Observed or phenotypic trends in animal performance can be readily quantified from information collected from research or field populations. Phenotypic performance is determined by the collective impact of systematic effects that vary by trait, but may include herd, year, sex, and age; additive genetic effects; and a remainder that is referred to as the lack-of-fit or unexplained residual. It is of interest to partition observed performance into these respective components to determine the extent to which genetic or environmental trends or both are responsible for any observed phenotypic trends. An animal breeding approach to separate these components from field data involves the use of a linear model that includes fixed effects for systematic terms and random effects for genetic and residual contributions. The fitted random effects are predicted using a shrinkage estimator known as BLUP that relies only on a translation invariant subset of the field data that does not involve the unknown fixed effects. Fixed effects can then be estimated by adjusting observations for estimates of the random effects. Reliable estimation of trends using this approach requires that relevant fixed effects are recorded, cohorts representing different fixed effects classes are genetically related or connected, and that any records used as the basis for selection in the population are included in the data set.