Premium
SYNTHETIC VALIDITY AND ITS APPLICATION TO THE UNIFORM GUIDELINES VALIDATION REQUIREMENTS
Author(s) -
TRATTNER MARVIN H.
Publication year - 1982
Publication title -
personnel psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.076
H-Index - 142
eISSN - 1744-6570
pISSN - 0031-5826
DOI - 10.1111/j.1744-6570.1982.tb02203.x
Subject(s) - construct validity , criterion validity , validation test , external validity , selection (genetic algorithm) , construct (python library) , psychology , data validation , test validity , test (biology) , reliability engineering , computer science , artificial intelligence , psychometrics , social psychology , engineering , database , clinical psychology , paleontology , biology , programming language
The Uniform Guidelines for Employee Selection Procedures have served to create an urgent need for efficient validation methods that can be generalized to a class of occupations. The one method currently authorized for such a purpose by the Guidelines is synthetic validation. (The Guidelines erroneously describe the synthetic validity paradigm as construct validity.) Approaches to synthetic validity employed by Lawshe, Guion, McCormick, and Primoff are described. Their extent of conformance to the Guidelines validation requirements is noted. Primoff's J ‐Coefficient approach is recommended for two reasons; it meets the Guidelines requirements and under certain circumstances it permits the test user to estimate the traditional validity coefficient. An illustrated example of Primoff's method is presented.