The Retest Reliability of the Observation of Certain Traits

Publisher Website
Google Scholar
Abstract
This paper presents some of the methodological problems of retest reliability of traits commonly used in genetic studies. Measures of repeatability of observations on 8 dichotomous traits were obtained. These included chi-square analysis for independence between original test and retest, chi-square for the symmetry of errors in the 2 trials, and estimates of lower and upper bounds for the probability of an observation falling into its "true" category in an infinite number of trials. All 8 variables showed high levels of repeatability. Tasting PTC (phenylthiocarbamide) gave indications of a possible asymmetric change from not tasting in the 1st to tasting in the 2nd trial. Possible sources of unreliability are discussed.

This publication has 0 references indexed in Scilit: