The easiest way to check for the homogeneity of variance (homoskedasticity)
across different levels of the independent variables is to examine a scatterplot of
the residuals against the predicted values. If the spread of the residuals appear to
be constant across the different levels of the dependent variables, there is not
much likelihood of a problem. It is important to mention that regression is also
robust to violations of homoskedasticity. However, mild violations of this
assumption are not worrisome (Tabachnick and Fidell, 2001).
From Figures 10.3 and 10.4, it is evident that, while the assumption of
homoskedasticity is not violated for the weight for age regressions, it is violated
for the case of height for age regressions, since the spread thickens around the
center.