Typos and other errors in Regression and Linear Models
(R. Darlington, McGraw-Hill, 1990)

1/18/97

Errors in formulas are the most important to change; they are marked with asterisks.
This list does not include errors that are unlikely to mislead anyone, such as misspellings.
Note the separate section titled "Typos in Homework Answers".

p. 19, Fig. 1.5B. The X-values for the 3 columns of dots are 5, 6, 7, not 1, 2, 3.

p. 46 line 4. Change "X1 and X2" to "X2 and X1".

p. 61, 5 lines above the formula appears the phrase "standardized variance of Xj." Change Xj to Y.

p. 62. In Question 5 change "means" to "marginal means".

p. 74. Title of Section 3.3.3 should be "Three Versions of the Phenomenon"

p. 90, statement 2a. Change "offer" to "differ"

p. 91, statement 2e, end of second line. Change "worse" to "worst"

p. 121, line 5. "Sec. 1.3" should be "Sec. 1.5.2"

*p. 136, very bottom of page. The denominator (1+r) should be (1-r).

p. 137, 3 lines above start of sec. 5.5.5. prj should be pr1

p. 141. Three lines above the formula "Raw total - mean + model + error" appears a Y with a hat, and on the next line down is a Y with no hat. The hat belongs on the second Y, not the first.

p. 141, next to last line. Reference to "Table 5.1" should be to Table 5.3.

p. 150. In the long formula, the final R should be r.

p. 158, figure caption. b1 and b2 should be sr1 and sr2

p. 160, line 7. Y should be "hat-Y"

*p. 162, formula for RS. In the denominator, (N - 2 x P + 2) should be (N - 2 x P - 2)

p. 193, third line of Sec. 8.1.3. Change "Singularity always occurs..." to "Singularity occurs only..."

p. 198, line 9. Change "prj" to "prj and srj"

p. 201, sec. 8.1.8, line 6. Add 3 more 0's to the first number, making it 10,000,000,000,001

*p. 254, first large formula. The entire right side of the formula should be enclosed within a radical (a square root sign). This makes the formula fit the example with t = 2 given a few lines later.

p. 254, second large formula. Add "|z| =" to the left, making it "|z| = (df -...

p. 285. Just below the middle of the page, a paragraph begins "The limitation..." At the end of that same sentence, change "a square term" to "the particular nonlinear term you choose"

p. 289, last line of text. "linearity" should be "nonlinearity"

p. 313, Sec. 13.1.3. Add to the end of the section the sentence, "When the addition of a covariate C changes bX, there is still just one slope bX that applies to all cases, while interaction means that different slopes apply for different values of C."

*p. 345, formula for Cooki. Replace P by (P+1).

*p. 356, large formula. The denominator of the first fraction should be P, not N.

*p. 359, formula for Cooki. Replace P by (P+1).

*pp. 366-367. Delete the entire last paragraph of Sec. 14.2.3, which begins "Normal-scores correlations tend to be..." This entire paragraph resulted from a bug in a computer program. This changes the example at the end of p. 367. Delete the "-4" from the formula for t, and change the final sentence on the page to "We have N = 20, P = 3, dfr = N - P - 1 = 16, t = 2.435, p = .027 (two-tailed)."

pp. 371, 373. The last paragraphs of these two pages use the word "unidirectional" to mean two different things. Not really wrong, but certainly confusing.

p. 385, about 60% of the way down the page, on the far right. Change "critz = .424" to "critfz = .424".

p. 396, Table 16.1, column V. Add 10.0 to the numbers given for Dallas-Fort Worth, Denver, Houston, Salt Lake City, New Orleans, Tucson, Tampa, and Phoenix. This amounts to adding a leading "1" to each of these numbers. For instance, Dallas-Fort Worth is changed from 6.0 to 16.0, Denver from 4.0 to 14.0, etc.

p. 510. In line 5 change "IF DF < 4" to "IF DF2 < 4"

TYPOS IN HOMEWORK ANSWERS

p. 486. Chap. 2, #1. In the fourth inequality, change ">" to "<"

p. 488. Chap. 3, #5, line 1. -2/3 should be +2/3

p. 494, line 12. Replace "The last also yields MS(error)" with "The first three yield MS(error)"

p. 494, fourth line from bottom. Change ".12" to ".08"

p. 495, fifth line from bottom. Change "model pop" to "model logpop"

p. 498, 11th line from bottom. Change "idf" to "zif"