Bibliographic record and links to related information available from the Library of Congress catalog.

**Note:** Contents data are machine generated based on pre-publication provided by the publisher. Contents may have variations from the printed book or be incomplete or contain other coding.

CONTENTS List of Tables, Figures, and Exhibits xv Preface xxxi For instructors xxxiv Acknowledgments xxxv The Author xxxvii Introduction 1 Overview of Chapters 2 1. Cross-tabulations 7 What this Chapter Is about 7 Introduction to the Book via a Concrete Example 7 Cross-tabulations 19 In Which Direction to Percentage the Table? 24 Control Variables 25 What this Chapter Has Shown 36 2. More on Tables 37 What this Chapter Is about 37 The Logic of Elaboration 37 Suppressor Variables 41 Additive and Interaction Effects 43 Direct Standardization 47 Example 1: Religiosity by Militancy among U.S. Urban Blacks 47 Example 2: Belief That Humans Evolved from Animals (Direct Standardization With Two or More Control Variables) 51 Example 3: Occupational Status by Race in South Africa 61 Example 4: Level of Literacy by Urban Vs. Rural Residence in China 65 A Final Note, on Statistical Controls vs. Experiments 67 What this Chapter Has Shown 71 3. Still More on Tables 73 What this Chapter Is about 73 Reorganizing Tables in Order to Extract New Information .73 Collapsing Dimensions 74 Transforming Property Space 75 When to Percentage a Table ¿Backwards¿ 78 Cross-tabulations in Which the Dependent Variable Is Represented by a Mean 80 Inferences from Information on Missing Data 85 Still Another Way of Presenting the Same Data 85 Index of Dissimilarity 89 Writing about Cross-tabulations 92 What this Chapter Has Shown 96 4. On the Manipulation of Data by Computer 98 What this Chapter Is about 98 Introduction 98 How Files Are Organized 100 A Digression on Card Desks and Card Image Files 105 Transforming Data 110 Recoding 110 Arithmetic Transformations 114 Contingency Transformations 115 Missing Data 116 Analyzing Surveys with Missing Data 119 What this Chapter Has Shown 121 Appendix 4.A: Doing Analysis Using Stata 10.0 123 Tips on Doing Analysis Using Stata 123 Do Everything with -do- Files 123 Build in Extensive Checks of Your Work 127 Document Your -do- File Exhaustively 127 Include ¿Side¿ Computations in Your -do- File 128 Rerun Your -do- File as a Final Check 129 Make Active Use of the Stata Manuals 129 Some Particularly Useful Stata 10.0 Commands 129 Appendix 4.B: The Stata -for- Command 134 5. Introduction to Correlation and Regression (Ordinary Least Squares) 135 What this Chapter Is about 135 Introduction 135 Quantifying the Size of a Relationship: Regression Analysis 136 Assessing the Strength of a Relationship: Correlation Analysis 139 The Relationship Between Correlation and Regression Coefficients 142 Factors Affecting the Size of Correlation (And Regression) Coefficients 143 Outliers and Leverage Points 143 Truncation 145 Regression Toward the Mean 147 Aggregation 149 Correlation Ratios 150 Testing Linearity 153 What this Chapter Has Shown 154 6. Introduction to Multiple Correlation and Regression (Ordinary Least Squares) 155 What This Chapter Is About 155 Introduction 155 Metric Regression Coefficients 158 Testing the Significance of Individual Coefficients 160 Standardized Coefficients 162 Coefficient of Determination (R2) 166 Standard Error of Estimate (Root MSE) 169 A Worked Example: the Determinants of Literacy in China 170 Graphic Representation of Results 178 Dummy Variables 181 A Strategy for Comparisons Across Groups 185 Re-expressing Variables as Deviations from Their Means 196 Testing Additional Hypotheses: Constraining Coefficients to Zero or to Equality 197 A Bayesian Alternative for Comparing Models 198 Independent Validation 202 What this Chapter Has Shown 204 7. Multiple Regression Tricks: Techniques for Handling Special Analytic Problems 206 What this Chapter Is about 206 Non-linear Transformations 