Analysis of variance (ANOVA) developed by R. A. Fisher 100 years ago, brought a new age of field experimentation. ANOVA revolutionized field experimentation in modern agriculture, providing ways to minimize or avoid bias, improve precision and secure valid conclusions. While these conventional techniques remain widely used, new statistical methods, that take advantages of the increased computing power, have an increasing role in agricultural research. Nowadays the development and implementation of mixed models has had a great impact on agricultural research.

Designed experiments conducted by crop scientists often give rise to several random sources of variation. Examples are split-plot designs, series of experiments and repeated measurements taken on the same field plots. Data arising from such experiments may be conveniently analysed by mixed models. Despite the presence of both fixed and random effects in most crop production experiments, many crop researchers use of the conventional analysis of variance model, that provides correct analysis only if all the effects are fixed.

This paper points out the need for mixed model analysis, describes some features and properties of mixed model analysis to emphasize why it is more flexible and powerful than the conventional methods. The second part is to carry out detailed mixed model analysis for two specific examples (split-plot design and unbalanced data) to demonstrate such flexibility and power. We are convinced that crop scientists can produce valid and useful mixed model analysis if equipped with the appropriate software and an understanding of some basic principles.

Keywords: analysis of variance, fixed vs. random effects, mixed models, repeated measurements, split-plot design