Social Science Methodology

Sociological Methodology 2001 edited by Michael Sobel, Mark P. Becker (Sociological Methodology, Vol 31: Blackwell) This volume of Sociological Methodology touches on both long‑standing and more recent themes in social statistics. In the first chapter, Adrian Raftery discusses all of these themes, reviewing the analysis of cross-tabulations and statistical methods developed to analyze survey data on individuals. Turning to the future, he discusses the limitations of some current work and outlines developments important for the field in the coming years, such as social networks, the analysis of longitudinal network data, spatial statistics, and social interactions. Durlaf's outstanding and forward­ looking work on what is perhaps the most important issue in sociology, the interdependence between group and individual, proposes models that can be used to inform important sociological issues, such as whether to have child, to move, or how many years of education one seeks. The editors of this volume have taken exceptional care to be mindful of the needs and interests of readers of Sociological Methodology and offer here a very rich and rewarding spectrum of issues and perspectives.

CONTENTS: RAFTERY Statistics in Sociology, 1950‑2000: A Selective Review; DURLAUF A Framework for the Study of Individual Behavior and Social Interactions (Discussion by Bowles, Tao and Winship, Dechter, and Durlauf); GILULA AND HABERMAN Analysis of Categorical Response Profiles by Informative Summaries; GOODMAN AND HOUT Statistical Methods and Graphical Displays for Analyzing How the Association Between Two Qualitative Variables Differs Among Countries, Among Groups, or Over Time. Part II: Some Exploratory Techniques, Simple Models, and Simple Examples. MAGIDSON AND VERMUNT Latent Class Factor and Cluster Models, Bi‑Plots and Related Graphical Displays; SCOTT AND HANDOCK Covariance Models for Latent Structure in Longitudinal Data; WHITE AND HARARY The Cohesiveness of Blocks in Social Networks: Connectivity and Conditional Density; SIJNDERS The Statistical Evaluation of Social Network Dynamics.

The eight chapters in this volume of Sociological Methodology 2001touch on both long‑standing and more recent themes in social statistics. In one way or another, all these themes (as well as others) are discussed at greater length in Chapter 1, "Statistics in Sociology, 1950‑2000: A Selective Review." Dividing this period into three generations based on the type of data structure most typically featured, Adrian Raftery first reviews the analysis of cross‑tabulations, a theme reflected here in the papers by Haberman and Gilula, Goodman and Hout, and Magidson and Vermunt. Next, he reviews statistical methods developed to analyze survey data on individuals‑for example, structural equation models and survival (event‑history) models. Scott and Handcock, who propose new methods for mod­eling longitudinal data, nicely combine a number of themes that have emerged primarily during this second generation. Finally, Raftery turns his attention to the future, discussing limitations of some current work and outlining some developments he thinks important for social statistics during the years ahead. One of the topics discussed is social networks, a theme reflected directly in the article by White and Harary and the piece by Snijders on the analysis of longitudinal network data. Raftery also dis­cusses recent work in spatial statistics, noting that approaches based on random fields have proved productive; Durlauf uses random fields to study social interactions. Raftery also outlines the need for further advances in the analysis of textual and qualitative (not tabular) data, narrative and sequence analysis, and the need for putting simulation modeling on a firmer inferential footing, going so far as to make specific types of proposals on the types of advances and how these might be achieved. Social statistics will blossom if the many leads identified by Raftery are followed, and the next generation of social statisticians should be grateful to him for seeing so clearly some of the steps ahead.

Steven Durlauf's outstanding and forward‑looking work on what is perhaps the most important issue in sociology‑the interdependence between group and individual‑is just the kind of advance Raftery would appreciate (even if it is not anticipated in his essay). Although economists have long recognized the interdependence of preferences, such considerations have not been incorporated into standard economic models of individual decision making (in part because it is hard to do). Sociologists, who often use these models, nevertheless complain that economists treat individuals as atomized actors. If ever such complaints were legitimate, Durlauf's paper amply evidences that it is time to cease and desist. To be sure, one might not like the way Durlauf incorporates social interactions into the models he proposes or think that this is the only productive way to do so. The point, however, is that Durlauf has put forth a coherent framework in which such objections can be raised, and the models improved and extended, both in directions Durlauf might imagine and in directions he might not. Furthermore, Durlauf's framework is not merely theoretical. The models he proposes can be estimated and used to inform important sociological issues, such as decisions to have (not have) a child or to move (not move) neighborhoods, or how many years of education to obtain. Use of this framework should also lead future workers to use more interesting and informative data structures than those routinely used in current work.

