
Like many psychology students of recent times I was brought up in
pointandclick style, patting myself on the back whenever output
magically appeared and pretending I knew what I was talking about
with regard to such legerdemain. What I came to learn was
that simply getting output is something wholly different than doing
statistical analysis, and that to do a good analysis requires
getting one's hands dirty.
However, I also don't think it's necessary to become a computer
programmer outright (or at least shouldn't be). One can take
what is necessary and use it in an applied fashion, doing much
better analyses and more efficient ones. As such code is
provided for lab content below so that you can spend more time on
understanding the analysis, with some additional stuff for R, my
preferred stat program. The SAS will not be great as it is my preference to
not use it if I can help it. I find I like
it as a program better than SPSS in terms of what it can do, but
just about everything regarding it is 'ugly' in appearance. Neither syntax
is as flexible as R, and almost always requires much more code to do
the same thing. Also, if you come across a more efficient
way to do something definitely let me know and I'll put it up here.
For lab specific code click here.
Link breakdown
Purpose: General statement on what's going on, things to look out
for etc.
Stat program name: the page with all the code
Marks within cells: specific code relevant to purpose
Code relevant to labs

Purpose 
R 
SPSS 
SAS 
Entering Data 
X 
X 
X 
Importing Data 
X 
X 
X 
Manipulating Data 
X 
X 

Frequencies 
X 
X 
X 
Central Tendency and Variability 
X 
X 
X 
Summary Statistics 
X 
X 
X 
ttests 
X 
X 

Correlation 
X 
X 
X 
Simple Regression 
X 
X 

Power and Effect Size 
X 
N/A 

Datasets
 Most datasets used in class and lab
demonstration may now be accessed
here.
Text files automatically load into the browser,
just access and save when you see it on your
screen. For SPSS files, download when the
dialog box presents itself to.
Note that by default Rcmdr imports the factor
labels (i.e. the words rather than the numbers
if there is a choice). If you prefer them
or need them to stay numeric, uncheck that box
that says 'convert value labels to factor
levels' during the import process.


Howell datasets.

Karl Wuensch's datasets.

Useful R packages/functions
 General

Pdf for installing R
at home

Quick R: an R website for SAS/SPSS/Stata Users
R wiki

R
web interface and notes

Using R in Psychological Research

ANOVA and Regression in R (book)

R
graphics
R colors

R for SAS and SPSS Users

Comparing those three packages

 Functions
 General
 Cran Task Views
 Wilcox,
Workshop
 Fox,
Robust appendix
 Basics

t.test

cor

aov
 Regression

General regression related functions

Robust regression functions
 Bootstrapping

boot

bootstrap

simpleboot

 Packages

BMA:
Bayesian model averaging

Boot,
simpleboot:
basic bootstrap

Design:
useful functions for
anova/regression, validation

Hmisc:
general purpose

lme4 and
nlme:
mixed effects/multilevel
modeling

ltm:
Item response theory

MASS:
evolved from Venerable and
Ripley's Modern Applied Statistical Analysis

MBESS:
Behavioral Sciences
specific, tons of effect size goodies
 Mike's Miscellaneous

Multtest:
multiple comparisons in
ANOVA

prettyR:
make descriptive output
'pretty'

Psy:
psychometric stuff

Psych: includes
functions for personality and
psychological research

QuantPsyc:
for testing moderation
and mediation

Quantreg:
quantile regression

Rcommander:
menu system


Rcmdr.HH

Relaimpo:
measures relative
importance of variables in multiple regression
in a meaningful way

Robust
library that makes it
easy,
notes

Robustbase:
more robust
regression

sem:
structural equation modeling

Wilcox library (a
host of robust functionality)
 Rallfunv8
 For Mike Miscellaneous and Wilcox libraries, right click and save them in your main R folder, and whenever you
want to use the functions within, from the menu
File/Source R code, then go find your file. Or at
the command line

 source("Rallfunv1v8")
 Where the quotes have the address of the
file location. To see what's in them just
open them up like any other script. See the Wilcox link for more
info. With the Mike misc, Macs may have an
issue sourcing the file, but if you just cut and
paste the functions (just click the link rather
than save) at the console that will work until I
can figure out the issue.
Help file for projects


 General psych/social sciencerelated stuff in
R

Psychometrics: IRT, SEM etc.

Social Science: general

Multivariate

Robust

Some examples of code I find useful:

Cohen's d, R^{2}, Intervals for them

Plotting Cohen's d

False Discovery Rate

Whizbang intervals

Robust correlation and simple regression

Bootstrap ttest
(requires Wilcox
libraries above)

Robust regression (general)

Simple Mediation

Testing multivariate normality

All subsets regression:
note that
it is better accomplished in the Rcommander menu
system if you install the Rcmdr.HH library (see
above). You even get a nice graphic.

Validation of a linear model

Path analysis of lecture notes example

Creating a dataset

Plotting interactions in regression

 Other
 Simple effects SPSS
syntax
 R, SPlus, SAS and SPSS scripts for
CIs for effect sizes (Smithson)


