Multilevel and Longitudinal Modeling Using Stata, Volumes I and II, Third Edition Paperback – 11 May 2012
- Choose from over 13,000 locations across the UK
- Prime members get unlimited deliveries at no additional cost
- Find your preferred location and add it to your address book
- Dispatch to this address when you check out
Frequently Bought Together
Customers Who Bought This Item Also Bought
Enter your mobile number below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.
Getting the download link through email is temporarily not available. Please check back later.
To get the free app, enter your mobile phone number.
What Other Items Do Customers Buy After Viewing This Item?
Top Customer Reviews
Most Helpful Customer Reviews on Amazon.com (beta)
Using Stata" was published in 2005. The second edition was released in 2008, and now this
third edition in 2012. With each edition the scope of the model's discussed in the text
has increased. This release is in fact a 2-volume work, with the first volume devoted to
panel models having a continuous response (or dependent variable), and the second to
discrete response panel models.
Nearly every panel model in the literature is addressed in the text. Discussion begins
with a review of basic linear regression and provides the basics upon which more complex
models will be developed. Then variance-components are addressed, explaining concepts such
as between-subject heterogeneity and within-subject dependence. Following this the authors
introduce fixed and random effects models, and then delve into the details of both random
intercept and random coefficient models. Mixed effects models is given a thorough examination.
Following a discussion of subject-specific models, the authors turn to population-averaged
or marginal models, as well as growth curve models. The first volume concludes with chapters
devoted to higher-level models with nested random effects and crossed random effects mdoels.
Throughout the volume each model, where appropriate, is approached as one, two, and higher
Volume two addresses one, two, and higher level categorical response models, count models,
and survival or duration models. The emphasis, of course, is on understanding data that
is structured as panels - whether clustered or longitudinal. Zero-inflated count models,
for example, are not discussed, nor are generalized binomial, generalized Poisson, or
generalized negative binomial models. But near every type of categorical response panel model
is discussed -- in full.
Volume one is over 500 pages in length; volume two is a bit shorter. Together the two volumes
consist of 974 pages plus nearly 40 Roman numeral pages. Stata statistical software is used
throughout the text, which is dually published by Stata Press and Chapman & Hall/CRC. Stata
and Limdep econometric software are in my opinion the two most compehensive panel-modeling
statistical packages available, with SAS the next best in this regard. Stata, as a general
purpose stat package has a much wider range of capabilities, as does SAS. It is therefore a
very good choice of software to use for examining this class of models. It is also a comparatively
easy programming language. The authors have written 'gllamm', a Stata command that allows estimation
of many of the more complex models disussed in the text, including, for example, a
three-level random coefficient logistic regression model. Most examples though rely on Stata's
built-in commands, plus it's Mata matrix programming facility.
This two-volume work is in my opinion the foremost text on multilevel models.
It uses Stata for examples, but any text that uses examples to explain difficult statistical
concepts and methods needs to use some type of statistical software. Stata is ideal for this
type of modeling, so has been used in this text. Researchers who use other software for
modeling; eg SAS, R, SPSS, etc, can use the methods taught in this volume with their preferred
package, insofar as it has the capability to estimate a particular type of model.
I highly recommend this two-volume set of books to anyone with an interest in modeling
multilevel and longitudial models, regardless of their preferred statistical software. It is
the most comprehensive work available on applied multilevel modeling. It is also very well
written, with each model examined in a very clear manner. Data sets and author-written code is
provided on the book's web site. Readers therefore are able to replicate the exmaples in the book,
or to adapt them for their own projects.
The book contains lots of output and allows to "work along" with chapter pretty easily. I only have two complaints about the book; both linked to the exercise sections. Firstly, some exercises were confusing. Most of these were fine, but I felt that each chapter had one/two exercises that contained vague questions that were not always relevant to the chapter. Secondly, worked examples are scarce. The majority of the exercises do not contain a "worked example" or a set of solutions. Only two or three of the exercises per chapter contain a set of solutions which could be found online. Although I had the chance to test myself at the end of a chapter, I rarely had the opportunity to confirm whether I was on the right track or not. I found this a bit annoying. Overall, a great resource.