Due to the potential importance and novelty of Durlauf's work, we believed that further discussion of his paper would be most interesting and useful for Sociological Methodology readers. We are grateful to Samuel Bowles, Lin Tao and Christopher Winship, and Aimee Dechter, for agreeing to serve as discussants and for contributing to this volume. We also thank Steven Durlauf for graciously consenting to this discussion and for responding to the discussants.

Chapters 3 through 5 discuss the modeling of categorical data. Haberman and Gilula consider estimation and description of a vector of dependent categorical variables (for example, binary responses on ten items) using as predictors one or more covariates. When there are many dependent variables and/or categories of the dependent variables and large samples, no nontrivial log‑linear model typically fits the data using X² tests; furthermore, such tests might not always be appropriate. Nevertheless, one wants to know how much the covariates help in predicting the response. This paper discusses appropriate estimation methods for fitting log‑linear models in this case and using them to describe the resulting predictive power of the model and compare competing sets of predictors. Goodman and Hout take up several special cases of the model for comparing differences in association across a third variable that they proposed in their 1998 Sociological Methodology paper. Magidson and Vermunt discuss exploratory latent class models and their relationship to latent class factor models, arguing that the latter often have interpretive and empirical advantages over the former. They also propose using "bi‑plots" to graphically display the results from fitting these models.

Longitudinal data analysis is now a large area of interest, and many different types of models and modeling strategies have been proposed, as nicely documented in the paper by Scott and Handcock. Readers of Sociological Methodology are most likely to be familiar with methods designed to describe the mean of a population (population average analysis) or subpopulation or methods designed to describe the mean of given persons for example, random coefficient models. In the former case, the covariance is a nuisance parameter and one wants to be able to model the mean structure properly without worrying about modeling the covariance structure correctly. In the second case, it is usual to endow individuals with unobserved components that reflect individual differences. Other approaches to modeling longitudinal data discussed in this paper include latent curve modeling and latent class models, in which individuals belong to latent classes with different trajectories. However, these methods do not yield a population average analysis of variation, which is what one would need to assess dispersion in wage profiles, a subject of great interest to labor economists. Scott and Handcock propose new models (proto-spline models) that combine features of the approaches above, enabling them to address such issues.

Chapters 7 and 8 take up themes in network analysis. In the first of these, White and Harary take up the topic of cohesion, which they regard as one component of solidarity (the other being adhesion), arguing that connectivity and conditional density are two measures of cohesion. In Chapter 8, Snijders proposes a continuous‑time Markov chain model for studying the evolution of a network of actors on which a directed relationship is defined. The dynamics are generated by actors maximizing objective functions (which include a random term) by adding new relations and/or dropping old ones. The model is estimated using MCMC methods.

Social Statistics for a Diverse Society. Third Edition by Chava Frankfort-Nachmias, Anna Leon-Guerrero (The Pine Forge Press Series in Research Methods and Statistics: Pine Forge) You may be reading this introduction on the first day or sometime during the first week of your statistics class. You probably have some questions about statistics and concerns about what your course will be like. Math, for­mulas, calculations? Yes, those will be part of your learning experience. However, there is more.

Throughout our text, we emphasize the relevance of statistics in our daily and professional lives. In fact, statistics is such a part of our lives that its importance and uses are often overlooked. How Americans feel about a variety of political and social topics‑safety in schools, gun control, abor­tion, affirmative action, or our president‑are measured by surveys and polls and reported daily by the news media. Consider how news programs are already predicting the Republican and Democratic front runners for the next presidential election based on early poll results. The latest from a health care study on women was just reported on a morning talk show. And that outfit you just purchased‑it didn't go unnoticed. The study of con­sumer trends, specifically focusing on teens and young adults, helps deter­mine commercial programming, product advertising and placement, and ultimately, consumer spending.

Statistics is not just a part of our lives in the form of news bits or infor­mation. And it isn't just numbers either. Throughout this book, we encour­age you to move beyond being just a consumer of statistics and begin to recognize and utilize the many ways that statistics can increase our under­standing of our world. As social scientists, we know that statistics can be a valuable set of tools to help us analyze and understand the differences in our American society and the world. We use statistics to track demographic trends, to assess differences among groups in society, and to make an im­pact on social policy and social change. Statistics can help us gain insight into real‑life problems that affect our lives.

The student will be expected to read and interpret statistical information presented by others in professional and scholarly publications, in the workplace, and in the popular media. This book will help you understand the concepts behind the statistics so that you will be able to assess the circumstances in which certain statistics should and should not be used.

Our second goal is to demonstrate that substance and statistical techniques are truly related in social science research. A special quality of this book is its integration of statistical techniques with substantive issues of particular relevance in the social sciences. Your learning will not be limited to statistical calculations and formulas. Rather, you will become proficient in statistical techniques while learning about social differences and inequality through numerous substantive examples and real‑world data applications. Because the world we live in is characterized by a growing diversity‑where personal and social realities are increasingly shaped by race, class, gender, and other categories of experience‑this book teaches you basic statistics while incorporating social science research related to the dynamic interplay of social variables.

