Exploring Social Issues: Using SPSS for Windows


Joseph F. Healey, John Boli, Earl R. Babbie & Fred Halley

  • Citations
  • Add to My List
  • Text Size

  • Chapters
  • Front Matter
  • Back Matter
  • Subject Index
  • Copyright

    View Copyright Page

    About the Authors

    Joseph F. Healey, PhD, is Professor of Sociology at Christopher Newport University. He is author of Statistics: A Tool for Social Research (2009, 8th ed.), and Race, Ethnicity, Gender and Class (2009, 5th ed.). He is coauthor of Sociology for a New Century (with York Bradshaw, 2001) and of Exploring Social Issues Using SPSS For Windows (with Earl Babbie, Fred Halley, and John Boli). He received his AB and MA degrees from The College of William and Mary (sociology and anthropology) and PhD from the University of Virginia (sociology and anthropology). In his spare time, he plays and records music for hammer dulcimer, banjo, and concertina.

    John Boli is Professor of Sociology at Emory University. He has held visiting appointments at Lund University, the University of Copenhagen, and the University of Santa Clara. A native Californian and graduate of Stanford University, he has published extensively on globalization, world culture, international nongovernmental organizations, education, citizenship, and state power and authority in the world polity. His books include World Culture: Origins and Consequences (2005) and The Globalization Reader (2008, 3rd ed.), both with Frank Lechner, as well as Constructing World Culture: International Nongovernmental Organizations Since 1875 (1999), with George Thomas. His current research concerns the origins and development of world culture and transnational structuration since the 12th century, with a special focus on Christendom and the Roman Catholic Church. Married with three children, he has lived for 8 years in Sweden, his wife's native country.

    Earl R. Babbie was born in Detroit, Michigan, in 1938, although he chose to return to Vermont 3 months later, growing up there and in New Hampshire. In 1956, he set off for Harvard Yard, where he spent the next 4 years learning more than he initially planned. After 3 years with the U.S. Marine Corps, mostly in Asia, he began graduate studies at the University of California, Berkeley. He received his PhD from Berkeley in 1969. He taught sociology at the University of Hawaii from 1968 through 1979, took time off from teaching and research to write full time for 8 years, and then joined the faculty at Chapman University in Southern California in 1987. He retired from teaching in 2006 and moved to Hot Springs Village, Arkansas, the next year. Although an author of research articles and monographs, he is best known for the many textbooks he has written, which have been widely adopted in colleges throughout the United States and the world. He also has been active in the American Sociological Association throughout his career for 40 years and served on the ASA's executive committee. He is also past president of the Pacific Sociological Association and California Sociological Association. He is married to Suzanne Babbie, a joyful partner in all aspects of his life, and he has a son, Aaron, who would make any parent proud. As partial proof, Aaron and his wife, Ara, produced the world's two greatest grandchildren: Evelyn and Henry.

    Fred Halley, State University of New York College at Brockport, has been developing computer-based tools for teaching social science since 1970. He has served as a collegewide social science computer consultant, directed Brockport's Institute for Social Research, and now directs the college's Data Analysis Laboratory.


    This workbook is a “hands-on” introduction to the craft of social research. It is intended for use in introductory sociology courses and may be combined with most of the standard textbooks in the field. Students are involved in the process of social research from the first chapter: Indeed, the student is the principal researcher. Most exercises are open ended and the student is required to frame hypotheses, choose variables, and interpret results. The text provides abundant explanations, examples, and hints, but ultimately this is a “self-writing” textbook, and no two students will complete it in exactly the same way. Students take an active role and test their own ideas about topics such as divorce, abortion, crime, inequality, and prejudice.

    Links to the sociological literature are established in two ways. First, students are always urged to use their textbook and other course materials to develop hypotheses and interpret results. Second, about half the chapters include exercises drawn from sociological research, giving students the opportunity to test or expand on ideas and theories from the literature. Appropriate, accessible sources are cited from the social science literature to strengthen the link with the professional research literature.

