inst/errata/errata.md

In jverzani/UsingR: Data Sets, Etc. for the Text "Using R for Introductory Statistics", Second Edition

Errata for Using R for Introductory Statistics, second edition.

• Thanks to William Kitto, Antelope Valley College for finding errors in the Solutions Manual (9.2 and 9.4).
• Thanks to Demetrios Papanastassiou for finding several issues in Chapter 3 and elsewhere.
• Thanks to Mr. Demetrios Kostakis for finding several issues.
• Thanks to Matthew Gurney for finding an error in Chapter 8.

Chapter 3

• p121: the explanation of $y_i = e^{error_i} \cdot e^{b_0}x_i^{b_1}$ is a confusing. Better wording would be:

This is roughly saying that the mean response in the brain is a multiple, ($e^2.15$), of the body weight raised to $0.752$ and the errors are multiplicative, not additive. (Thanks to Demetrios Papanastassiou.)

• p122. The line fm <- weight/2.2 ~ I(height*2.54/100^2) is missing a parentheses so that the squaring is of the entire expression:

fm <- weight/2.2 ~ I( (height*2.54/100)^2 )

This will change the regression values to

Call:
lm(formula = fm, data = kid.weights, subset = idx)

Coefficients:
             (Intercept)  I((height * 2.54/100)^2)  
                  15.571                     4.239  

The slope of 4 is more reasonable. (Thanks to Demetrios Papanastassiou.)

• p125: The formulas on robust regression should read $\sum(y_i + (b_0+b-1x_i))^2$ and $\sum \rho(y_i + (b_0+b-1x_i))$. (Thanks to Demetrios Papanastassiou.)

• p214, 215: The example on sample uses k to refer to the data vector, but ks in the sample code. These should be the same. IN addition, the labeling of the "spinner" does not match the labeling of the spike plot illustrating the distribution. The values should be shifted by 1.

• p218 line -1: The are to the left of $b$ is being referred to, not to the right of $b$.

• p222: The comments explaining dunif and punif are switched. The explanation that dunif(x=1, min=0, max=3) is 0.3333 is because $1/(b-a)$ takes that value. The explanation for punif(q=2, min=0, max=3) is 0.6667 is because $2/3$rds of the area is to the left of $2$.

• p229: figure 6.6. The illustration was to show that the area to the left of one standard deviation from the mean is the same, but actually we see 1.5 standard deviations, not 1.

• p231 line -9: spurious % symbol following 50.

• p262 line 9: an error of 55.9 should be $0.0559$, not $0.059$.

• p267 line 4: Should be "What is a 90% condidence interval" and not "confidence level."

• p270, line -3: Divisor should be M and not n. This would give a value of $0.9$.

• p273 Question 8.6 Making Coffee. The question asks for an 90% confidence interval, but the answer calculates using 80%.

• p280, line -2: the data set is nym.2002 not nyc.2002.

• p 308, line -16: the argument to qnorm is named mean, not mu.

jverzani/UsingR documentation built on Aug. 3, 2020, 11:57 a.m.