Seven Rules for Social Research Book

Preface     xi
The First Rule: There Should Be the Possibility of Surprise in Social Research     1
Selecting a Research Question     2
Researchable Questions     2
Interesting Questions     4
Selecting a Sample     18
Samples in Qualitative Studies     23
Is Meaningful Social Research Possible?     26
Summary     29
Student Exercises on Rule 1     31
The Second Rule: Look for Differences That Make a Difference, and Report Them     36
You Can't Explain a Variable with a Constant     37
Maximizing Variance to Find the Effect of a Cause     39
Size versus Statistical Significance     41
Comparing Effects Where There Is a Common Metric     42
Calibration: Converting Explanatory Variables to a Common Metric     44
Substantive Profiling: The Use of Telling Comparisons     46
Visual Presentation of Results     51
Policy Importance     53
Importance for Theory     54
Conclusion     56
Student Exercises on Rule 2     58
The Third Rule: Build Reality Checks into Your Research     64
Internal Reality Checks     65
Reality Checks onData-Dubious Values and Incomplete Data     65
Reality Checks on Measures-Aim for Consistency in Conceptualization and Measurement     69
Reality Checks on Models-The Formal Equivalence Check     71
External Reality Checks: Validation with Other Data and Methods     76
Using Causal-Process Observations to Test Plausibility of Results     77
Using Ethnographic Data to Help Interpret Survey Results     79
Other Examples of Multiple-Method Research     81
Concluding Remark     82
Student Exercises on Rule 3     84
The Fourth Rule: Replicate Where Possible     90
Sources of Uncertainty in Social Research     91
Overview: From Population to Sample and Back to Population     93
Measurement Error as a Source of Uncertainty     100
Illustration: Two Methods for Estimating Global Poverty     101
Toward a Solution: Identical Analyses of Parallel Data Sets     105
Meta-analysis: Synthesizing Results Formally across Studies     106
Summary: Your Confidence Intervals Are Too Narrow     109
Student Exercises on Rule 4     111
The Fifth Rule: Compare Like with Like     120
Correlation and Causality     121
Types of Strategies for Comparing Like with Like      129
Matching versus Looking for Differences     130
The Standard Regression Method for Comparing Like with Like     131
Critique of the Standard Linear Regression Strategy     132
Comparing Like with Like Through Fixed-Effects Methods     134
First-Difference Models: Subtracting Out the Effects of Confounding Variables     134
Special Case: Growth-Rate Models     138
Sibling Models     140
Comparing Like with Like through Matching on Measured Variables     146
Exact Matching     146
Propensity-Score Method     147
Matching as a Preprocessing Strategy for Reducing Model Dependence     151
Comparing Like with Like through Naturally Occurring Random Assignment     152
Instrumental Variables: Matching through Partial Random Assignment     153
Matching Through Naturally Occurring Random Assignment to the Treatment Group     158
Comparison of Strategies for Comparing Like with Like     159
Conclusion     162
Student Exercises on Rule 5     165
The Sixth Rule: Use Panel Data to Study Individual Change and Repeated Cross-section Data to Study Social Change     172
Analytic Differences between Panel and Repeated Cross-section Data     173
Three General Questions about Change     175
Changing-Effect Models, Part 1: Two Points in Time     176
Changing-Effect Models, Part 2: Multilevel Models with Time as the Context     182
What We Want to Know     183
The General Multilevel Model     183
Convergence Models     185
The Sign Test for Convergence: Comparing Your [phi]s and [delta]s     186
Convergence Model versus Changing-Effect Model     191
Bridging Individual and Social Change: Estimating Cohort Replacement Effects     195
An Accounting Scheme for Social Change     197
Linear Decomposition Method     198
Summary     201
Student Exercises on Rule 6     203
The Seventh Rule: Let Method Be the Servant, Not the Master     207
Obsession with Regression     209
Naturally Occurring Random Assignment, Again     209
Decomposition Work in the Social Sciences     218
Decomposition of Variance and Inequality     220
Decomposition of Segregation Indexes     222
The Effects of Social Context     226
Context Effects as Objects of Study     227
Context Effects as Nuisance     230
Critical Tests in Social Research      231
Conclusion     235
Student Exercises on Rule 7     236
References     241
Index     253