    This text uses SPSS for Windows, Student Version to analyze the 2006 General Social Survey (GSS). (The standard or professional version of SPSS for Windows may also be used.) It includes an overview of the research process, an introduction to SPSS for Windows, and a description of the 2006 GSS. The package is self-contained (no additional documentation is needed), and students are guided step-by-step through all exercises. No previous experience with computers, Windows, SPSS, statistics, or social research is required to use this text successfully. The data sets used in the text (two versions of the 2006 GSS and a version of the 1972 GSS) can be downloaded from the Web site for this text at http://www.pineforge.com/textbooksProdSampleMaterials.nav?prodId=Book232268. Click on the name of the data file and save the file to your hard drive or to some other storage medium (such as a flash drive).

    Chapters are arranged in an order that roughly parallels the organization of most introductory sociology texts. Following the first chapter, we cover the research process, culture, socialization, deviance, inequality, and social institutions. Issues and problems that are of interest to students are stressed throughout.

    Each chapter includes explanations of basic research principles and techniques, exercises, research reports, and end-of-chapter independent projects. The exercises are basically demonstrations, and students are expected to “follow along” on their own. In the research reports, students apply what they learned in the explanatory sections and exercises. The research reports follow a standardized, fill-in-the-blank format for presenting and analyzing results, but space is always left for students to summarize their results in their own words. This format is designed to ease the often burdensome chore of deciding “what to say” about the results (and will also ease the instructor's burden of checking the reports).

    End-of-chapter projects are of two types. Independent Projects afford an opportunity to further pursue projects begun in the chapter. Comparative Analyses provide historical depth and analyze trends over time by comparing 2006 results and patterns with data from the 1972 GSS.

    Data analysis begins with univariate frequency distributions and percentages and builds through descriptive statistics, charts and graphs, bivariate tables, tests of significance and measures of association, scatterplots, and Pearson's r. Students are introduced to multivariate analysis in the form of controlling for a third variable in bivariate tabular analysis. Statistical presentations stress the simpler techniques and more “intuitive” (vs. mathematical) understandings, as is appropriate at the introductory level. Most chapters rely on bivariate tables for statistical analysis, an approach that is consistent with the type of data contained in the GSS. Each of the first 10 chapters introduces a new technique, SPSS procedure, or statistic. The last two chapters apply these techniques to new material and may be covered in any order. To provide maximum flexibility, there is considerable choice of topics within each chapter and in the end-of-chapter exercises.


    We would like to thank the many people who have been instrumental in making this book a reality. We thank Steve Rutter, formerly of Pine Forge Press, for his guidance in the early stages of this project and Ben Penner for his help in preparing this edition of the text. Jamie Perkins of Pine Forge Press was instrumental in bringing the project to completion and we thank him for his assistance.

    We would also like to thank the many reviewers who helped us along the way: Marybeth Ayella, St. Joseph's University; Ray Daville, Stephen F. Austin University; David Karp, University of Washington; Peter Lehman, University of Southern Maine; Joe Lengermann, University of Maryland, College Park; Brad Lyman, Baltimore City Community College; Edgar (Ted) Mills, University of Connecticut; Elizabeth Nelson, California State University, Fresno; and Assata Zerai, Syracuse University.

    Reviewers for This Edition
    • Rae Banks, Syracuse University
    • Carolyn Bond, Westchester University
    • Kevin Demmitt, Clayton State University
    • Whitney Garcia, Towson University
    • Richard Harris, University of Texas, San Antonio
    • Jason Leiker, Utah State University
    • Mario Renzi, Hiram College
  • Appendix A: Codebooks for All Data Sets

    Codebook for the GSS-2006-Tabular Data Set

    NOTE: The wordings of some questions have been edited and shortened.

    Codebook for the GSS-2006-Numerical Data Set

    NOTE: The wordings of some questions have been edited and shortened.

    Appendix B: SPSS Commands Used in This Text

    NOTE: The procedures (Crosstabs, etc.) that you run will be performed only on the selected cases. Once you have finished, be sure to return to the Select Cases dialog box and click Reset to restore the sample to its original composition.

