The Analysis of Experiments on Growth Rate

Abstract

A method is presented for the analysis of experiments in which successive measurements of the same quantity are made on the same organism at equal intervals of time. The specification and estimation of treatment effects are discussed and demonstrated by a numerical example. When a treatment effect can be specified by a curve of degree p passing through the origin, curves of degree p are fitted to the observations and 2 estimates are obtained of each coefficient of the curve specifying the treatment effect. These can be combined by a method giving minimum variance. Thus when the treatment effect is linear in time, straight lines are fitted to the data and two estimates of the linear coefficient specifying the treatment effect are combined with minimum variance to give an improved estimate. Covariance analysis on initial observations is also discussed.