Many of you may lack substantial math background, and some of you may suffer from "math anxiety syndrome." This anxiety often leads to a less‑than‑optimum learning environment, with students trying to memorize every detail of a statistical procedure rather than attempting to understand the general concept involved. Hence, our third goal is to address math anxiety by using straightforward prose to explain statistical concepts and by emphasizing intuition, logic, and common sense over rote memorization and derivation of formulas.

The three learning goals we emphasize are accomplished through a variety of specific and distinctive features throughout this book.

A Close Link Between the Practice of Statistics, Important Social Issues, and RealWorld Examples A special quality of this book is its integration of statistical technique with pressing social issues of particular concern to society and social science. We emphasize how the conduct of social science is the constant interplay between social concerns and methods of inquiry. In addition, the examples throughout the book‑most taken from news stories, government reports, scholarly research, and the NORC General Social Survey‑are formulated to emphasize to students like you that we live in a world in which statistical arguments are common. Statistical concepts and procedures are illustrated with real data and research, providing a clear sense of how questions about important social issues can be studied with various statistical techniques. This focus on the richness of social differences within society is manifested in the application of statistical tools to examine how race, class, gender, and other categories of experience shape our social world and explain social behavior.

Reading the Research Literature In your student career and in the workplace, you may be expected to read and interpret statistical information presented by others in professional and scholarly publications. The statistical analyses presented in these publications are a good deal more complex than most class and textbook presentations. To guide you in reading and interpreting research reports written by social scientists, most chapters include a section presenting excerpts of published research reports utilizing the statistical concepts under discussion.

Integration and Review Chapters Two special review chapters are included. The first is Chapter 9, a review of descriptive statistical methods (Chapters 2‑8), and the second, Chapter 15, reviews inferential statistics (Chapters 1014). These review chapters provide an overview of the interconnectedness of the statistical concepts in this book and help test your abilities to cumulatively apply the knowledge from previous chapters. Both chapters include flowcharts that summarize the systematic approach utilized in the selection of statistical techniques as well as exercises that require the use of several different procedures.

Tools to Promote Effective Study Each chapter closes with a list of main points and key terms discussed in that chapter. Boxed definitions of the key terms also appear in the body of the chapter, as do learning checks keyed to the most important points. Key terms are also clearly defined and explained in the index/glossary, another special feature in our book. Answers to all the odd‑numbered problems in the text are included in the back of the book. Complete step‑by‑step solutions are in the manual for instructors, available from the publisher upon adoption of the text.

Emphasis on Computing SPSS® for Windows© is used throughout the book, although the use of computers is not required to learn from the text. Real data are used to motivate and make concrete the coverage of statistical topics. These data, from the General Social Survey, are included on a CD packaged with every copy of the text. At the end of each chapter, we feature a demonstration of a related SPSS procedure, along with a set of exercises.

Clearer and more concise presentation of topics. We have carefully edited the discussion of statistical procedures and concepts, reducing the redundancy of statistical procedures and clarifying examples, while at the same time preserving the book's easily understood style. For examples, please refer to our expanded discussion on standard deviation in Chapter 5 and our discussion of control and intervening variables in Chapter 6.

Revisions to Chapter 13, "Testing Hypotheses about Two Samples." We reintroduced, from the first edition, the detailed discussion of the principles of hypothesis testing, including the Z test for one sample. In addition, we have added a discussion of the t statistic in the context of cane‑sample tests. Although one‑sample tests are seldom used in practice, we introduce them here as a pedagogical tool to clarify statistical inference before moving to the more complicated concept of the two‑sample test. For simplification, the discussion of unequal variances has been boxed and shortened considerably.

Supplemental electronic chapters. Instead of adding additional printed chapters (and pages) to our text, we provide two additional chapters on the accompanying CD and on the Pine Forge web site to increase flexibility for instructors and their students. The first electronic chapter is an expanded version of Chapter 8, "Bivarzate Regression and Correlation." The electronic version includes new sections on inference in regression, analysis of variance, and a brief overview of multiple regression techniques. This chapter can be used in conjunction with Chapter 8 in the textbook or used alone. The second electronic chapter is a stand‑alone chapter on analysis of variance (ANOVA). The chapter begins with a detailed computational example, highlights two SPSS ANOVA applications, and reviews two examples from research literature. As in our text chapters, each electronic chapter concludes with an SPSS demonstration, SPSS exercises, and review exercises.

Real‑world examples and exercises. A hallmark of our first two editions was the extensive use of real data from a variety of sources for chapter illustrations and exercises. Throughout the third edition, we have updated the majority of exercises and examples based on General Social Survey and U.S. Census data.

