Thursday, April 2, 2015

Methods Matter Improving Causal Inference in Educational and Social Science Research






Methods Matter Improving Causal Inference in Educational and Social Science Research
Buku ini diterbitkan tahun 2011  Oleh  Oxford University Press., adalah buku edisi Pertama.



Judul:  Methods Matter Improving Causal Inference
in Educational and Social Science Research
Oleh:  Richard J. Murnane, et al
Penerbit:   Oxford University Press.
Tahun: 2011
Jumlah Halaman: 414  hal.


Penulis:

Richard J. Murnane
John B. Willett


Lingkup Pembahasan:
Buku ini  mengemukakan 14 pokok bahasan utama, yaitu  1 Tantangan Penelitian Pendidikan,
2 Pentingnya Teori, 3 Merancang Penelitian untuk  Menetapkan Pertanyaan kausal,  4 Penyidik-Merancang Percobaan Acak, 5 Tantangan dalam Merancang, Pelaksana, dan Belajar dari Percobaan Acak, 6 statistik Power dan Sampel, 7 Experimental Research Ketika Peserta Apakah Clustered Dalam Grup Utuh, 8 Menggunakan Percobaan Alam Memberikan Penyelesaian Variabilitas "Diperdebatkan eksogen",  9 Memperkirakan kausal Efek Menggunakan Pendekatan Regresi-Diskontinuitas, 10 Memperkenalkan Instrumental-Variabel Estimasi, 11 Menggunakan IVE untuk Recover Pengaruh Perlakuan dalam Kuasi-Eksperimen, 12 Berurusan dengan Bias di Treatment Efek Perkiraan dari data Non eksperimental, 13 Pelajaran metodologis dari Long Quest, dan 14 Pelajaran substantif dan Pertanyaan Baru.


Daftar Isi:

