Invited review: Assessing experimental designs for research conducted on commercial dairies
- 1 January 2009
- journal article
- review article
- Published by American Dairy Science Association in Journal of Dairy Science
- Vol. 92 (1) , 1-15
Abstract
Because of increasing constraints placed on conducting large studies at universities, more research is being conducted on commercial dairies, thereby raising some implications for experimental designs and data analysis. For example, experimental units are often specified to be pens of animals in on-farm studies, thereby requiring that at least 2 pens be used per treatment group in a single-dairy study. Even when treatments are compared within pens, the precision of inference on treatment differences is still primarily limited by the number of pens in the study, rather than the number of cows per treatment in each pen. Other challenges with on-farm studies include proper blocking and randomization of cows or pens to treatments. On the other hand, multiple farm studies are attractive, because they facilitate a broader scope of inference on treatment effects across a wider range of management or climatic conditions and genetic backgrounds compared with single-site university studies. Furthermore, studies based on multiple farms or multiple pens within a single large farm can facilitate greater power for treatment comparisons on binary reproduction or health responses than can be achieved at a smaller research herd. Because quantitative geneticists have been analyzing commercial dairy data for decades, they have developed useful data analysis techniques that should be harnessed to facilitate even greater statistical scope and power for on-farm studies.Keywords
This publication has 29 references indexed in Scilit:
- Using genomic biology to study liver metabolism*Journal of Animal Physiology and Animal Nutrition, 2008
- What Can Genome-Wide Association Studies Tell Us about the Genetics of Common Disease?PLOS Genetics, 2008
- SNP discovery and allele frequency estimation by deep sequencing of reduced representation librariesNature Methods, 2008
- The Many Faces of ReplicationCrop Science, 2006
- An evaluation of bivariate random‐effects meta‐analysis for the joint synthesis of two correlated outcomesStatistics in Medicine, 2006
- A Unified Approach to Mixed Linear ModelsThe American Statistician, 1991
- Best Linear Unbiased Prediction in the Generalized Linear Regression ModelJournal of the American Statistical Association, 1962
- Best Linear Unbiased Prediction in the Generalized Linear Regression ModelJournal of the American Statistical Association, 1962
- The Estimation of Environmental and Genetic Trends from Records Subject to CullingBiometrics, 1959
- Estimation of Variance and Covariance ComponentsBiometrics, 1953