    Answers to Selected Exercises

    Answers to some items in selected research reports are provided so that you can check your work and make sure that you are using SPSS correctly. In general, answers are provided for reports that use an SPSS procedure for the first time. For research reports in which students have a choice of variables, results for the first variable mentioned in the text are shown.

    Research Report 1.1

    1. (Religious affiliation or RELIG) For the 2006 GSS sample, the most common religious affiliation was Protestant with 51.7 % of the sample. The second most common was Catholic with 25.5%. The least common religious affiliation was Jewish with 2.0% of the sample. About 16.2% of the respondents have no religious affiliation (“None”).

    Research Report 2.2

    1. In 2006, the percentage of all respondents who have been divorced is 24.9%.

    If you chose degree as your independent variable:

    Research Report 3.1

    1. About 48% of the sample are married, and about 8% are widowed.

    2. The largest social class is the working class, and the smallest is the upper class.

    Research Report 3.2

    1. The sample has completed an average of 13.27 years of schooling. The most educated respondents completed 20 years, and the least educated finished 0 years. The range of the scores was 20–0 or 20.

    Research Report 3.3

    1. For years of education (educ), the highest peak (most common score) was 12. The lowest point of the chart was at 1 year of education. Most cases seemed to be grouped around 12 years of schooling.

    Research Report 4.2

    1. A total of 35.2% of high-income respondents are strong in their religious faith versus 36.1% of middle-income and 34.1% of low-income respondents. Religious intensity is strongest for middle-(low/middle/high) income respondents, and there is not (is/is not) a relationship between these two variables. This table does not (does/does not) support the deprivation theory.

    Research Report 4.3

    1. Record the percentage of respondents who said “yes” to POSTLIFE for each independent variable. Record the significance (“p” or “Asymp. Sig. 2-sided”) of chi-square and the value of Cramer's V in the spaces provided.

    2. For INCOME: The column percentages do (do/do not) change, so there is (is/is not) a relationship between INCOME and POSTLIFE. The significance of chi-square is more than (less than/more than) .05, so the relationship between INCOME and POSTLIFE is not (is/is not) statistically significant. The value of Cramer's V for INCOME and POSTLIFE is .06, so this is a weak (weak/moderate/strong) relationship.

    Research Report 5.4

    Summarize and interpret these results:

    The column percentages do change, so there is a relationship between these variables. The significance of chi-square for this relationship is .000, and the value for p is less than 0.05, so this relationship is statistically significant. The value of Cramer's V is 0.15, so this is a moderate relationship.

    Research Report 6.5

    4. Summarize and interpret these results:

    • The column percentages do change, so there is a relationship between these variables.
    • The value of chi-square for this relationship is 6.555, and the value for p is more than 0.05, so this relationship is not statistically significant.
    • The value of gamma is 0.10, so this is a weak relationship.
    • The sign of gamma is positive. This means that support for spanking decreases as social class standing increases.

    NOTE: Be careful when interpreting direction for ordinal-level variables. In this case, SPANKING is scored so that support (“agree”) is a lower score than opposition (“disagree”). The independent variable, CLASS, is scored with higher scores indicating higher class. The positive sign of gamma tells us that opposition to spanking (the higher score of “disagree”) increases as class increases.

    Research Report 7.2

    Fear (Percent “Yes”) by Income Level and Race or Ethnicity

    Use the guidelines in Exhibit 7.3 to help you describe the effect of controlling for RACEETH.

    • For whites: FEAR decreases as INCOME increases. The gamma of .41 indicates a moderate relationship.
    • For blacks: FEAR decreases as INCOME increases. The gamma of .08 indicates a weak relationship.
    • For Hispanics: FEAR decreases as INCOME increases. The gamma of .35 indicates a moderate relationship.
    Research Report 8.1

    1. The relationship between income and wealth is positive. This appears to be a moderate relationship.

    2. Pearson's r for the relationship between WEALTH and INCOME06 was .60. The relationship between these variables is positive and moderate to strong. The relationship is statistically significant. The value of Pearson's r squared times 100 is 0.36. This means that income explains 36% of the variance in wealth.