Preface xi
1     The Challenge for Educational Research 3

       The Long Quest 3
       The Quest Is Worldwide 9
       What This Book Is About 10
       What to Read Next 12
2     The Importance of Theory 14
       What Is Theory? 15
       Theory in Education 19
       Voucher Theory 21
       What Kind of Theories? 24
       What to Read Next 24
3     Designing Research to Address Causal Questions 26
       Conditions to Strive for in All Research 27
       Making Causal Inferences 29
       Past Approaches to Answering Causal Questions in Education 31
       The Key Challenge of Causal Research 33
       What to Read Next 39
4     Investigator-Designed Randomized Experiments 40
       Conducting Randomized Experiments 41
           The Potential Outcomes Framework 41
           An Example of a Two-Group Experiment 45
       Analyzing Data from Randomized Experiments 48
           The Better Your Research Design, the Simpler Your Data Analysis 48
           Bias and Precision in the Estimation of Experimental Effects 52
       What to Read Next 60
5     Challenges in Designing, Implementing, and Learning from Randomized Experiments 61
        Critical Decisions in the Design of Experiments 62
            Defi ning the Treatment 64
            Defi ning the Population from Which Participants Will Be Sampled 66
            Deciding Which Outcomes to Measure 67
            Deciding How Long to Track Participants 68
        Threats to the Validity of Randomized Experiments 69
            Contamination of the Treatment–Control Contrast 70
            Cross-overs 70
            Attrition from the Sample 71
            Participation in an Experiment Itself Affects Participants’ Behavior 73
        Gaining Support for Conducting Randomized Experiments: Examples from India 74
            Evaluating an Innovative Input Approach 75
            Evaluating an Innovative Incentive Policy 79
        What to Read Next 81
6     Statistical Power and Sample Size 82
        Statistical Power 83
            Reviewing the Process of Statistical Inference 83
            Defi ning Statistical Power 92
        Factors Affecting Statistical Power 96
            The Strengths and Limitations of Parametric Tests 101
            The Benefi ts of Covariates 102
            The Reliability of the Outcome Measure Matters 103
            The Choice Between One-Tailed and Two-Tailed Tests 105
        What to Read Next 106
7     Experimental Research When Participants Are Clustered Within Intact Groups 107
       Random-Intercepts Multilevel Model to Estimate Effect Size When Intact Groups Are
            Randomized to Experimental Conditions 110
       Statistical Power When Intact Groups of Participants Are Randomized to Experimental Conditions 120
       Statistical Power of the Cluster-Randomized Design and Intraclass Correlation 122
       Fixed-Effects Multilevel Models to Estimate Effect Size When Intact Groups of
            Participants Are Randomized to Experimental Conditions 128
       Specifying a Fixed-Effects Multilevel Model 128
       Choosing Between Random- and Fixed-Effects Specifi cations 131
       What to Read Next 134
8     Using Natural Experiments to Provide “Arguably Exogenous”  Treatment Variability 135
       Natural- and Investigator-Designed Experiments: Similarities and Differences 136
       Two Examples of Natural Experiments 137
           The Vietnam-Era Draft Lottery 137
           The Impact of an Offer of Financial Aid for College 141
       Sources of Natural Experiments 145
       Choosing the Width of the Analytic Window 150
       Threats to Validity in Natural Experiments with a Discontinuity Design 152
           Accounting for the Relationship Between the Outcome and the Forcing Variable in a
                Discontinuity Design 153
           Actions by Participants Can Undermine Exogenous Assignment to Experimental
               Conditions in a Natural Experiment with a Discontinuity Design 163
       What to Read Next 164
9     Estimating Causal Effects Using a Regression-Discontinuity Approach 165
       Maimonides’ Rule and the Impact of Class Size on Student Achievement 166
           A Simple First-Difference Analysis 170
           A Difference-in-Differences Analysis 171
           A Basic Regression-Discontinuity Analysis 174
       Choosing an Appropriate Bandwidth 181
           Generalizing the Relationship Between the Outcome and the Forcing Variable 186
            Specifi cation Checks Using Pseudo-Outcomes and Pseudo–Cut-offs 192
            Regression-Discontinuity Designs and Statistical Power 195
       Additional Threats to Validity in a Regression-Discontinuity Design 197
       What to Read Next 202
10   Introducing Instrumental-Variables Estimation 203
       Introducing Instrumental-Variables Estimation 204
           Bias in the OLS Estimate of the Causal Effect of Education on Civic Engagement 206
           Instrumental-Variables Estimation 215
       Two Critical Assumptions That Underpin Instrumental-Variables Estimation 223
          Alternative Ways of Obtaining the Instrumental-Variables Estimate 226
           Obtaining an Instrumental-Variables Estimate by the Two-Stage Least-Squares Method 227
           Obtaining an Instrumental-Variables Estimate by Simultaneous-Equations Estimation 233
       Extensions of the Basic Instrumental-Variable Estimation Approach 238
           Incorporating Exogenous Covariates into Instrumental-Variable Estimation 238
           Incorporating Multiple Instruments into the First-Stage Model 243
           Examining the Impact of Interactions Between the Endogenous Question
                Predictor and Exogenous Covariates in the Second-Stage Model 247
          Choosing Appropriate Functional Forms for Outcome/Predictor
                Relationships in First- and Second-Stage Models 251
       Finding and Defending Instruments 252
           Proximity of Educational Institutions 253
           Institutional Rules and Personal Characteristics 257
           Deviations from Cohort Trends 261
           The Search Continues 263
      What to Read Next 264
11   Using IVE to Recover the Treatment Effect in a Quasi-Experiment 265
       The Notion of a “Quasi-Experiment” 267
       Using IVE to Estimate the Causal Impact of a Treatment in a Quasi-Experiment 269
       Further Insight into the IVE (LATE) Estimate, in the Context of Quasi-Experimental Data 274
       Using IVE to Resolve “Fuzziness” in a Regression-Discontinuity Design 280
       What to Read Next 285
12   Dealing with Bias in Treatment Effects Estimated from Nonexperimental Data 286
       Reducing Observed Bias by the Method of Stratifi cation 289
            Stratifying on a Single Covariate 289
            Stratifying on Covariates 299
       Reducing Observed Bias by Direct Control for Covariates Using Regression Analysis 304
       Reducing Observed Bias Using a Propensity-Score Approach 310
            Estimation of the Treatment Effect by Stratifying on Propensity Scores 316
            Estimation of the Treatment Effect by Matching on Propensity Scores 321
            Estimation of the Treatment Effect by Weighting by the Inverse of the Propensity Scores 324
       A Return to the Substantive Question 328
       What to Read Next 331
13   Methodological Lessons from the Long Quest 332
       Be Clear About Your Theory of Action 333
       Learn About Culture, Rules, and Institutions in the Research Setting 335
       Understand the Counterfactual 337
       Always Worry About Selection Bias 338
       Use Multiple Outcome Measures 340
       Be on the Lookout for Longer-Term Effects 341
       Develop a Plan for Examining Impacts on Subgroups 342
       Interpret Your Research Results Correctly 344
       Pay Attention to Anomalous Results 346
       Recognize That Good Research Always Raises New Questions 348
       What to Read Next 349
14   Substantive Lessons and New Questions 350
       Policy Guideline 1: Lower the Cost of School Enrollment 352
           Reduce Commuting Time 352
           Reduce Out-of-Pocket Educational Costs 353
           Reduce Opportunity Costs 354
       Policy Guideline 2: Change Children’s Daily Experiences in School 355
           More Books? 355
           Smaller Classes? 356
           Better Teaching? 357
       Policy Guideline 3: Improve Incentives 359  
           Improve Incentives for Teachers 359
           Improve Incentives for Students 361
       Policy Guideline 4: Create More Schooling Options for Poor Children 364
           New Private-School Options 365
           New Public-School Options 365
       Summing Up 367
       Final Words 368

References 369
Index 381



Berminat?
Email: zanetapm@gmail.com





Methods Matter Improving Causal Inference in Educational and Social Science Research Rating: 4.5 Diposkan Oleh: Unknown

0 comments:

Post a Comment