LINGUIST List 14.963

Tue Apr 1 2003

Review: Discourse Analysis/Socioling: Paolillo (2002)

Editor for this issue: Naomi Ogasawara <naomilinguistlist.org>


FUND DRIVE 2003 To give you an incentive to donate, many of our Supporting Publishers have generously donated some amazing linguistic prizes. As a donor you are automatically entered into this prize draw. To find out what's on offer and the rules etc., visit: http://linguistlist.org/prizedraw.html We still have a long way to go, however, to reach our target of $50,000. Please make a donation at: http://linguistlist.org/donation.html The LINGUIST List depends on the generous contributions from subscribers like you; we would not be able to operate without your help. The moderators, staff, and student editors at LINGUIST would like to take this opportunity to thank you for your continuous support. What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Simin Karimi at siminlinguistlist.org.

Directory

  1. John Stevens, Analyzing Linguistic Variation: Statistical Models and Methods

Message 1: Analyzing Linguistic Variation: Statistical Models and Methods

Date: Tue, 01 Apr 2003 14:46:02 +0000
From: John Stevens <stevensjuncw.edu>
Subject: Analyzing Linguistic Variation: Statistical Models and Methods

Paolillo, John C. (2002) Analyzing Linguistic Variation: Statistical
Models and Methods, CSLI Publications.

Announced at http://linguistlist.org/issues/13/13-2564.html


John J. Stevens, University of North Carolina at Wilmington

''Analyzing Linguistic Variation: Statistical Models and Methods'' by
John C. Paolillo explains the statistics of logistic regression within
the context of variationist linguistics. The book's main purpose is
to go beyond the simple instructional manuals of the popular
statistical software packages by providing a convenient resource to
researchers who seek answers to common questions as well as a more
comprehensive explanation of the principles underlying statistical
models of variation. The book examines all aspects of the most
commonly used analytical tool in sociolinguistic variationist studies,
VARBRUL, a multiple regression computer program developed by David
Sankoff based on William Labov's (1969) notion of the variable rule.
The author evaluates VARBUL in the light of other statistical
techniques used in the social sciences and attempts to relate
variationist methods to more formal models of linguistics.

This book assumes no previous familiarity with statistics. It is
written with three different audiences in mind: graduate students and
researchers who are looking for a guide that explains how to conduct
variationist linguistic analyses; more experienced researchers seeking
answers to recurring problems not adequately addressed in the
currently available literature; and researchers from other areas of
linguistics who need to relate variationist analyses to the
theoretical models current in their own sub-fields.

The volume is organized into ten chapters and includes appendices,
references, and an index. The first chapter serves as a general
introduction and provides information about the fundamental concepts
that are key to understanding variationist linguistic analysis and its
goals such as variable rules, chance occurrence, probability,
hypothesis testing, modeling, and the nature of different types of
data. Chapter 1 also gives a partial taxonomy of statistical analysis
procedures and explains the reasons why VARBRUL is the most widely
used program for variationist analysis.

Chapters 2, 3, and 4 address most directly the first audience
mentioned above by focusing on the practical aspects of performing
variationist linguistic analysis. Chapter 2 outlines the goals of
variationist research and considers issues that may affect the
interpretation of the results of statstical analysis. Chapter 3
discusses the types of data required for analysis by VARBRUL and other
software programs (i.e., raw data, tokens, and contingency tables) and
describes procedures for recoding data in order to achieve different
research goals such as testing for significant factor groups and
identifying interaction among factor groups. Chapter 4 explains the
steps involved in running an actual VARBRUL analysis and includes
information on how to read the results and assess the model of
variation in terms of ''goodness-of- fit,'' or how well the
statistical model fits the variation observed in the data (see Young
and Bayley 1996).

