What works : a new approach to program and policy analysis / Kenneth J. Meier, Jeff Gill.

Format:

Book

Author/Creator:

Meier, Kenneth J., 1950-

Contributor:

Gill, Jeff.

Language:

English

Subjects (All):

Policy sciences--Statistical methods.

Policy sciences.

Physical Description:

xiv, 157 pages : illustrations ; 23 cm

Place of Publication:

Boulder, Colo. : Westview Press, [2000]

Contents:

1 Zen and the Art of Policy Analysis: Substantively Weighted Analytical Techniques 1

1.2 Sciences of the Artificial 2

1.3 A Graphical Overview 3

1.4 Heterogeneity Is Good 5

1.5 The Zen of Weighting 7

1.6 Why Not Just Run Separate Regressions? 8

1.7 Don't Throw Away Your Other Tools 10

1.8 When Not to Use SWAT 10

1.8.1 Case 1: No Residual Variation 11

1.8.2 Case 2: When There Is a Clear Break Between Two Sets of Programs 11

1.8.3 Case 3: An Incorrect Functional Form 12

2 An Introduction to Substantively Weighted Least Squares 17

2.2 Child Support Enforcement 18

2.3 The Original Study 19

2.4 Regression Diagnostics 21

2.5 Robust and Resistant Regression 24

2.6 SWLS 27

2.8 SWAT Versus Best Practices 34

2.9 SWAT Versus L-Regression: A Methodological Interlude 35

2.10 But So What? 36

2.11 Addendum: Data Description 37

3 The Theory and Application of Generalized Substantively Reweighted Least Squares 41

3.2 Distribution of the Jackknifed Residuals 43

3.3 Relationship to the F Distribution 45

3.4 Five Assumptions and a Warning 47

3.5 Using the GSRLS Procedure 48

3.7 SWAT Versus the Chow Test 55

3.9 Addendum: Splus/R Code for SWAT 57

4 Substantively Weighted Analytical Techniques for Successes and Failures: Swls and Gsrls 59

4.2 The Dataset 61

4.2.1 The Outcome Variable 61

4.2.2 The Explanatory Variables 61

4.3 SWLS Findings 64

4.4 GSRLS Results 71

4.5 Differences Between SWLS and GSRLS 71

5 Separating Excellent Agencies from the Good Ones: Pushing the Extremes of the Data Distribution 83

5.2 An Education Production Function 85

5.3 Linear Model Findings 85

5.4 The Pretty Good Agencies 86

5.5 The Super Agencies 89

5.6 Lucky or Good? 90

5.7 What the Super Agencies Do Differently 92

5.8 Extensions 96

6 Weighting with an Exogenous Variable or with Two Exogenous Variables: Equity Versus Excellence in Organizations 99

6.2 The Theoretical Setting for the Study 100

6.3 Educational Performance 101

6.3.1 The Outcome Variable 102

6.3.2 Explanatory Variables 103

6.4 Substantively Weighted Least Squares 104

6.5 Findings 104

6.6 Weighting in Two Dimensions 112

6.7 Two-Dimensional Weights: An Illustration 113

7 Swat in Pooled Analysis 119

7.2 Zen and the Art of Pooling 120

7.3 The Full

and Then Some

Prison Blues 123

7.4 Looking at the Data 125

7.5 Applying SWAT to the Data 127

7.5.1 A Note of Caution 130

7.5.2 The Winners: The Low Crime States 131

7.5.3 The Losers: High Crime States 134

7.6 Path Dependence 138

7.7 A Second Example: Educational Performance 138

7.9 Addendum: Data Description 143

8 The Zen and the Practice: Some Final Remarks 147

8.2 The Secret Life of Residuals 148

8.3 Weighting the Artificial 149

8.4 Hypotheses 150

8.5 If I Had a Hammer 151

8.6 The Final Word 152.

Notes:

Includes bibliographical references and index.

ISBN:

0813397820

0813397812

OCLC:

44791298