As kappa corrects for chance agreement among multiple coders, use of kappa is likely to underestimate reliability (Steinijans, Diletti, Bomches, Greis, & Solleder, 1997).
The interval after which the second version is to be given is a relevant decision; a too long interval may underestimate reliability by allowing actual changes in patients' degree of disability to occur, while a too short interval may overestimate it because of recall bias.
The new ICCs underestimated observer reliability compared with previously published results (intraclass ICCs lower by 0.01-0.02 and interclass ICCs lower by 0.03-0.10).
However, because α underestimates reliability when the items of a scale are not tau-equivalent (ie, have unequal factor loadings) and because the assumption of tau-equivalence was not met by most of the new OAV scales, ρSEM was slightly higher than α in these scales.
According to Gherunpong et al. [ 59], alpha is not a perfect indicator of reliability, as it tends to underestimate the reliability of multidimensional scales and because lower values can be expected from health-related measures.
The Cronbach's Alpha level reached (0.65) was below the recommended level of 0.70 [ 21]; however, as cited by Gherunpong et al. [ 20], alpha is not a perfect indicator of reliability as it tends to underestimate the reliability of multidimensional scales, and lower values can be expected from health-related measures.
Therefore, the reliabilities reported in this study might underestimate the reliability for a random group of bulls, especially for strongly selected traits.
This indicates that the allelic R measure provided by Beagle gives a good indication of imputation reliability in general, although in specific cases it may severely underestimate imputation reliability.
For reliability assessment, in contrast to the conventional Cronbach's alpha (which tends to underestimate the reliability of the total score), we calculated model based coefficients omegas[ 25].
However Cronbach's alpha coefficient will underestimate the reliability of SF-36 because this scale has eight domains which showed somewhat heterogeneity [ 27, 28].
