normally distributed Solution: Robust standard errors Equal Variance (homoscedasticity) The probability distribution of the residuals has constant variance (does not look like a fan) Solution: Heteroskedasticity-consistent standard err No multicollinearity There are no disturbingly high positive or negative correlations between IVs (0.7 or -0.7) If such correlations are present among your IVs, retain only one of the highly correlated variables