The proportion of variance explained (PVE) is another measure of how good the regression line predicts obtained (actual scores). The values of PVE range from zero (the regression line has no predictive value) to one (the regression line predicts each obtained score with perfect accuracy). The definitional equation for proportion of variance explained is
PVE = ——
There is also a computational equation for PVE, which is
PVE = ———
PVE is sometimes called the Coefficient of Determination and is represented by the symbol
The proportion of unexplained variance (or variability) is sometimes called the Coefficient of Indeterminance and is represented by
The notion of explained variance is important. It is used in discussions of reliability of measurement as well as in investigations of effect size and in statistical power analyses.
NOTE: This means that it is a very good idea to understand this concept of explained and unexplained variance.
|©1999 by J.T. (Tim) Stocks, Ph.D.