Premium
A new approach to determine the reliability of the rater technique by use of Smith's rating scales
Author(s) -
De Freitas Antǒnio Carlos,
Ribeiro Guilherme Henrique Caspary
Publication year - 1977
Publication title -
journal of clinical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.124
H-Index - 119
eISSN - 1097-4679
pISSN - 0021-9762
DOI - 10.1002/1097-4679(197710)33:4<1032::aid-jclp2270330424>3.0.co;2-v
Subject(s) - psychology , inter rater reliability , rating scale , reliability (semiconductor) , clinical psychology , psychoanalysis , reliability engineering , applied psychology , developmental psychology , engineering , thermodynamics , power (physics) , physics
Used as the base for the study a sample of ratings of 15 pilots made by supervisors during a 6‐month period. This report of that evaluation confirms the validity of technique (Smith, 1974) using intraclass correlation and Spearman's rho to identify which raters were unreliable. More rigorous criteria for the elimination of unreliable raters were developed by applying the Kuder‐Richardson reliability coefficient (KR‐20) as a more precise measure of reliability than the intraclass correlation criteria used by Smith. The proposed technique permits improved reliability by the criterious removal of raters who are not contributing significantly to the total consistency of the behavioral scale and thus eliminates partially the inconvenience of the unavailable significant table of data to test the acceptance of the reliability.