206 Curvilinear Relationships: Age and Income 207 Semi-log Transformations: Income 209 Mobility Effects 214 Testing the Equality of Coefficients 215 Trend Analysis: Testing the Assumption of Linearity 218 Predicting Variation in Gender Role Attitudes over Time: A Worked Example 220 Linear Splines 223 A Worked Example: Trends in Educational Attainment over Time in the U.S. 224 A Second Worked Example, with a Discontinuity: Quality of Education in China Before, During, and after the Cultural Revolution 229 Expressing Coefficients as Deviations from the Grand Mean (Multiple Classification Analysis) 235 Other Ways of Representing Dummy Variables 239 Effect Coding 243 Contrast Coding 244 Sequential Coefficients 246 Decomposing the Difference Between Two Means 247 A Worked Example: Factors Affecting Racial Differences in Educational Attainment 249 Additional Reading on Decomposing the Difference Between Means 256 What this Chapter Has Shown .256 9. Multiple Imputation of Missing Data 258 What this Chapter Is about 258 Introduction 258 Casewise Deletion 260 Weighted Casewise Deletion 262 Mean Substitution 262 Hotdeck Imputation 264 Full Bayesian Multiple Imputation 265 A Worked Example: the Effect of Cultural Capital on Educational Attainment in Russia 268 Creating the Substantive Model 269 Creating the Imputation Model 273 Comparing Casewise Deletion and Multiple Imputation Results 275 What this Chapter Has Shown 277 9. Sample Design and Survey Estimation 278 What this Chapter Is about 278 Survey Samples 278 Simple Random Samples 279 Multistage Samples 281 Population Register Samples 289 Household Samples 290 Random Walk Samples 294 Design Effects 296 Stratifying to Offset the Effect of Clustering 301 Weighting 303 Survey Estimation Using Stata 307 Literacy in China 308 Meff 311 Deff 312 Analysis of Subpopulations: Effect of Education and Race on Income among Women 316 Combining GSS Datasets for Multiple Years 320 Conclusion 322 What this Chapter Has Shown 323 10. Regression Diagnostics 324 What this Chapter Is about 324 Introduction 324 A Worked Example: Societal Differences in Status Attainment Processes 328 Preliminaries 329 Leverage 331 Outliers 331 Influence 332 Plots for Assessing Influence 332 Added-variable Plots 333 Residual-versus-fitted Plots and Formal Tests for Patterns in the Data. 334 Robust Regression 338 Bootstrapping Standard Errors 340 What this Chapter Has Shown 343 11. Scale Construction 344 What this Chapter Is about 344 Introduction 344 Validity 345 Reliability 347 Scale Construction 351 Additive Scaling 351 Factor-based Scaling 353 A Worked Example: Religiosity and Abortion Attitudes (Again) 357 Seemingly Unrelated Regression 366 Effect-proportional Scaling 367 Errors-in-variables Regression 374 What this Chapter Has Shown 377 12. Log-linear Analysis 379 What this Chapter Is about 379 Introduction 379 Choosing a Preferred Model 382 Model Selection Based on Goodness-of-fit 383 Theory-based Model Selection 390 Effect Parameters 391 Another Worked Example: Anti-communist Sentiment 393 Analysis Strategy 394 Implementation 395 Doing Log-linear Analysis with Polytomous Variable 399 Doing Log-linear Analysis with Individual Level Data 401 More Parsimonious Models 401 Topological, or Levels, Models 407 Quasi-independence Models 408 Quasi-symmetry Models 409 Crossings Models 410 Uniform Association Models 412 Linear-by-linear Association Models 413 Row-effects (And Column-effects) Models 414 Row-and-column-effects Model I 417 Row-and-column-effects Model Ii (The RC or Log-multiplicative Model) 418 Extensions 423 A Bibliographic Note 423 What this Chapter Has Shown 425 Appendix 12.A: Derivation of the Effect Parameters 426 Appendix 12.B: Introduction to Maximum Likelihood Estimation 429 Mean of a Normal Distribution 430 Log-linear Parameters 432 13. Binomial Logistic Regression 434 What this Chapter Is about 434 Introduction 434 Relation to Log-linear Analysis 436 A Worked Logistic Regression Example: Predicting Prevalence of Armed Threats 438 A Second Worked Example: Schooling Progression Ratios in Japan 454 A Third Worked Example (Discrete-time Hazard-rate Models): Age at First Marriage 461 A Fourth Worked Example (Case-control Models): Who Became a Member Of the Nomenklatura in Russia? 