[ Aris Spanos (2012), “Philosophy of Econometrics”, in Uskali Mäki (ed.)(2012), Philosophy of Economics: Handbook of the Philosophy of Science 13 (North Holland), pp. 329-393. ]
1. Introduction
2. Relevant Philosophical/Methodological Issues
2.1. Probing the different ways an inference might be in error
2.2. The Pre-Eminence of Theory (PET) perspective
2.2.1. Statistical misspecification vs. the ‘realisticness’ issue
2.3. Reflecting on textbook econometrics
3. Philosophy of Science and Empirical Modeling
3.1. Logical positivism/empiricism
3.2. The fading of logical empiricism
3.3. The New Experimentalism
3.4. Learning from Error
4. Statistical Inference and Its Foundational Problems
4.1. Frequentist statistics and its foundational problems
4.2. Bayesianism and its criticisms of the frequentist approach
5. Error-Statistics (E-S) and Inductive Inference
5.1. Induction by enumeration vs. model-based induction
5.2. The frequentist interpretation of probability
5.3. Statistical induction: factual vs. hypothetical reasoning
5.4. Factual reasoning: estimation and prediction
5.5. Hypothetical reasoning: testing
5.6. Post-data error probabilities in confidence intervals
5.7. Severity: a post-data evaluation of inference
5.7.1. Severity reasoning
5.7.2. Severe testing and the p-value
5.7.3. The fallacies of acceptance and rejection
5.7.4. Revisiting observed Confidence Intervals (CI)
5.8. Revisiting Bayesian criticisms of frequentist inference
6. Statistical Adequacy and the Reliability of Inference
6.1. A statistical model can have ‘a life of its own’
6.2. Relating the substantive and the statistical information
6.3. Mis-Specification (M-S) testing and Respecification
6.4. Methodological problems associated with M-S testing
6.4.1. Illegitimate double-use of data
6.4.2. The infinite regress and circularity charges against M-S testing
6.4.3. Revisiting the pre-test bias argument
7. Philosophical/Methodological Issues Pertaining to Econometrics
7.1. Statistical model specification vs. model selection
7.2. The reliability/precision of inference and robustness
7.3. Weak assumptions and the reliability/precision of inference
7.4. Statistical ‘Error-fixing’ strategies and data mining
7.5. Unreliable strategies for ‘upholding’ a theory
7.6. Revisiting the omitted variables bias argument
7.7. If everything else fails, blame multicollinearity
8. Summary and Conclusions
1. Introduction
2. Relevant Philosophical/Methodological Issues
2.1. Probing the different ways an inference might be in error
2.2. The Pre-Eminence of Theory (PET) perspective
2.2.1. Statistical misspecification vs. the ‘realisticness’ issue
2.3. Reflecting on textbook econometrics
3. Philosophy of Science and Empirical Modeling
3.1. Logical positivism/empiricism
3.2. The fading of logical empiricism
3.3. The New Experimentalism
3.4. Learning from Error
4. Statistical Inference and Its Foundational Problems
4.1. Frequentist statistics and its foundational problems
4.2. Bayesianism and its criticisms of the frequentist approach
5. Error-Statistics (E-S) and Inductive Inference
5.1. Induction by enumeration vs. model-based induction
5.2. The frequentist interpretation of probability
5.3. Statistical induction: factual vs. hypothetical reasoning
5.4. Factual reasoning: estimation and prediction
5.5. Hypothetical reasoning: testing
5.6. Post-data error probabilities in confidence intervals
5.7. Severity: a post-data evaluation of inference
5.7.1. Severity reasoning
5.7.2. Severe testing and the p-value
5.7.3. The fallacies of acceptance and rejection
5.7.4. Revisiting observed Confidence Intervals (CI)
5.8. Revisiting Bayesian criticisms of frequentist inference
6. Statistical Adequacy and the Reliability of Inference
6.1. A statistical model can have ‘a life of its own’
6.2. Relating the substantive and the statistical information
6.3. Mis-Specification (M-S) testing and Respecification
6.4. Methodological problems associated with M-S testing
6.4.1. Illegitimate double-use of data
6.4.2. The infinite regress and circularity charges against M-S testing
6.4.3. Revisiting the pre-test bias argument
7. Philosophical/Methodological Issues Pertaining to Econometrics
7.1. Statistical model specification vs. model selection
7.2. The reliability/precision of inference and robustness
7.3. Weak assumptions and the reliability/precision of inference
7.4. Statistical ‘Error-fixing’ strategies and data mining
7.5. Unreliable strategies for ‘upholding’ a theory
7.6. Revisiting the omitted variables bias argument
7.7. If everything else fails, blame multicollinearity
8. Summary and Conclusions
(2026.07.24.)
댓글 없음:
댓글 쓰기