内容

目次

Research Designs and Causal Inferences: On Lord’s Paradox.- 1. Introduction.- 2. A Model for Causal Inference.- 2.1 The elements of the model.- 2.2 Discussion of the elements of the model.- 2.3 Three kinds of studies.- 2.4 Causal effects and related quantities in causal studies.- 3. Lord’s examples.- 3.1 Example 1: Two subpopulations receiving one treatment.- 3.2 Example 2: A descriptive study.- 3.3 Example 3: Contemplating new treatments.- 3.4 Example 4: Two explicit treatments.- 4. Discussion.- Appendix: Randomization and Inference for Causal Effects.- A.1 The completely randomized experiment.- A.2 Causal effects in subpopulations.- A.3 Randomization within populations.- A.4 Randomization based on a covariate.- A.5 Randomization based on a covariate within subpopulations.- A.6 Descriptive studies.- References.- Tables and Figures.- Toward Conducting Benefit-Cost Analysis of Data Programs.- 1. An Illustration of the Use of Benefit-Cost Analysis.- 2. A Closer Look at Benefit-Cost Analysis.- 3. Data Quality.- 4. Benefits.- 5. Uses of Data in Decisionmaking.- 6. Nonoptimal Data Use.- 7. Summary and Conclusions.- Tables and Figures.- The Design and Analysis of Longitudinal Surveys: Controversies and Issues of Cost and Continuity.- Abstract.- 1. Introduction.- 2. Conceptual Framework.- 3. The Key Controversy: It’s All a Matter of Perspective.- 4. Costs and the Effects of Scale.- 5. Longitudinal Missingness, Attrition, and Imputation.- 6. Drawing Formal Inferences from Sample Survey Data.- 7. Summary and Answer to Questions.- References.- The Role of Panel Studies in a World of Scarce Research Resources.- 1. Analytic Benefits of Panel Designs.- 1.1. The role of panel data in describing and explaining change.- 1.2. Panels as successive cross-sections.- 2. Panel Designs and Data Quality.- 2.1. Effects of memory on response error.- 2.2. Effect of panel participation.- 2.3. Analytic costs of “cheaper” panel studies.- 3. Relative Monetary Costs of Longitudinal Studies.- 4. Summary.- References.