... immediately apparent. In the calculation of reliability, one needs to treat I and II as repeated obser-vations of the same scale, in order to estimate true and error variance. So it makes no sense to then ... all variation is error.If you want to look at reliability or discrimination, youmust separate error variance – differences between indi-vidual observations and the average for each individual,from ... calculations it's0.7 04) . This comes about because, to estimate the reliabil-ity, we use both sets of observations to estimate the aver-age score for each individual, then calculate the...