PDF version of this document
Epub version of this document
Preface
1
Background and R Setup
1.1
A note on R functions and usage style
1.2
Resources
2
Import data set and do Exploratory Data Analysis
2.1
Data Import
2.2
Numerical Exploratory Data Analysis
2.3
Graphical EDA
2.4
Graphical EDA with
ggplot2
: boxplots and bar graphs with error bars
2.5
What are the proper error bars for a plot of a within-subjects design?
2.5.1
The method of Loftus and Masson (1994)
2.5.2
Alternatives to the Loftus and Masson approach
2.5.3
Franz and Loftus 2012 approach for pairwise comparisons
2.6
Conclusions on error bars and plotting graphs and specific errors for analytical/orthogonal contrasts
2.7
Conclusions from EDA
3
Traditional Approaches to One Factor Repeated Measures Designs
3.1
The traditional “univariate” GLM approach to the repeated measures problem and the Multivariate approach
3.1.1
Method I: Use of the
aov
function
3.1.2
Method II, a multivariate linear model
3.1.3
Method III, also a Multivariate Linear Model but using
Anova
from
car
3.1.4
Method IV: Implement the univariate analysis with
ezanova
from the
ez
package
3.1.5
Method V: Implement the univariate analysis with the
afex
package
3.2
Commentary on the Univariate/Multivariate methods
4
Analytical Contrasts
4.1
Partitioning the omnibus effect into orthogonal contrasts
4.2
Contrasts with the
aov
model object
4.3
Contrasts with the Method III “mlmfit” object
4.4
Recall how to “manually” implement contrasts for repeated factors.
4.4.1
Create a plot of these contrasts and their specific errors
4.5
Working with
afex
and
emmeans
for contrasts and pairwise comparison follow ups
4.5.1
Use of
emmeans
on
aov
fit objects.
4.6
Commentary on Contrast Analysis with the Univariate/Multivariate methods
5
Linear Mixed Models
5.1
Basic LMM Analysis using
lme
5.2
Basic LMM Analysis using
lmer
5.3
A modeling approach
5.4
Alternative covariance structures
5.4.1
Contrast evaluation from the alternative covariance structure model
5.5
Post hoc pairwise comparisons and planned/orthogonal contrasts
5.5.1
Use of
emmeans
on the lme model
6
Bayesian Approaches
6.1
Bayes Factor analysis
6.2
Contrasts with BayesFactor methods?
7
Robust and Resampling Methods
7.1
Robust Tests
7.2
Resampling Methods
7.2.1
Howell’s permutation approach
7.3
Bootstrappping
8
Nonparametric Approaches
9
Trend Analysis and A Pre-Post Design
9.1
Trend Analysis
9.1.1
Import the data and perform EDA.
9.1.2
The omnibus analysis for the McCarthy Scale by Age dataset
9.1.3
Implementing Orthogonal Polynomial Trend Analysis
9.1.4
Specific error terms for the Trend contrasts
9.1.5
Visualization of the trend contrasts
9.1.6
Conclusion on Univariate approach to trend contrasts and specific error terms
9.1.7
Bayes Factor evaluation for trend.
9.1.8
Linear Mixed Effects Modeling and Trend
9.1.9
General conclusion regarding trend for the Age factor.
9.2
A Pre-Post design
9.2.1
Import the data and perform EDA
9.2.2
The omnibus analysis: Why?
9.2.3
Building contrasts
9.2.4
Visualization of the contrasts
9.2.5
Bayes Factor values for the pre-post contrasts
References
10
Reproducibility and history
Published with bookdown
One factor Repeated Measures ANOVA with R
References