Understanding Data

Understanding Data

Understanding Data

Understanding Data

Synopsis

For statistics to be used by sociologists, and especially by students of sociology, they must first be easy to understand and use. Accordingly this book is aimed at that legion of professional sociologists and students who have always feared numbers; it employs much visual display, for example, as an easy way into the data. Also, the book is written in a relaxed and enthusiastic way that reassures apprehensive students without watering down what they must be taught.

Classical statistics were developed to meet the requirements of the natural sciences; as such they reflect the more deductive nature of hypothesis development in these sciences. However, they have offered the sociologists little in the way of techniques for exploring messy data in the context of incomplete theories.

This book attempts to remedy those weaknesses, and it emphasizes exploratory data techniques which sociologists will find useful in their day-to-day research. The primary characteristics of exploratory techniques discussed by the authors are simplicity, resistance and elucidation. Its coverage is from basic statistics up to multiple regression and two-way anova. The inter-relationship between exploratory and confirmatory techniques is stressed, and, through the alternating presentation of each, the students learn to master data analysis: to be and to feel in control.

Excerpt

It has now been more than 15 years and several thousand students since the original edition of Understanding Data was completed. the original premise of the text was that insights gained from data exploration not only would be important in their own right but could carry over to analogous confirmatory statistics. We anticipated that social science students, who are often very weak mathematically, would comprehend the traditional materials better than if taught either through mathematical demonstrations or in a cookbook fashion. As a consequence, the text was built around a set of core elements from John Tukey’s (1977) highly innovative text Exploratory Data Analysis. in our view the approach works well, a view generally supported by students, other teachers and several reviewers. But 15 more years of teaching, including much trial and error, plus comments from students, other users and reviewers have indicated the need for some changes. Other changes have been mandated by the increased access to personal computers and the more general availability of high quality software for exploratory data analysis.

The overall organization of the text as well as our views about how best to teach the material are largely unchanged. We have added sections on a variety of topics though, including interval estimation (Chapter 9), tabular analysis (Chapter 14), transformation as a way of treating interaction effects (Chapter 16), dummy variable analysis and high influence points (Chapter 20). in addition, nearly every chapter contains a section on using the computer. This has been keyed to minitab (release 7), one of the most user-friendly of the many packages available, with good exploratory capabilities. Consequently, we have also slightly de-emphasized handwork in the text although we continue to feel that at least some handwork is important for developing a sense of mastery and control. Several chapters have been substantially reorganized, most notably Chapters 5 and 6 on batch transformations, and several text examples have been simplified, resulting in somewhat clearer exposition.

For teachers experienced only with traditional statistics courses, selecting effective data sets can be a problem. Along with the data being real (‘warts and all’) and appropriate to the task, experience has shown that good data sets meet most of the following criteria:

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.