Correlation and Regressssion - Agricultural Statistical Data Analysis Using Stata - page 188

500

0

-500

4500 5000 5500 6000

Fitted Values

6500 7000

Figure 10.4 Fitted values versus residuals for the rice tiller and yield data.

7000

Grain yield (kg/ha) Fitted values

Grain yield (kg/ha) Fitted values

6500

6000

5500

5000

150 200 250 300

Tillers (no./sq. m.)

350 400

Figure 10.5 Scatter plots and fitted lines for two rice varieties: Milfor 6(2) and Taichung Native 1.

Looking at Figure 10.5, it is quite obvious that there are two dis-
tinct varieties. Running the regression separately for each variety and
examining the residuals will show they are independent and occur
randomly for each. To run both regressions simultaneously, enter the
following command:

by variety , sort : regress yield tiller