    Research Report 9.1

    For years of education (EDUC):

    1. For lower years of education (less than 12), men are generally the same as women.

    2. For high school (12 years), men are generally the same as women.

    3. For higher years of education (especially for a college degree or 16 years or education), men are generally the same as women.

    4. For the very highest levels of education (above 16), men are generally higher than women.

    5. This chart indicates that men and women are equal in level of education.

    Research Report 9.2

    Complete the following summary table for the results of the T Test.

    1. Men's average occupational prestige is the same as women's, and the difference is not significant.

    Research Report 10.2

    1. Complete the following summary tables for the results of the T Test. This table is based on Exhibit 10.2 but includes only the most essential information.

    Research Report 11.1

    3. Now, analyze the crosstabulation of marital and hhtype.

    a. For “married” respondents, 55.0% are childless couples, whereas 43.3% are nuclear families.

    b. The household arrangement for the great majority of the “widowed” respondents is single adult.

    c. Most “divorced” respondents are living as single adults, but 17.6% are single parents and 9.3% are cohabiting.

    d. Most “separated” respondents are living as single adults but 24.2% are single parents and 9.1% are cohabiting.

    e. Most people who have “never married” are living as single adults, but some (14.6%) are cohabiting and others (13.2%) are single parents.

    Research Report 12.4

    1. Compare and contrast the percentages of each group voting for Bush for each class. High-income whites had the highest percentage of support for Bush and middle-income Asian Americans (or, because there are so few Asian Americans in the sample) middle-income blacks had the lowest.

