217 / 762
Chapter 5: Bivariate Data
217
We can also use the
Simple Regression
workbook to help compute the
coefficient of determination. One of the statistics that is displayed in the
output is the coefficient of determination.
Example 5-9:
What is the value of the coefficient of determination for the
model in
Example 5-6
.
Solution:
When we use the
Simple Regression
workbook, the computed
coefficient of determination is
= 0.9597. This value is displayed in
Figure 5-21
. This value is close to one, so the model presented in
Example
5-6
would be a very useful model for the data.
Recall, when we studied the theory associated with the simple linear
regression model, we minimized the error sum of squares to find estimates
for the intercept and the slope for the line of best fit. We also refer to these
errors as residuals. Next, we will analyze graphs of these residuals. These
graphs may help us determine whether we are using the appropriate models
for the data.
Section Review
5-8 Residual Plots
Residuals are just the observed errors. In particular, a residual is the
difference between an actual observed
y
value and the corresponding
predicted
̂
value.
e-Self Review Click here for the Simple Regression Workbook