Exam 1 Study Guide
Setting up the problem in R for analysis by
WinBUGS/JAGS
WinBUGS likes things simple. For instance WinBUGS accepts only numeric
data. Character data and factors are not allowed so factors need to be converted
to a set of dummy variables

Lecture 12
Analysis of Covariance
To illustrate the use of WinBUGS to fit statistical models I introduce a standard
statistical model that was historically known as analysis of covariance. Analysis
of covariance is just a regression model in which the mai

Lecture 11
Fitting count models as generalized linear models
The Poisson and negative binomial models that we've fit using maximum
likelihood are examples of Poisson regression and negative binomial
regression. Poisson regression is also a type of general

Lecture 10
Calculating profile likelihood confidence intervals by hand
As we just saw when we fix all the parameters in the log-likelihood at their
maximum likelihood estimates except for one, we end up with a log-likelihood
function that is a function of

Lecture 9
Profile likelihood confidence intervals
Profile likelihood confidence intervals were introduced in lecture 8. There we saw
that a (1)% profile likelihood confidence interval for a parameter is
constructed by first determining the value of the lo

HW 2
1.
If some of the observed values of the response variable are zero, we have
to modify our protocol slightly because the logarithm function is
undefined at zero.
2.
The standard approach is to add a small constant c, e.g. c = 1, to all
values before

HW 1
1.
When using log-likelihood or AIC to compare models all the models
being compared must use exactly the same responseall y, all log y, etc.
2.
Two models are fit to the same count data set using maximum
likelihood.
3.
In the first case the model is

Final Exam Study Guide
Comparing Bayesian and frequentist results
With an entire posterior distribution at our disposal we have to choose what we
want to call the Bayesian point estimate. Obvious choices are the mean or the
median of the posterior distrib

Exam 2 Study Guide
Fig. 6 WinBUGS report on the model
1. data = ipo.data The first argument to bugs is the data object that
2.
3.
4.
5.
6.
was created above listing the variable names.
inits = ipo.inits The second argument is the function that sets
the in

Quiz 4 Study Guide
Fitting species area curves to the Galapagos Islands
plant richness data set
Johnson and Raven (1973) provide data on plant species richness for 29 islands
in the Galapagos Islands archipelago. Their paper was presented as a rebuttal to