Print PageMost ecological studies focus on the changes in the means of response variables across groups or as a function of covariates and aim for low variability in order to maximize statistical power. However, patterns of variability also contain information about ecological processes. In particular, the patterns of variation in body size over time within a cohort of growing individuals can reveal whether there is significant among-individual variation in growth rates. Individuals in a cohort may grow consistently faster or slower than average (positive growth autocorrelation) due to genetic differences, phenotypic plasticity, or behavior syndromes. Previously, detecting such a pattern required data from marked individuals, but we have developed a method that only requires the cohort’s mean and variance through time.
Results/Conclusions
Based on simulations, our method can detect significant growth autocorrelation at moderate levels with moderate sized cohorts. For example, we get statistical power of 80% to detect growth autocorrelation levels of 0.5 (on a scale from 0 to 1) in a cohort of 100 individuals measured on 16 occasions. We discuss a range of feasible sampling designs for detecting growth autocorrelation in terms of time points and cohort size.