Chapters 5, 6, 7, and 8 present the underlying statistical principles
involved in variationist linguistic methodology. Chapter 5 describes
the fundamentals of contingency tables analysis, introduces the chi-
square test of independence, and lays the foundation for logistic
regression modeling, which is further developed in subsequent
chapters. Chapter 6 examines the analytical structure of regression
models and explains how their components are used to assess competing
models of a set of data. Chapter 7 describes the measure of variance
known as log- likelihood and explains how the likelihood ratio test
can be used to achieve an optimal model of variation. Chapter 8
presents the logistic regression model itself and discusses the
procedures and software used to assess model fit.

The final two chapters consider variationist linguistic methods within
the larger theoretical context of statistics and linguistics. Chapter
9 relates the logistic regression model to log-linear models and
general linear models and compares and contrasts the assumptions,
uses, and limitations of each. Chapter 10 examines the notion of the
variable rule within the larger context of formal models of language
and extends the Variable Rule model to other models of language such
as Optimality Theory and Default Inheritance.

This book is written in an engaging, straightforward style appropriate
for graduate students and professionals in the field who are
interested in conducting variationist linguistic research. Although
the text contains numerous typographical errors, context generally
allows the reader to identify them as such without any loss in
meaning.

Paolillo uses many tables and figures to illustrate his explanations
and main points. Particularly useful are the figures and tables used
in Chapter 4, ''Conducting Variationist Analyses,'' which exemplify
the actual print-outs given in a multivariate analysis as performed by
the VARBRUL software program. In addition to leading the reader step
by step through the analysis, the author provides information on how
to read and interpret the results files of the print-outs.

Unlike other books on statistical methods, this manual provides
linguistic examples from actual variationist studies, many of which
may already be familiar to the reader (e.g, Labov's [1972] study of
(r) deletion in New York City department stores). These examples may
serve as potential models for researchers and assist them in the
conception of their own investigations of linguistic variables. Each
chapter ends with a ''Further Reading'' section, which provides
references to works that may provide answers to questions that go
beyond the scope of this volume.

In Chapter 3, ''Variable Linguistic Data,'' Paolillo discusses the
management of different forms of data and explains procedures for
coding and recoding data. However, for most of the technical details
of preparing data for statistical analysis, the reader is referred to
the specific instructions for whatever software package is used. The
author does provide invaluable information on where to obtain
statistical analysis software in Appendix 1, which gives the addresses
(URLs) where several versions of VARBRUL may be downloaded free of
charge from the Internet. These VARBRUL software programs currently
include versions for the Macintosh and PC platforms as well as a
newly- available version re-written for Microsoft Windows called
GoldVarb 2001.

The book provides complete references for all works cited. Although
the index is comprehensive, the volume would have benefited form the
inclusion of a glossary of terms. Such a glossary would be
particularly useful to those unfamiliar with the specialized
vocabulary of statistical analysis. In addition, instructions on how
to write up the results of a variationist linguistic analysis would
have been helpful, especially for novice researchers not sure about
what information, including tables, figures, and measures of fit,
should be reported in a written results section.

''Analyzing Linguistic Variation'' brings together a wealth of
information from several fields to explain the principles of logistic
regression as it is applied in the analysis of linguistic variation.
It will be especially suitable as a textbook for graduate students
learning how to perform variationist linguistic analyses for the first
time. It will also prove to be an indispensable resource for more
experienced researchers who seek a deeper understanding of the
statistical bases of VARBRUL.

REFERENCES

Labov, William. (1969) ''Contraction, deletion, and inherent
variability of the English copula.'' Language, 45:715-762.

Labov, William. (1972) ''The social stratification of (r) in New York
City department stores.'' In Sociolinguistic Patterns, pp. 43-69.
Philadelphia: University of Pennsylvania Press.

Young, Richard, and Robert Bayley. (1996) ''VARBRUL analysis for
second language acquisition research.'' Second Language Acquisition
and Linguistic Variation, ed. by R. Bayley and D. Preston, 253-306.
Philadelphia: John Benjamins.

ABOUT THE REVIEWER

John J. Stevens is Assistant Professor of Spanish at the University of
North Carolina at Wilmington. He is currently conducting research on
variation in the interlanguage of learners of Spanish as a second
language.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue