Kohler and Kreuter's third editon of "Data Analysis Using Stata", is 497 pages in length and is
published only in paperback form. It is the most recent publication in the series of texts on statistics
published by Stata Press. Stata is a well-used general-purpose commercial statistical application, and
its series of books on statistics have been popular among both Stata users and others who are simply
interested in the subject matter of a book in the Series. In fact, many of the Series books can be
thought of as texts on a particular area of statistics --- using Stata for examples. Others
are more Stata-specific, eg. books dealing with Stata programming, graphics, or data
management. Kohler and Kreuter's text is of the latter variety. On the other hand, many of the data
analytic principles discussed in the book are applicable to data management in general, so have
application beyond a particular software application. Moreover, the author's discuss more than just how
to manage data; they do so with respect to the preparation of graphics, and managing both
descriptive and inferential statistics.
Although a number of data sets are used for examples in the text, most examples relate to the 5411 observation
German Socioeconometric Panel data, a survey of German households containing demographic, income, employment,
and other similar type information. The author's guide the reader on setting up the data for various
types of analysis and for using Stata's graphical interace facilities. Stata code for all examples is
provided, and explained in detail. The reader gains an ability to use Stata for data management, graphical
setup and analysis, as well as model setup and testing.
I should also mention that the book includes a chapter on creating Stata do and ado files. "do" files
are simply batch files -- what are called 'scripts' in R and 'macro's' in SAS. "ado" files are
"automatic do files" or "commands", which are similar to R functions. The majority of the official commercial
Stata pacakge is programmed using Stata commands. Commands authored by Stata users look the same as those
authored by Stata developers. Stata also consists of binary code, but much of it is in terms of commands,
many of which were originally contributed by users.
finally, I should also say something more about Stata Press, which gives an insight into the various books
in the Series, including the text of this review. The Series began as a single text in 1999 authored by
Stata staff on maximum likelihood estimation using Stata. The remainder of the Series began in 2001 with
the publication of a user-authored text on Generalized Linear Models. Eleven years later there are 20 separate
titles in the Series, which includes one text in French and another in Korean. Many of the texts have
multiple editions. In total, with the publication of Kohler and Kreuter's third edition text, Stata Press
has published 40 books in the Series. Nine more separate-title books are scheduled to be published by the
end of 2013.
I highly recommend this book for anyone who has an interest in data management, in particular to Stata
users who must also manipulate data as part of job.