Using R and RStudio for Data Management, Statistical Analysis, and Graphics, 2nd Edition

i

i

“K23166” — 2015/1/28 — 9:35 — page 87 — #113

i

i

i

i

i

i

6.6. EXAMPLES 87

The AIC criterion also suggests that the model without the interaction is most appropriate.

It may be useful to change the default reference level for variables. The default R design

matrix (see 6.1.4) can be changed and the model re-fit.

> contrasts(ds$sub) = contr.SAS(3)

> aov3 = lm(cesd ~ sub + genf, data=ds)

> summary(aov3)

Call:

lm(formula = cesd ~ sub + genf, data = ds)

Residuals:

Min 1Q Median 3Q Max

-32.13 -8.85 1.09 8.48 27.09

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 33.52 1.38 24.22 < 2e-16

sub1 5.61 1.46 3.83 0.00014

sub2 5.32 1.34 3.98 8.1e-05

genfM -5.62