Scatter plot of residuals on the y axis and fitted values on the x axis to detect non-linearity, unequal error variances, and outliers.
Details
Characteristics of a well behaved residual vs fitted plot:
The residuals spread randomly around the 0 line indicating that the relationship is linear.
The residuals form an approximate horizontal band around the 0 line indicating homogeneity of error variance.
No one residual is visibly away from the random pattern of the residuals indicating that there are no outliers.
See also
Other residual diagnostics:
ols_plot_resid_box()
,
ols_plot_resid_hist()
,
ols_plot_resid_qq()
,
ols_test_correlation()
,
ols_test_normality()
Examples
model <- lm(mpg ~ disp + hp + wt, data = mtcars)
ols_plot_resid_fit(model)