Please enable JavaScript.
Coggle requires JavaScript to display documents.
2 Two-way within-subjects ANOVA Split plot (Split plot anova…
2
Two-way within-subjects ANOVA
Split plot
2 way WS
Assumptions
Same as 1way
dep=quantitative
everything normally distributed
Steps
first interaction
sig
simple effects
not sig
main effects
Design
WS facotrs are crossed
each person is observed in each combination of factors
Laboratory experiments
listwise deletion
if participant doesn't do all conditions, delete completely
otherwise non-orthogonal
If WS factor has 2 levels
Multi- and uni-variate equal
Sphericity always met
Split plot anova
combines within and between subjects factors
Add person as a random factor
Assumptions
Dependent = quantative
Split-plot for groups
BS should be independent
WS should be dependent
Normality
for sample and difference scores
Sphericity
same variance for difference scores
Equal covariance matrices
equal population matrices
Box's test
not robust if no normality
Also called mixed design
Field experiments
Univariate method
2-way WS=3-way BS
A and b fixed
Person = random factor
Every fixed effect must be interacted with random effect
Assumptions
Sphericity
for each effect
Met if ws factor has 2 levels
Multivariate method
Transform repeated measure into
WS average
K-1 contrasts