Coefficient of Determination
Categories: Metrics
Sadly, this isn’t a measure of how hard we’ll try to beat that really tough level on Candy Crush. Instead, the Coefficient of Determination (or r2 as those in the know call it) is the percentage of change in the y variable that is due exclusively to changes in the x variable for bivariate, or two-variable, data.
Let’s say that the value of r2 for data taken on the price per pound of grapes and the weight of grapes sold that day is 0.791. That means 79.1% of changes in the weight of grapes sold are due exclusively to changes in the price per pound of the grapes. And 28.9% of the changes in the weight of grapes sold are due to changes in other variables that we probably didn’t measure...because we were too busy eating the grapes. Also, more simply, it’s one measure of how closely the points in a scatterplot are to the best fit line calculated via linear regression.