1
$\begingroup$

Residuals vs Fitted

Hi there, I tried to interpret and evaluate the given plot. Most statistic-books say that if homoscedasticity in regression models is present, it would be recognizable through a triangle-shape in a residual vs fitted plot. For me, there is no clear triangle-shape present. What I find more complex to interpret is that most of the residuals are in the very beginning. What does that tell me regarding the regression-assumptions? I could neither find such an example in a statistic book, nor online.

$\endgroup$
1
$\begingroup$

If the predictions are reasonably accurate, then the fact that most of the predictions are lower values implies also that most of your dependent variable's values are also lower values. (I.e., there is a positive skew in the marginal distribution of the DV.)

Your assessment of heteroscedasticity seems reasonable.

The diagonal lines are evidence of discreteness, which means that many observations on the dependent variable take the same numeric value. This is evidence of non-normality, but it's not a gross deviation.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.