Cm. Russell et al., SIMULATION STUDY OF A PANEL OF RELIABILITY INDICATORS APPLIED TO PAIRED MEASUREMENTS, American journal of human biology, 6(3), 1994, pp. 311-320
Reliability is a subject of continuing discussion in biomedial special
ty areas, including physical anthropology and nutritional epidemiology
. The purpose of this study was to explore techniques of detecting dif
ferences between two evaluators or methods. A field study in which ant
hropometric dimensions would be taken by two independent evaluators on
each participant in a study group was simulated. A panel of reliabili
ty indicators was applied across a broad range of parameters using sim
ulation, and then the panel was applied to field anthropometric data.
The panel consisted of the intraclass correlation coefficient (ICC), p
aired t-test, a simultaneous test of evaluator means and variances, te
chnical error of measurement, mean absolute difference, and mean diffe
rence. The simultaneous test for equal evaluator means and variances u
ses regression to model paired differences versus paired sums. The sim
ulation demonstrated general properties of the reliability indicators
across many conditions of population variance, measurer bias, and meas
urer error variance. High values of ICC often exist in cases in which
the measurers are different. The simultaneous test is thus a powerful
method for detecting measurer differences, especially when combined wi
th the paired t-test. However, a single reliability indicator that is
sufficient to determine all measurer inconsistencies was not identifie
d. The field study and the simulation permitted the development of a l
ogical approach to determining the source and magnitude of measurer di
fferences using the panel of reliability indicators. (C) 1994 Wiley-Li
ss, Inc.