STAT3012 Lecture Notes - Lecture 18: National Institute Of Justice, Covariate, Dependent And Independent Variables
Lecture 18 – 2-way ANOVA
New concepts
✷2-way analysis of variance
✷Main effects
✷Interaction effects
Applied Linear Models: Lecture 18 1
find more resources at oneclass.com
find more resources at oneclass.com
New topic – Two-way analysis of variance
Theory – An additive two factor model
✷Suppose that there are two factors, A and B, which may be related to the
response variable.
✷The most straightforward way of modelling these factors is to assume that their
effects are additive.
✷This leads to what can be termed the main effects two-way ANOVA model.
✷The rationale for using the phrase main effects will become clear when we study
interactions later.
Applied Linear Models: Lecture 18 2
find more resources at oneclass.com
find more resources at oneclass.com
Theory – Main effects model for two factors
A (main effects) two-way ANOVA model with factors A and B is
Yijk =µ+αi+βj+ǫijk, ǫijk ∼NID(0, σ2)(1)
where i= 1, . . . , a,j= 1,,...,b and k= 1, . . . , nij.
Components
✷Yijk is the response for the kth unit at level iof factor A and level jof factor
B (i= 1, . . . , a, j = 1, . . . , b, k = 1, . . . , nij).
✷α1, . . . , αaand β1, . . . , βbare parameters describing the ‘main effects’ of A and
B respectively.
✷Necessary to constrain parameters α1, . . . , αaand β1, . . . , βb.
⇒Treatment constraint or sum constraint.
Applied Linear Models: Lecture 18 3
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Suppose that there are two factors, a and b, which may be related to the response variable. The most straightforward way of modelling these factors is to assume that their e ects are additive. This leads to what can be termed the main e ects two-way anova model. The rationale for using the phrase main e ects will become clear when we study interactions later. Theory main e ects model for two factors. A (main e ects) two-way anova model with factors a and b is. Yijk = + i + j + ijk, Ijk n id(0, 2) (1) where i = 1, . Yijk is the response for the kth unit at level i of factor a and level j of factor. , b are parameters describing the main e ects" of a and. Theory writing the model as a multiple linear regression.