Examples of possible questions
to the oral exam
The list is
not exclusive; it only gives examples of questions.
Several questions
will be posed to each student.
Some of the
questions may be supplemented with computer output and graphs.
?
Explain
what is meant by a counting process and its intensity process. Illustrate with
examples.
?
Explain
how a martingale may be derived from a counting process. What are the
predictable and optional variation processes of the martingale?
?
Explain
what is meant by a stochastic integral with respect to a counting process martingale,
and why the stochastic integral is itself a martingale. What are the
predictable and optional variation processes of the stochastic integral?
?
Explain
what is meant by independent right censoring.
?
Explain
what is meant by the multiplicative intensity model for counting processes and
give examples of situations that may be described by the multiplicative
intensity model.
?
Give
a motivation for the Nelson-Aalen estimator for the
multiplicative intensity model for counting processes and show that the estimator
is approximately unbiased. Derive an estimator for its variance.
?
Show
that the Nelson-Aalen estimator is approximately
normally distributed and use this to derive a log-transformed confidence interval
for the cumulative hazard.
?
How
are the survival function and the (cumulative) hazard rate defined for the
continuous case, and how are they related? How can the relations be generalized
to general distributions?
?
Give
a motivation for the Kaplan-Meier estimator and describe how it may be used to
estimate quartiles of the survival distribution. How can one derive confidence
limits for the quartiles?
?
Explain
the relation between the Kaplan-Meier and Nelson-Aalen
estimators.
?
Show
that the Kaplan-Meier estimator is approximately normally distributed and use
this to derive a log-log-transformed confidence interval for the survival
function.
?
Give
a motivation for the logrank test for two samples,
and describe alternative tests.
?
Describe
Cox's regression model and discuss the model assumptions.
?
Derive
Cox's partial likelihood and Breslow's estimator for
the cumulative baseline hazard.
?
Describe
?
Describe
situations where occurrence/exposure rates apply, derive the
occurrence/exposure rates and discuss their statistical properties.
?
Describe
situations where Poisson regression applies, and describe how the model fit may
be performed by "standard software"
?
Explain
what is meant by the proportional gamma frailty model. Derive the population
survival function and population hazard rate and interpret the results.