TY - JOUR
T1 - Statistical options for the analysis of in vitro gas production profiles illustrated using rumen liquor as the inoculum
AU - Dhanoa, Dan
AU - Sanderson, R.
AU - Lister, Sue
AU - Mauricio, Rogerio Martins
AU - López, Secundino
AU - Ellis, Jennifer
AU - Powell, Chris
AU - France, James
N1 - Funding Information:
This work was funded in part by the Canada Research Chairs program, grant number 045867 (Natural Sciences and Engineering Research Council of Canada, Ottawa).
Publisher Copyright:
Copyright © The Author(s), 2023. Published by Cambridge University Press.
PY - 2023/11/29
Y1 - 2023/11/29
N2 - The use of repeated measures analysis of variance (ANOVA) options for the analysis of in vitro ruminal fermentation gas production profiles is illustrated. Because of the different variances and covariance structures among profile observations, ordinary ANOVA for more than two-time points is not recommended. To mitigate this problem, the Greenhouse-Geisser epsilon correction can be applied to reduce the degrees of freedom, inflated by violation of the sphericity assumption, for F ratio probability calculations. After this correction, the Box-Greenhouse-Geisser ANOVA (modified ANOVA) layout appears similar to the layout of a split-plot design ANOVA with whole plots divided into subplots (incubation time). Any F tests in the main plot part are valid but F tests involving the time factor from the subplot part need modification because time factor, by its very nature, cannot be allocated at random. Application of multivariate ANOVA, distance multivariate ANOVA, ante-dependence and mixed model analysis are also considered. All these options lend themselves to wide application in the applied biological sciences.
AB - The use of repeated measures analysis of variance (ANOVA) options for the analysis of in vitro ruminal fermentation gas production profiles is illustrated. Because of the different variances and covariance structures among profile observations, ordinary ANOVA for more than two-time points is not recommended. To mitigate this problem, the Greenhouse-Geisser epsilon correction can be applied to reduce the degrees of freedom, inflated by violation of the sphericity assumption, for F ratio probability calculations. After this correction, the Box-Greenhouse-Geisser ANOVA (modified ANOVA) layout appears similar to the layout of a split-plot design ANOVA with whole plots divided into subplots (incubation time). Any F tests in the main plot part are valid but F tests involving the time factor from the subplot part need modification because time factor, by its very nature, cannot be allocated at random. Application of multivariate ANOVA, distance multivariate ANOVA, ante-dependence and mixed model analysis are also considered. All these options lend themselves to wide application in the applied biological sciences.
KW - ante-dependence analysis
KW - modified analysis of variance
KW - multivariate analysis of variance
KW - multivariate distance methodology
KW - residual maximum likelihood
UR - http://www.scopus.com/inward/record.url?scp=85179042511&partnerID=8YFLogxK
U2 - 10.1017/S0021859623000588
DO - 10.1017/S0021859623000588
M3 - Article
AN - SCOPUS:85179042511
SN - 0021-8596
SP - 686
EP - 695
JO - Journal of Agricultural Science
JF - Journal of Agricultural Science
ER -