 SPSS Version 11.0. Packaged with this text, on an optional basis, is SPSS Student Version 11.0. SPSS demonstrations and exercises have been updated, using Version 11.0 format. An appendix on how to use a statistical package has also been updated to highlight SPSS 11.1) features and has been moved to the accompanying CD.

 

Persuasion: Theory and Research by Daniel J. O'Keefe (Sage) This comprehensive text provides a thorough and critical treatment of persuasion theory and research from a social science perspective. O'Keefe includes a discussion of research on the production of persuasive messages as well as more traditional research on the study of message effects. The new edition contains more coverage of the theory of reasoned action, a new chapter on functional approaches to attitude, a new chapter on behavioral change, and new material on persuasive campaigns. The research citations and examples are also updated.

Persuasion surveys social‑scientific theory and research concerning per­suasive communication. The relevant work, as will become apparent, is scattered across the academic landscape‑in communication, psychol­ogy, advertising, marketing, political science, law, and so on. Although the breadth and depth of this literature rule out a completely comprehensive and detailed treatment, the main lines of work are at least sketched here.

This introductory chapter begins, naturally enough, with a discussion of the concept of persuasion. But because social‑scientific treatments of per­suasion have closely linked persuasion and attitude change, the concept of attitude is discussed as well, some common attitude assessment procedures are described, and the relationship of attitudes and behavior is considered; a concluding section discusses the assessment of persuasive effects.

Contents: Preface  1. Persuasion, Attitudes, and Actions  2. Functional Approaches to Attitude  3. Belief-Based Models of Attitude  4. Cognitive Dissonance Theory  5. Theories of Behavioral Intention  6. Elaboration Likelihood Model  7. The Study of Persuasive Effects  8. Source Factors  9. Message Factors  10. Receiver and Context Factors References Author Index Subject Index About the Author

Data Analysis Using SPSS for Windows Version 8 to 10: A Beginner's Guide by Jeremy Foster (Sage Publications) is a new edition of this best‑selling introductory book to cover the latest SPSS versions 8 to 10.

This book is designed to teach beginners how to use SPSS for Windows, the most widely used computer package for analyzing quantitative data. Written in a clear, readable and non‑technical style, the author explains the basics of SPSS including the input of data, data manipulation, descriptive analyses and inferential techniques, including:
Ø      creating, using and merging data files
Ø      creating and printing graphs and charts
Ø      parametric tests including t‑tests, ANOVA, GLM
Ø      correlation, regression and factor analysi
Ø      non-parametric tests and chi square
Ø      reliability
Ø      obtaining neat print outs and tables
Ø      a Disk containing example data files, syntax files, output files and Excel spreadsheets.

The book begins with a brief introduction to statistical analysis ‑ starting with basic definitions through to interpretation and significance of the results. The author shows in a simple step‑by‑step method how to set up SPSS data files in order to run your analysis, as well as how to graph and display data. The book shows the reader how to use SPSS for all the main statistical approaches you would expect to find in an introductory statistics course.

Also included are detailed worked examples‑with SPSS output and screen shots, along with diagrams and a useful decision chart to select the appropriate statistical test. The differences between versions 8 to 10 are also listed to help those upgrading to the latest version.

This book will prove invaluable to all those students and researchers who need to learn to use SPSS effectively in their research.

Regression Basics by Leo H. Kahane (Sage) Although many people have PCs with software capable of performing regression techniques, only a few know how to capitalize on the flexibility and wide application of regression analysis. This book shows readers how to get the most from regression by providing a friendly, non‑technical introduction to the subject. Accessible to anyone with an introductory statistics background, the book begins with the simplest, two‑variable linear model and gradually builds towards models of more complexity, such as multivariate regression. Kahane uses three engaging examples to illustrate regression concepts. These examples show the creative way in which regression analysis can be used to determine why some professional sports players earn higher salaries than others, the factors that affect voting patterns in presidential elections, and the factors that explain the difference in abortion rates across the United States. Data for these examples are provided in an appendix so that readers can have tangible, hands‑on experience in performing linear regression analysis.

Additional outstanding features:

• End-of-chapter problems based on two additional completely worked-out examples: the effects of education and other factors on a person's salary and the factors that determine automobile prices
• Data for end-of-chapter problems provided at the end of the book
• Complete solutions to all problems so readers can check their own work
• Glossary of terms so the reader can become more confident in using the language of regression.

The book is a comprehensive, thought‑provoking resource for Ph.D. students, academics, and professionals working to minimize or eliminate the sources of stress in the workplace.

Headline 3

insert content here

WT Main | About WT | Review Links | Contact | Review Sources | Search