Statistical aspects of aquaculture research: pond variability and pseudoreplication |
| |
| Authors: | J Riley P Edwards |
| |
| Institution: | Statistics Department, IACR-Rothamsted, Harpenden, UK;Agricultural and Aquatic Systems Program, School of Environment, Resources and Development, Asian Institute of Technology, Bangkok, Thailand |
| |
| Abstract: | Pond experiments with one treatment per pond are discussed for their sources of variability. Replication of treated ponds, randomization of treatment allocation to ponds and the importance of independence between ponds are emphasized for efficient estimation of both treatment and inherent pond differences. Appropriate models are given for the analysis of data for pond experiments with whole-pond treatments. Corresponding analyses of variance are given to illustrate the estimation of differences between treatments, between ponds and between fish within ponds. Appropriate significance tests for treatment effects are demonstrated to avoid pseudoreplication, the incorrect use of within-pond variation. Using a fertilizer experiment, simple graphical methods are presented for examining between- and within-pond variation to identify ponds with unusual data, and to determine adequacy of design and appropriate analyses. |
| |
| Keywords: | |
|
|
