Bernt P. Stigum
- Published in print:
- 2014
- Published Online:
- September 2015
- ISBN:
- 9780262028585
- eISBN:
- 9780262323109
- Item type:
- chapter
- Publisher:
- The MIT Press
- DOI:
- 10.7551/mitpress/9780262028585.003.0009
- Subject:
- Economics and Finance, Econometrics
In philosophy, a scientific explanation has four components: a family of sentences, H, to be explained, a list of antecedent conditions, a list of general laws, and arguments that demonstrate that H ...
More
In philosophy, a scientific explanation has four components: a family of sentences, H, to be explained, a list of antecedent conditions, a list of general laws, and arguments that demonstrate that H is a logical consequence of the antecedent conditions and the laws. The chapter explains why such a characterization of scientific explanations is of little use in economics, and presents two different formal characterizations of logically and empirically adequate scientific explanations – one (SE1) for economics and one (SE2) for econometrics. In SE1, H is a family of sentences; the antecedent conditions are axioms of the data universe; the laws are axioms or theorems of an empirically relevant economic theory; and the arguments are the arguments of first-order logic. In SE2, H is a family of statistical relations; the antecedent conditions are axioms of a data universe; the laws are axioms or theorems of an empirically relevant economic theory; and the arguments are the arguments of mathematical statistics. The chapter exemplifies the two schemes with an SE1-explanation of regularities in experimental economics and with an SE2-explanation of regularities in a financial market in which the statistical arguments of formal econometrics are contrasted with the statistical arguments of present-day econometrics.Less
In philosophy, a scientific explanation has four components: a family of sentences, H, to be explained, a list of antecedent conditions, a list of general laws, and arguments that demonstrate that H is a logical consequence of the antecedent conditions and the laws. The chapter explains why such a characterization of scientific explanations is of little use in economics, and presents two different formal characterizations of logically and empirically adequate scientific explanations – one (SE1) for economics and one (SE2) for econometrics. In SE1, H is a family of sentences; the antecedent conditions are axioms of the data universe; the laws are axioms or theorems of an empirically relevant economic theory; and the arguments are the arguments of first-order logic. In SE2, H is a family of statistical relations; the antecedent conditions are axioms of a data universe; the laws are axioms or theorems of an empirically relevant economic theory; and the arguments are the arguments of mathematical statistics. The chapter exemplifies the two schemes with an SE1-explanation of regularities in experimental economics and with an SE2-explanation of regularities in a financial market in which the statistical arguments of formal econometrics are contrasted with the statistical arguments of present-day econometrics.
