Analysis showed that these two variables provided the best
predictive regression modes, explaining 22% of the variance
in self-care behaviors. Similarly, after using univariate analysis
to identify significant variables, variance items in the qualityof-
life model (shown in Table 4, Model 2) included age,
functional class, self-care behaviors, illness knowledge, living
independently, and personal income. Analysis showed that
functional class, living independently, and age (adjust R2 =
.41) provided the best predictive regression modes, explaining
41% of the variance in quality of life.