Content uploaded by Ron Fricker

Author content

All content in this area was uploaded by Ron Fricker on Aug 08, 2016

Content may be subject to copyright.

Book Review: A Handbook of Statistical Analyses using Stata, 2nd Edition

Sophia Rabe-Hesketh and Brian Everitt, Boca Raton, Fl: Chapman & Hall/CRC, 2000. 210 pp. $39.95

This book is neither a statistical textbook nor a Stata software reference manual. Rather, A

Handbook of Statistical Analyses using Stata is more of a concise Stata “how-to” manual

that takes readers through the analysis of a dozen different data sets, each requiring the

application of one or more statistical methods. It progresses from producing simple

descriptive statistics in Stata through the Stata commands for more advanced analytical

techniques such as generalized linear models, survival analysis, and maximum likelihood

estimation.

In its second edition, A Handbook of Statistical Analyses using Stata is well written and

nicely organized. It is designed for readers either already familiar with the statistical

techniques it presents or for those sufficiently sophisticated to be able to learn the requisite

statistical details from other sources. Readers with solid statistical training, but unfamiliar

with Stata, will find this book to be a succinct and pleasant introduction to the software. As

the authors state in their Preface, “Our hope is that this approach will provide a useful

complement to the excellent but very extensive Stata manuals.” The second edition is an

update of the original edition from Stata 5 to Stata 6.

In the first chapter, the reader is given a concise overview (24 pages) of the Stata interface

and syntax. This chapter provides all of the fundamental commands necessary to become

immediately functional in the software. Then, each subsequent chapter presents a problem

and data set followed by a step-by-step analysis in Stata. The data sets are largely taken

from clinical trials or epidemiological studies. While each chapter does attempt to briefly

describe the particular statistical technique used, the level of detail varies and is often

relatively superficial. However, this allows the authors to get right to point – presenting

Stata in the context of real problems and real analysis.

The scope and focus of the book is best illustrated by its Table of Contents:

Chapter 1: A Brief Introduction to Stata

Chapter 2: Data Description and Simple Inference: Female Psychiatric Patients

Chapter 3: Multiple Regression: Determinants of Pollution in U.S. Cities

Chapter 4: Analysis of Variance I: Treating Hypertension

Chapter 5: Analysis of Variance II: Effectiveness of Slimming Clinics

Chapter 6: Logistic Regression: Treatment of Lung Cancer and Diagnosis of Heart Attacks

Chapter 7: Generalized Linear Models: Australian School Children

Chapter 8: Analysis of Longitudinal Data I: The Treatment of Postnatal Depression

Chapter 9: Analysis of Longitudinal Data II: Epileptic Seizures and Chemotherapy

Chapter 10: Some Epidemiology

Chapter 11: Survival Analysis: Retention of Heroin Addicts in Methadone Maintenance

Treatment

Chapter 12: Principal Components Analysis: Hearing Measurement using an Audiometer

Chapter 13: Maximum Likelihood Estimation: Age of Onset of Schizophrenia

Readers interested in a further description of each chapter should consult the recent review

by Tony Lachenbruch in The American Statistician (2000). While this review was for the

first edition, the chapter descriptions are still relevant.

A Handbook of Statistical Analyses using Stata could be useful in filling specialized

pedagogical niches. While the text assumes a level of technical statistical sophistication that

makes is unsuitable for introductory or lower-level courses, it might form the basis for an

advanced undergraduate or graduate course in applied data analysis. It could be particularly

relevant in a curriculum where students must first take a number of theoretical statistical

courses, especially in schools of public health and other health-related settings, followed by

a course focused on applied data analysis. It could also be used as a supplementary text for

other courses that require the use of Stata.

Each chapter has exercises at the end. However, I found the exercises relatively simple and

limited to posing minor variants of the analysis presented within the chapter, mainly to

motivate the use of alternate Stata syntax. These exercises were sufficient in the context of

learning to use Stata. However, I suggest that expanding the exercises would improve the

text’s usefulness in the classroom. In particular, I would like to see new exercises for each

chapter based on a second data set that is not used in the chapter. This data set would be

similar in scope and analytical requirements to the one presented in the chapter, but also

sufficiently different so that students could do their own analysis (prompted by the exercise

questions) from scratch. This would allow the instructor to use the chapter data set and

analysis in a lecture and yet still have other material for student homework and exercises.

Readers should note that the book is focused on the “Intercooled” version of Stata for

Windows. Also, in the never-ending race of software manuals to keep up with the latest

software revisions, Stata 7 has recently been released, which may date some of the material

presented. However, since I do not have access to Stata 7 as of this writing, I cannot judge

how much different that version is and how much of an effect those changes will have on

this text. None-the-less, this book provides an excellent, interesting, and efficient

introduction to Stata that I can easily recommend. And, I am sure a third edition is already

in the works.

Ronald D. Fricker, Jr.

RAND

References

Cox, Nick. 2000. Book Review: A Handbook of Statistical Analysis using Stata.

Biometrics, 56, pp. 650-651.

Lachenbruch, Peter A. 2000. Book Review: A Handbook of Statistical Analysis using Stata.

The American Statistician, 54, pp. 153-154.