How to tell if there is a homoscedasticity of the model based on this plot?

Question

I am building regression model of cholesterol predicted by 4 dietary components. I want to check if the assumption of Homoscedasticity is satisfied. I plotted Residuals vs Laverage plot.

Unfortunatelly I cannot tell from this plot whether we can say Homoscedasticity or not... I know in the perfect scenario the residuals should be spread quite equally, though this data set is quite big (25 dietary components), therefore I was wondering where is the limit of deciding: we can/cannot day there is Homoscedasticity.

Maybe there are some other tests I can make to test that?

enter image description here

• Adrian Keister: You can plot the residuals against the fitted values, and you can also plot the residuals against the predictors. You're looking for relatively equal variance from a baseline throughout your plots.
• mdewey: The scale-location plot is ideal for what you are planning.

Answer 1

Scale-Location is used to check the homoscedasticity of residuals (equal variance of residuals). If the residuals are spread randomly and the see a horizontal line with equally (randomly) spread points, then the assumption is fulfilled. Check the link: https://bookdown.org/jimr1603/Intermediate_R_-_R_for_Survey_Analysis/testing-regression-assumptions.html

• Nick Cox: What you mean is better phrased as points equally scattered around a horizontal line. The SEE itself is a constant and itself can only plot as horizontal.