    Glossary of Key Concepts

    • Bar chart: A type of graph that represents the distribution of scores of a variable by means of bars. The horizontal axis presents the scores and the vertical axis represents frequencies or percentages. The higher the bar, the more common the score.
    • Bivariate tables: Tables that show the relationship between two variables.
    • Causal relationships: There is a causal relationship between variables if the presence of one variable makes the other occur or if changes in the value or score of one variable result in changes in the value or score of the other variable.
    • Cells: In a bivariate (or Crosstab) table, each combination of row and column is a cell. Each cell contains the number and / or percentage of cases that have a specific combination of scores on the two variable in the table (e.g., males born in the South, Protestant Republicans).
    • Chi-square: A statistic computed on bivariate tables that tells us the probability that the pattern of cell frequencies in the table occurred by random chance alone. In SPSS output, we look especially at the value of p or Asymp. Sig. (2-sided). Values less than .05 are taken to be statistically significant.
    • Column percentages: In a bivariate table, percentages are computed so that they total 100% within each column.
    • Column: The vertical dimension of the SPSS Data Editor window. Each column contains the scores of all cases on a particular variable.
    • Control variable: A “third variable.” In the multivariate technique called elaboration, we observe how the relationship between an independent and dependent variable changes across the various categories of the control variable.
    • Cramer's V: A statistic computed on bivariate tables that tell us the strength of a relationship. Use Exhibit 4.6 to judge the strength of a relationship using Cramer's V.
    • Crime: Behavior that violates the criminal law.
    • Crosstab tables: See Bivariate tables.
    • Culture: All aspects of a society's way of life. Culture includes values, language, rules for behavior, and other elements that are contained in the way of life and heritage of the society.
    • Data Editor: The window of SPSS that contains the scores of all cases on all variables.
    • Dependent variable: The variable in a relationship that is assumed to be the effect or result.
    • Deviance: Behavior that violates any cultural norm.
    • Epsilon: The difference between the largest and smallest column percentages in any row of a bivariate table. Epsilon shows the strength of the relationship between two variables: The higher its value, the stronger the relationship.
    • File: An organized collection of information saved in the memory of a computer (or other storage medium).
    • Frequency distribution: A table that presents the number of times each value or score of a variable occurred.
    • Gamma: A measure of association that provides two pieces of information about a bivariate relationship between two numerical variables: the strength of the relationship and the direction of the relationship.
    • Hypothesis: An explanation that is related to a theory but is more specific and exact.
    • Income: One dimension of inequality. Income is usually defined as “take-home” pay as distinct from wealth.
    • Independent variable: The variable in a relationship that is assumed to be the cause.
    • Inequality: The uneven distribution of valued goods and services across a population.
    • Interval-ratio variable: A variable with mathematical scores. Scores can be used in any mathematical procedure (addition, multiplication, etc.).
    • Line chart: A type of graph that represents the distribution of scores of a variable by means of a line. The horizontal axis presents the scores and the vertical axis represents frequencies or percentages. Higher points of the line represent more common scores.
    • Mean: The average score of a variable.
    • Measures of association: A class of statistics that includes Cramer's V. Measures of association indicate the strength and (sometimes) the direction of a relationship between two variables. See Chapter 4 for more information.
    • Multivariate analysis: Statistical techniques that allow us to investigate relationships between more than two variables simultaneously.
    • Negative relationship: In a negative relationship between two numerical variables, the variables change in the opposite direction. As one decreases, the other increases and high scores on one variable are associated with low scores on the other variable.
    • Nominal variable: A type of variable that is non-numerical. The scores of the variable are labels and have no mathematical qualities.
    • Norms: Standards or cultural expectations for behavior.
    • Nuclear family: A type of family in which two generations (parents and children) live together.
    • Operationalization: The process of deciding how to measure abstract concepts. During operationalization, we match specific variables to the concepts in our theories and hypotheses.
    • Ordinal: A type of variable whose scores have some limited mathematical qualities. Scores on ordinal variables can be ranked as higher or lower.
    • Pearson's r: A measure of association that shows the strength and direction of a bivariate relationship. Best used with numerical variables with many scores.
    • Population: The entire group of cases in which the researcher is interested. In this text, the population is all adult Americans.
    • Positive relationship: In a positive relationship between two numerical variables, the variables change in the same direction. As one decreases, the other decreases as well and high scores on one variable are associated with high scores on the other variable.
    • Prejudice: Negative attitudes, feelings, and stereotypes about other groups.
    • Prestige: Respect or honor. Like income and wealth, prestige is unequally distributed among people.
    • Probabilistic causal relationship: A causal relationship stated in terms of probabilities or tendencies.
    • Range: The distance from the highest to the lowest score of a variable.
    • Regression line: On a scatterplot, the straight line that touches all points or comes as close to doing so as possible.
    • Representative: A characteristic of carefully chosen, randomly selected samples. A representative sample has the same characteristics (within a small error margin) as the population from which it was selected.
    • Research: The careful gathering and evaluation of empirical evidence.
    • Row: The horizontal dimension of the SPSS Data Editor window. Each row contains the scores of a particular case on all variables.
    • Sample: A subgroup of a population. In this text, the sample is about 1,490 adult Americans who completed the General Social Survey in 2006.
    • Scatterplots: A graph showing the scores of cases on two variables. Best used with numerical variables with many scores.
    • Sexism: The belief that women are inferior to men.
    • Socialization: The process by which people learn the culture of their society. SPSS Viewer: The window of SPSS that contains all output.
    • Statistical significance: Results or relationships that are unlikely to have been produced by random chance are said to be statistically significant. The .05 level is the conventional indicator of statistical significance: results that have a probability of less than .05 are statistically significant.
    • t-Test: A test of significance that tells us the probability that a difference between sample means also exists in the populations from which the samples were selected. Generally, if the probability of getting the difference in sample means is less than .05, we conclude that the difference is significant (i.e., it also exists in the population).
    • Theory: A generalized explanation of a relationship.
    • Validity: A criteria we use to judge the adequacy of our operationalizations. A variable is a valid indicator if it actually measures the concept intended.
    • Values: Cultural ideas about what is moral and desirable.
    • Wealth: A dimension of inequality. Wealth is the total value of all assets as distinct from income.

    • Loading...
Back to Top