Errata for Multilevel and Longitudinal Modeling Using Stata, Second Edition
The errata for Multilevel and Longitudinal Modeling Using Stata, Second Edition, are
provided below. Click here for an
explanation of how to read an erratum. Click here
to learn how to determine the printing number of a book.
(1,2)  Chapter 1, p. 17, third line under equation 
Here t = 99.53, ...

Here t = 10.09, ...

(1)  Chapter 3, p. 132, last line of third paragraph 
SES) is often called a contextual effect or compositional effect.

SES) is often called a contextual effect in contrast to compositional effect.

(1,2)  Chapter 4, p. 165, first two histogram commands at the bottom of the page 
. histogram reff1 if pickone==1, normal xtitle(Predicted random slopes)
. histogram reff2 if pickone==1, normal xtitle(Predicted random intercepts)

. histogram ebs if pickone==1, normal xtitle(Predicted random slopes)
. histogram ebi if pickone==1, normal xtitle(Predicted random intercepts)

(1,2)  Chapter 4, p. 167, last group of Stata examples on page 
. gsort +reffm1 f
. generate rank = sum(f)
. generate labpos = reffm1 + 1.96*reffs1 + .4
. serrbar reffm1 reffs1 rank if f==1, addplot(scatter labpos rank, mlabel(school) msymbol(none) mlabpos(0)) scale(1.96) xtitle(Rank) ytitle(Prediction) legend(off)

. gsort +reffm1 pickone
. generate rank = sum(pickone)
. generate labpos = reffm1 + 1.96*reffs1 + .4
. serrbar reffm1 reffs1 rank if pickone==1, addplot(scatter labpos rank, mlabel(school) msymbol(none) mlabpos(0)) scale(1.96) xtitle(Rank) ytitle(Prediction) legend(off)

(1)  Chapter 5, p. 196, three lines above section 5.5.5 
can be eliminated by double differencing or “differenceindifference”, ....

can be eliminated by double differencing, ....

(1)  Chapter 6, p. 281, exercise 6.7.2, lefthandside of the equation 
(1) 
Chapter 6, p. 284, exercise 6.9.6c 
c. Run the gllamm command from step 3 ...

c. Run the gllamm command from step 2 ...

(1) 
Chapter 6, p. 284, exercise 6.9.6d 
d. Compare the estimates with those from step 3 ...

d. Compare the estimates with those from step 2 ...

(1) 
Chapter 10, p. 467, exercise 10.4.2 
2. Obtain the estimated intraclass correlations of the latent
responses for ...

2. Obtain the estimated intraclass correlations for ...

(1) 
Chapter 10, p. 470, exercise 10.9.5 
5. Perform likelihoodratio tests for the two variance components
using a 5% level of significance.

5. Perform likelihoodratio tests for each of the variance
components using a 5% level of significance.

(1) 
Chapter 11, p. 478, first sentence of paragraph above section 11.3.4 
The reason for choosing year to be at level 2 and to have a random
intercept for each firm at level 3 is to minimize the computational
burden.

The reason for choosing year to be at level 3 and to have a random
intercept for each firm at level 2 is to minimize the computational
burden.

(1) 
Chapter 11, p. 502, exercise 11.2.6 
6. Fit the model ... Explain how and why the estimated
coefficient of lnfuel changes compared with the model
that treats year as random.

6. Fit the model ... Explain how and why the estimated
coefficient of log fuel price changes compared with the model
that treats year as random.
