This is a statistics course for postgraduate students in the Faculty of Agricultural and Environmental Sciences at McGill University. The orientation of this course is to the application of statistics in the biometrical field; the use of various statistical methods related to Analysis of Variance in the interpretion of experimental research data. The SAS statistical package will be used extensively throughout this course, both to demonstrate how to carry out an analysis and to show the proper SAS statements.

Course description, as given in the McGill Graduate Programme Book

Principles of linear models, multiple regression equations and classification models. Introduction to Analysis of Variance and common statistical designs used in agricultural and environmental sciences. Emphasis on balanced and unbalanced designs and data structures; their analysis and tests of statistical significance.

Prerequisite: AEMA 310 Statistical Methods I, or an equivalent undergraduate statistics course.

Course outline and topics to be covered

Quantitiative traits, what they are and are not, examples

A brief introduction to matrices

Review of simple linear regression (from Statistical Methods I)

Multiple Regression

..Assumptions of the model

..Linear model in statistical, biological and biometrical terms

Least Squares

..Derivation of Least Squares, how we get the equations that we use

..Obtaining a solution

..Estimated model equations

..Parameter estimates and standard errors

..Sums of Squares and ANOVA

t-test and confidence intervals

Linear and quadratic regressions and interactions

Curve fitting

Correlations, when is a correlation appropriate?

..Partial correlations

..Confidence interval and tests of significance

Classification models

One-way classification

..Assumptions

..Linear model

..Obtaining a solution

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

..Multiple comparisons

..Homogeneity of variance and Normality

Two-way classification

..Assumptions

..Linear model

..Obtaining a solution

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

Type I and Type III Sums of Squares

Gains in efficiency due to use of 'blocks'

Subsampling and nested, hierarchical models

Factorial models

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

Latin Square models

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

Analysis of Covariance

..What is estimable

..Sums of Squares and ANCOVA

..Treatment differences

Split Plot models

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

Fixed effects or Random effects

Mixed models

..Assumptions

..Linear model

..Obtaining a solution

..What is estimable

..Sums of Squares and ANOVA

..Treatment differences

R.I. Cue, 2010 May 4