Premium
Searching Data for Predictive Variables: A Set‐Theoretical Approach *
Author(s) -
Muir Donal E.
Publication year - 1969
Publication title -
sociological inquiry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.446
H-Index - 51
eISSN - 1475-682X
pISSN - 0038-0245
DOI - 10.1111/j.1475-682x.1969.tb00936.x
Subject(s) - set (abstract data type) , task (project management) , computer science , variable (mathematics) , data set , mill , epistemology , mathematics , data mining , artificial intelligence , philosophy , history , engineering , mathematical analysis , systems engineering , archaeology , programming language
A common task of the scientific investigator is to search out factors which will predict the presence or absence of some variable attribute in which he is interested. Procedures for such searches are spelled out in Mill's famous canons of discovery. A set‐theoretical analysis of these, however, discloses unnecessary inconsistency and difficulty, and leads to a single canon which extracts maximal information concerning causal factors from arbitrary sequences of cases, while reminding the researcher of the limitations of finite samples of both subjects and measures.