Scribd
Upload
2K views

Doing Bayesian Data Analysis With R and BUGS

This document provides an overview of a full-day tutorial on doing Bayesian data analysis with R and BUGS software. The tutorial will cover familiarization with R and BUGS, Bayesian concepts like uncertainty and Bayes' rule, Markov chain Monte Carlo methods, hierarchical models, and Bayesian analyses of linear regression, analysis of variance, power analysis and replication probability. Attendees are instructed to install the necessary free software before the tutorial. The instructor, John Kruschke, is an experienced teacher of Bayesian statistics and researcher in models of attention and learning.

Uploaded by

Octavio Martinez
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2K views

Doing Bayesian Data Analysis With R and BUGS

This document provides an overview of a full-day tutorial on doing Bayesian data analysis with R and BUGS software. The tutorial will cover familiarization with R and BUGS, Bayesian concepts like uncertainty and Bayes' rule, Markov chain Monte Carlo methods, hierarchical models, and Bayesian analyses of linear regression, analysis of variance, power analysis and replication probability. Attendees are instructed to install the necessary free software before the tutorial. The instructor, John Kruschke, is an experienced teacher of Bayesian statistics and researcher in models of attention and learning.

Uploaded by

Octavio Martinez
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Tutorial: Doing Bayesian Data Analysis with R and BUGS

John K. Kruschke ([email protected])


Department of Psychological and Brain Sciences, Indiana University
1101 E. 10th St., Bloomington, IN USA

Before arriving, install free software from this web site:


http://www.indiana.edu/jkkteach/TutorialCogSci2010.html

Keywords: Data analysis (Bayesian); Statistics (Bayesian);


Markov chain Monte Carlo; Bayesian models; Hierarchical
models

Topics
Familiarization with software: R, BRugs, BUGS. See installation instructions before arriving at the tutorial.
Uncertainty and Bayes rule: Application to the rational
estimation of parameters and models, given data.
Markov chain Monte Carlo: Why its needed, how it works,
and doing it in BUGS.
Hierarchical models: Flexibility for modeling individual
differences, group effects, repeated measures, etc.
Bayesian (multiple) linear regression: Bayesian inference
reveals trade-offs in credible regression coefficients.
Bayesian analysis of variance: Encourages thorough multiple comparisons, with no need for balanced designs.
Bayesian power analysis and replication probability:
Straight forward meaning and computation.

Bayesian data analysis is not


Bayesian modeling of mind
Data analysis involves generic descriptive models, such as
linear regression, without any necessary interpretation as cognitive computation. The rational way to estimate parameters
in descriptive models is Bayesian, regardless of whether or
not Bayesian models of mind are viable.

Why go Bayesian?
Scientists in fields from astronomy to zoology are making
Bayesian data analysis their standard operating procedure.
Figure 1 (humorously) suggests this trend. Bayesian data
analysis delivers many practical benefits:

15
10
5

Happiness and Success

This full-day tutorial shows you how to do Bayesian data


analysis, hands on. The software is free. The intended audience is graduate students and other researchers who want
a ground-floor introduction to Bayesian data analysis. No
mathematical expertise is presumed. If you can handle a few
minutes
R of summation notation like i xi and integral notation
like x dx, youre good to go. Complete computer programs
will be worked through, step by step.

An introduction to doing
Bayesian data analysis

Data with credible regression lines

50

100

150

Knowledge of Bayesian data analysis


Figure 1: Bayesian linear regression reveals many credible
lines, instead of a single best line. Also revealed are the
inherent trade-offs in estimated slope and intercept: When
the slope is steeper, the intercept is lower.
Bayesian methods permit model flexibility and appropriateness: Hierarchical models can be built easily to suit the
design of the experiment and the type of data measured.
Figure 2 shows a diagram of the model used for simple
linear regression. Such models can be easily extended to
capture individual differences, group differences, repeated
measures, for various data types.
Bayesian methods reveal credibilities of all combinations
of parameter values, unlike traditional analysis which has
only point estimates. As a simple example, Figure 1
demonstrates that Bayesian linear regression reveals correlations in estimated values of slope and intercept. Knowledge of these trade offs can be especially useful in applications involving multiple correlated predictors and other
realistic situations.
Bayesian methods encourage thorough data analysis in-

Before arriving, install free software from this web site:


http://www.indiana.edu/jkkteach/TutorialCogSci2010.html

cluding multiple comparisons, because there are no penalizing corrections for multiple comparisons as in p-value
based decisions. Bayesian methods instead create rational
shrinkage informed by the data.
Model flexibility allows conceptual transition from generic
descriptive models to domain-specific models wherein parameters serve psychometric purposes.
Bayesian methods permit different sample sizes in different
groups, and different sample sizes per subject, unlike traditional ANOVA which has troubles with unbalanced designs.
Bayesian methods allow data collection to stop at any time,
unlike p-value based decisions that require a pre-set stopping criterion, such as fixed sample size, and no peeking at
the data.
Bayesian hypothesis testing permits a principled way to assess evidence in favor of a null hypothesis, unlike NHST.
Bayesian methods allow cumulative science and use of
prior knowledge for leveraged inference when data are
sparse, unlike traditional methods.
Power and replication probability are straight forward to
estimate with Bayesian methods, but difficult to assess in
p-value based methods.
20th century methods, based on p values, have numerous
deep problems that are avoided with Bayesian methods.
The tutorial does not address all the points listed above,
but does illustrate many of them with examples from linear
regression and ANOVA. For a brief discussion of several benefits of Bayesian data analysis, along with a worked example,
and an emphasis that Bayesian data analysis is not Bayesian
modeling of mind, see Kruschke (2010c). For a lengthier exposition that explains one of the primary pitfalls of null hypothesis significance testing and has a discussion of Bayesian
null hypothesis testing, along with different examples, see
Kruschke (2010a).

Before arriving, install necessary software


Well be doing the analyses, so bring your notebook computer. There will not be internet access from the tutorial room,
so you must prepare your computer before arriving at the tutorial. Please visit the tutorial web site, framed at the top
of this page, before arriving at the tutorial. Follow the instructions on the web site to install the free software on
your computer.

Figure 2: Hierarchical model for Bayesian linear regression.


The data are denoted yi at the bottom of the diagram. The
model assumes that the data are generated probabilistically
from a normal distribution, governed by parameters 0 (the
intercept), 1 (the slope), and (the noise). Bayesian estimation keeps track of combinations of 0 , 1 , and that credibly
account for the data.

The instructor
John Kruschke has taught introductory Bayesian statistics
to graduate students for several years (and traditional statistics and mathematical modeling for over 20 years). He is
five-time winner of Teaching Excellence Recognition Awards
from Indiana University, where he is Professor of Psychological and Brain Sciences, and Adjunct Professor of Statistics. He has written an introductory textbook on Bayesian
data analysis (Kruschke, 2010b); see also the articles linked
above. His research interests include models of attention in
learning, which he has developed in both connectionist and
Bayesian formalisms. He received a Troland Research Award
from the National Academy of Sciences. He chaired the Cognitive Science Conference in 1992.

References
Kruschke, J. K. (2010a). Bayesian data analysis. Wiley Interdisciplinary Reviews: Cognitive Science. (Early view
available online.)
Kruschke, J. K. (2010b). Doing Bayesian data analysis: A
tutorial with R and BUGS. New York: Academic Press /
Elsevier Science. (To appear, November)
Kruschke, J. K. (2010c). What to believe: Bayesian methods
for data analysis. Trends in Cognitive Sciences. (In press)

You might also like