The Art of R Programming

As usual, a more detailed accounting can be obtained via the callstr(lma).

The estimated values ofβiare stored inlma$coefficients. You can display

them by typing the name at the prompt.

You can also save some typing by abbreviating component names, as

long as you don’t shorten a component’s name to the point of being ambig-

uous. For example, if a list consists of the componentsxyz,xywa, andxbcde,

then the second and third components can be abbreviated toxywandxb,

respectively. So here we could type the following:

> lma$coef

(Intercept) examsquiz[, 1]

1.1205209 0.5899803

Sincelma$coefficientsis a vector, printing it is simple. But consider what

happens when you print the objectlmaitself:

> lma

Call:

lm(formula = examsquiz[, 2] ~ examsquiz[, 1])

Coefficients:

(Intercept) examsquiz[, 1]

1.121 0.590

Why did R print only these items and not the other components oflma?

The answer is that here R is using theprint()function, which is another

example of generic functions. As a generic function,print()actually hands

off the work to another function whose job is to print objects of classlm—

theprint.lm()function—and this is what that function displays.

We can get a more detailed printout of the contents oflmaby call-

ingsummary(), the generic function discussed earlier. It triggers a call to

summary.lm()behind the scenes, and we get a regression-specific summary:

> summary(lma)

Call:

lm(formula = examsquiz[, 2] ~ examsquiz[, 1])

Residuals:

Min 1Q Median 3Q Max

-3.4804 -0.1239 0.3426 0.7261 1.2225

Coefficients:

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

(Intercept) 1.1205 0.6375 1.758 0.08709.

examsquiz[, 1] 0.5900 0.2030 2.907 0.00614**

...

18 Chapter 1