Errata
On page 45 the standard deviations are used instead of the variances when computing the intraclass correlation coefficient (ICC). As given, the values is .626/(.626+2.246)=.218. However, the standard deviations were used instead of the variances. Thus, the equation given on page 45 should be .6262/(.6262+2.2462)=.392/(.392+5.045)=.072. This error of using the standard deviation instead of the variance continues for several others examples. The lme4 package reports information on random effects in terms of the standard deviations, not in terms of the variances. Thus, one needs to ensure that they square the terms in the output. This can be done via the VarCorr() function, which extracts the variance and correlation components from an lme4 fitted model. Thanks to Paul Bliese (University of South Carolina) for pointing this out.
On page 47 the denominator of R-square sub 1 should use M0, not M1. The numbers themselves are correct but the symbols in the denominator are wrong. Thanks to Hannah Müggenburg (Institut für Humangeographie) for pointing this out.
We say, on p. 60, that the following code produces uncorrelated random effects: "Model3.11 <- lmer(geread ~ gevocab + age + (gevocab|school) + (age|school), data=Achieve)" – but it does not! Rather, the following code should be used: "Model3.11 <- lmer(geread ~ gevocab + age + (1|school) + (-1 + gevocab|school) + (-1 + age|school), data=Achieve)." Thanks to Amir Kasaeian (Tehran University of Medical Sciences) for pointing out that our code creates an unintended random effect.
On p. 25 the equation after 2.2, "S^2" is missing from the left hand side of the equation. Thus, rather than "j" the left hand side should be "S^2_j" (as the preceding part of the sentence says. Additionally, the numerator on the right-hand-side is not squared but it should be – that is, the numerator should be the sum of squared deviations between the y_{ij} and the group mean. Thanks to Tae-Rae Kim (Samsung Advanced Institute for Technology) for pointing this and issues 5–8 out to us.
On p. 73 the t value in text (second paragraph) of 2.19 has an associated p-value of .005. It should be p< .05 (i.e., it should be "<.05" not "<.005).
On p. 74 (very bottom) the summary() and intervals() functions should be applied to "Model4.4" (yet they are applied to "Model4.3" instead).
On p. 96 in the last paragraph the text states "Returning to Model 5.3...." However, it should have said "Returning to Model 5.2."
On p. 131 in Equation 7.4 the first component on the right hand side should be b_i0 (not b_i1).
On p. 92 "Model_5.3" should have "na.action=na.omit" added to the model for lme (as there is missing data). Thanks to Camilo Alberto Caudillo Cos for pointing this out.
On p. 14 we say "we can see that the predicted GPA for the first individual in the sample was approximately 0.465 points below the actual GPA" but it should be "....465 points above" (i.e., Observed - Predicted = 2.5 - 2.94 = .465). Thanks to David Alexandro (University of Connecticut) for pointing this out.
Example data for the GPA example was provided by Jerrell Cassady (Ball State University). "Cassady" was misspelled as "Cassidy" for the data set in the R code.
The first line on p. 18 should say "negative slope" not positive slope, with the second line being -.152 for the regression coefficient (not .152). Thanks to Wei Wang (Monash University) for pointing this out.