Use of a Split-Plot Analysis of Variance Design for Repeated-Measures Fishery Data

Abstract

We present a repeated-measures split-plot analysis of variance (ANOVA) design that can be used to analyze fisheries data collected throughout time from fixed sampling stations. A high percentage (97%) of authors who published papers in the Transactions of the American Fisheries Society and the North American Journal of Fisheries Management during the past 10 years did not use repeated-measures ANOVA for analyzing their data when this technique was appropriate. When data are collected from fixed stations throughout time, the assumption that independent samples are randomly collected from the population is violated for simple ANOVA. Fixed stations may display population characteristics specific to that location, and environmental change or a management practice may influence stations differently. In addition, multiple measurements collected at a fixed station throughout time will likely be correlated. For a dependent variable of interest, a repeated-measures split-plot design will partition the variation due to (1) a treatment or manipulation, (2) sampling stations, (3) the temporal correlation that may occur among repeated measurements, and (4) the potential interactions among these class variables. This statistical approach will provide increased accuracy and likely provide investigators with new insights on the responses of fish populations to these variables.