Contents

Preface	ix

1 The Science in Social Science	3
  1.1 Introduction	3
    1.1.1 Two Styles of Research, One Logic of Inference	3
    1.1.2 Defining Scientific Research in Social Sciences	7
    1.1.3 Science and Complexity	9
  1.2 Major Components of Research Design	12
    1.2.1 Improving Research Questions	14
    1.2.2 Improving Theory	19
    1.2.3 Improving Data Quality	23
    1.2.4 Improving the Use of Existing Data	27
  1.3 Themes of This Volume	28
    1.3.1 Using Observable Implications to Connect Theory and Data	28
    1.3.2 Maximizing Leverage	29
    1.3.3 Reporting Uncertainty	31
    1.3.4 Thinking like a Social Scientist: Skepticism and Rival Hypotheses	32

2 Descriptive Inference	34
  2.1 General Knowledge and Particular Facts	35
    2.1.1 "Interpretation" and Inference	36
    2.1.2 "Uniqueness," Complexity, and Simplification	42
    2.1.3 Comparative Case Studies	43
  2.2 Inference: the Scientific Purpose of Data Collection	46
  2.3 Formal Models of Qualitative Research	49
  2.4 A Formal Model of Data Collection	51
  2.5 Summarizing Historical Detail	53
  2.6 Descriptive Inference	55
  2.7 Criteria for Judging Descriptive Inferences	63
    2.7.1 Unbiased Inferences	63
    2.7.2 Efficiency	66

3 Causality and Causal Inference	75
  3.1 Defining Causality	76
    3.1.1 The Definition and a Quantitative Example	76
    3.1.2 A Qualitative Example	82
  3.2 Clarifying Alternative Definitions of Causality	85
    3.2.1 "Causal Mechanisms"	85
    3.2.2 "Multiple Causality"	87
    3.2.3 "Symmetric" and "Asymmetric" Causality	89
  3.3 Assumptions Required for Estimating Causal Effects	91
    3.3.1 Unit Homogeneity	91
    3.3.2 Conditional Independence	94
  3.4 Criteria for Judging Causal Inferences	97
  3.5 Rules for Constructing Causal Theories	99
    3.5.1 Rule 1: Construct Falsifiable Theories	100
    3.5.2 Rule 2: Build Theories That Are Internally Consistent	105
    3.5.3 Rule 3: Select Dependent Variables Carefully	107
    3.5.4 Rule 4: Maximize Concreteness	109
    3.5.5 Rule 5: State Theories in as Encompassing Ways as Feasible	113

4 Determining What to Observe	115
  4.1 Indeterminate Research Designs	118
    4.1.1 More Inferences than Observations	119
    4.1.2 Multicollinearity	122
  4.2 The Limits of Random Selection	124
  4.3 Selection Bias	128
    4.3.1 Selection on the Dependent Variable	129
    4.3.2 Selection on an Explanatory Variable	137
    4.3.3 Other Types of Selection Bias	138
  4.4 Intentional Selection of Observations	139
    4.4.1 Selecting Observations on the Explanatory Variable	140
    4.4.2 Selecting a Range of Values of the Dependent Variable	141
    4.4.3 Selecting Observations on Both Explanatory and Dependent Variables	142
    4.4.4 Selecting Observations So the Key Causal Variable Is Constant	146
    4.4.5 Selecting Observations So the Dependent Variable Is Constant	147
  4.5 Concluding Remarks	149

5 Understanding What to Avoid	150
  5.1 Measurement Error	151
    5.1.1 Systematic Measurement Error	155
    5.1.2 Nonsystematic Measurement Error	157
  5.2 Excluding Relevant Variables: Bias	168
    5.2.1 Gauging the Bias from Omitted Variables	168
    5.2.2 Examples of Omitted Variable Bias	176
  5.3 Including Irrelevant Variables: Inefficiency	182
  5.4 Endogeneity	185
    5.4.1 Correcting Biased Inferences	187
    5.4.2 Parsing the Dependent Variable	188
    5.4.3 Transforming Endogeneity into an Omitted Variable Problem	189
    5.4.4 Selecting Observations to Avoid Endogeneity	191
    5.4.5 Parsing the Explanatory Variable	193
  5.5 Assigning Values of the Explanatory Variable	196
  5.6 Controlling the Research Situation	199
  5.7 Concluding Remarks	206

6 Increasing the Number of Observations	208
  6.1 Single-Observation Designs for Causal Inference	209
    6.1.1 "Crucial" Case Studies	209
    6.1.2 Reasoning by Analogy	212
  6.2 How Many Observations Are Enough?	213
  6.3 Making Many Observations from Few	217
    6.3.1 Same Measures, New Units	219
    6.3.2 Same Units, New Measures	223
    6.3.3 New Measures, New Units	224
  6.4 Concluding Remarks	229

References	231
Index	239