469 What this Chapter Has Shown 473 Appendix 13.A: Some Algebra for Logs and Exponents 475 Appendix 13.B: Introduction to Probit Analysis 476 14. Multinomial and Ordinal Logistic Regression and Tobit Regression 483 What this Chapter Is about 483 Multinomial Logit Analysis 483 A Worked Example: Foreign Language-competence in the Czech Republic 485 Independence of Irrelevant Alternatives (IIA) 490 Ordinal Logistic Regression 492 A Worked Example: Political Party Identification in the U.S., 1998 494 Converting the Logits to Y*-standardized Form 497 Getting Predicted Percentages 498 Constructing Odds Ratios. 500 Comparisons to Other Estimating Procedures: -Gologit2- 501 Ordinary Least Squares as an Alternative 505 Tobit Regression (and Allied Procedures) for Censored Dependent Variables 506 A Worked Example: Frequency of Sex 510 Other Models for the Analysis of Limited Dependent Variables 515 What this Chapter Has Shown 516 15. Improving Causal Inference: Fixed Effects and Random Effects Modeling 517 What this Chapter Is about 517 Introduction 517 Fixed Effects Models for Continuous Variables 519 Fundamental FE Equation 520 Allowing the Slopes of the x¿s to Vary 521 Testing Whether the Effects of the Time-invariant Variables Vary over Time 522 Interactions Between Time-constant and Time-varying Variables 522 Analyzing More than Two Time Points 523 Fixing Effects Across Individuals Rather than over Time 526 Limitations of Fixed Effects Approaches and Cautions to Keep in Mind 527 Random Effects Models for Continuous Variables 528 A Worked Example: the Determinants of Income in China 531 Fixed Effects Models for Binary Outcomes 536 Random Effects Models for Binary Outcomes 538 A Worked Example with a Binary Outcome: the Effect of Migration on School Enrollment among South African Blacks 538 A Bibliographic Note 544 What this Chapter Has Shown 544 16. Final Thoughts and Future Directions: Research Design and Interpretation Issues 545 What this Chapter Is about 545 Research Design Issues 545 Comparisons Are the Essence 545 Population Subgroups, Populations, and Historical Periods 546 Natural Experiments 552 Multilevel Analysis 553 Endogeneity, Sample Selection Bias, and Other Threats to Correct Causal Inference 555 Treatment Effects 555 Instrumental Variables (IV) Regression 557 Sample Selection Bias 559 Heckman Selection Model 560 Endogenous Switching Regression 561 Propensity Score Matching 562 Structural Equation Models 565 The Importance of Probability Sampling 570 The Concept of a ¿Super-population¿ 571 Pooling Data from Multiple Surveys 574 A Final Note: Good Professional Practice 576 Understand the Properties of Your Data 576 Explore Alternatives to Your a Priori Hypotheses 578 Conduct Sensitivity Analysis 580 Document Your Work 581 Error Checking 583 What this Chapter Has Shown 584 Appendix A: Data Descriptions and Download Locations for the Data Used in This Text 585 China 585 Eastern Europe 586 South Africa 587 U.S. General Social Survey 588 Appendix B: Doing Survey Estimation with the General Social Survey 590 Introduction 590 Analyzing Data from a Single Year 591 The 1972-1976 Block Quota Samples 591 The 1977-2002 Surveys, Except 1982 and 1987 592 The 1982 and 1987 Surveys with Oversamples of Blacks 594 The 2004 and 2006 Surveys 594 The FORMWT Variable 595 Pooling Surveys from More Than One Year 596 References 597 Index 635

Library of Congress Subject Headings for this publication:

Social sciences -- Research -- Statistical methods.

Sociology -- Research -- Statistical methods.

Sociology -- Statistical methods.

Social sciences -- Statistical methods -- Computer programs.

Stata.