| List of Figures | p. ix |
| List of Tables | p. x |
| Data Sources | p. xi |
| Introduction | p. 1 |
| 1 The Use of Linear Regression in Labour Economics | p. 4 |
| 1.1 The Linear Regression Model-A Review of Some Basic Results | p. 5 |
| 1.2 Specification Issues in the Linear Model | p. 10 |
| 1.3 Using the Linear Regression Model in Labour Economics-the Mincer Earnings Equation | p. 20 |
| 1.4 Concluding Remarks | p. 30 |
| Appendix: The Mechanics of Ordinary Least Squares Estimation | p. 32 |
| 2 Further Regression Issues in Labour Economics | p. 34 |
| 2.1 Decomposing Differences Between Groups-Oaxaca and Beyond | p. 35 |
| 2.2 Quantile Regression and Earnings Decompositions | p. 42 |
| 2.3 Regression with Panel Data | p. 44 |
| 2.4 Estimating Standard Errors | p. 48 |
| 2.5 Concluding Remarks | p. 51 |
| 3 Dummy and Ordinal Dependent Variables | p. 53 |
| 3.1 The Linear Model and Least Squares Estimation | p. 53 |
| 3.2 Logit and Probit Models-A Common Set-up | p. 56 |
| 3.3 Interpreting the Output | p. 61 |
| 3.4 More Than Two Choices | p. 68 |
| 3.5 Concluding Remarks | p. 74 |
| 4 Selectivity | p. 76 |
| 4.1 A First Approach-Truncation Bias and a Pile-up of Zeros | p. 77 |
| 4.2 Sample Selection Bias-Missing Values | p. 79 |
| 4.3 Marginal Effects and Oaxaca Decompositions in Selectivity Models | p. 84 |
| 4.4 The Roy Model-The Role of Comparative Advantage | p. 87 |
| 4.5 The Normality Assumption | p. 90 |
| 4.6 Concluding Remarks | p. 91 |
| Appendix: |
| 1 The conditional expectation of the error term under truncation | p. 93 |
| 2 The conditional expectation of the error term with sample selection | p. 94 |
| 3 Marginal effects in the sample selection model | p. 95 |
| 4 The conditional expectation of the error terms in two equations with selectivity bias | p. 96 |
| 5 Duration Models | p. 97 |
| 5.1 Analysing Completed Durations | p. 100 |
| 5.2 Econometric Modelling of Spell Lengths | p. 102 |
| 5.3 Censoring: Complete and Incomplete Durations | p. 108 |
| 5.4 Modelling Issues with Duration Data | p. 113 |
| 5.5 Concluding Remarks | p. 117 |
| Appendix: |
| 1 The expected duration of completed spell is equal to the integral of the survival function | p. 119 |
| 2 The integrated hazard function | p. 119 |
| 3 The log likelihood function with discrete (grouped) duration data | p. 120 |
| 6 Evaluation of Policy Measures | p. 122 |
| 6.1 The Experimental Approach | p. 123 |
| 6.2 The Quasi-experimental Approach-A Control Group can be Defined Exogenously | p. 125 |
| 6.3 Evaluating Policies in a Non-experimental Context: The Role of Selectivity | p. 131 |
| 6.4 Concluding Remarks | p. 136 |
| Appendix: |
| 1 Derivation of the average treatment effect as an OLS estimator | p. 138 |
| 2 Derivation of the Wald estimator | p. 139 |
| Conclusion | p. 141 |
| Bibliography | p. 143 |
| Index | p. 147 |