Centre for Applied Social Surveys
SHORT COURSE: EVENT HISTORY ANALYSIS:
Supported by the ESRC
Mac McDonald and Fiona Steele
16-18 September 1998
University of Southampton
CASS is an ESRC Resource Centre hosted by
SCPR and the University of Southampton,
with the University of Surrey
To provide a practical introduction to discrete-time methods for the
analysis of event histories as applied to survey data
To enable participants to analyse their own event history data using
Event histories and issues in their analysis
Discrete-time approach to hazard modelling
Data management for the analysis of event histories using SAS
Modelling the time to a single event
Competing risks for multiple types of event
Multilevel models for modelling the times to recurrent events
Models including unobserved heterogeneity
The course will introduce the principles and practice of event history
analysis for survey analysts using discrete-time models. These models
include logit, multinomial logit models and multilevel versions of
these models. Many illustrative examples taken from the instructors
own research will be provided, especially their recent analyses of
survey data. The course will have a strong practical emphasis, with
regular computer-based workshop sessions at which the participants
will work through examples and exercises using the Windows version of
MLn. The workshops aim to provide practical experience in the
preparation of data for a discrete-time event history analysis, model
fitting and interpretation of results. The course will also discuss
other software packages which could be used for discrete-time event
history analysis, e.g. GLIM, SABRE, EGRET and BUGS as well as their
advantages and disadvantages.
The course is aimed at researchers who need to model event history
data, especially those in the social, economic and medical sciences.
The methods to be taught will be useful for survey analysts who are
dealing with surveys that have collected event histories (eg birth
histories, employment histories etc). A good understanding of
standard statistical methods of regression analysis and logistic
models for binary responses will be assumed. some familiarity with
event history analysis may be useful, but is not essential. Some
familiarity with SAS and MLn is useful, but not essential.
Participants are very welcome to bring their own datasets and will
have the opportunity to speak to the instructors about their analysis.
Participants intending to bring their own data should contact Rebecca
Bundock well in advance of the course so that details of the
appropriate data format can be supplied.
Participants will receive copies of overhead transparencies and
handouts for the computer-based workshops.
John `Mac' McDonald is Senior Lecturer in Social Statistics at the
University of Southampton. His general research interests are in
demography and social statistics. Particular research interests
include categorical data analysis and discrete-time event history
models with unobserved heterogeneity. He has presented workshops on
the analysis of social science data in several countries.
Fiona Steele is a Lecturer in Statistics and Research Methodology at
the London School of Economics. Her research interests are in the
application of categorical data analysis, multilevel models and
discrete-time event history models in demography. She has presented
computer workshops on the multilevel analysis of categorical data
using MLn to researchers from developing countries.
The standard fee is stlg240. Because of ESRC support, a number of
bursaries are available which reduce the fee to stlg65. Applicants for
bursaries must be full-time students, UK academic staff or ESRC-funded
University of Southampton, Highfield, Southampton, SO17 1BJ.
Telephone: 01703 593048.
The course will begin with registration and coffee at 10.30am 14
September and end at 12.30pm 16 September 1998. Participants are
welcome to stay on after the end of the course to try out the methods
they have learnt on their own data and/or to discuss with the
instructors how to analyse their data.
A good understanding of standard methods of regression analysis and
logistic models for binary responses will be assumed. A good
textbook, covering all the background required, is:
Hosmer, D.W. and Lemeshow, S. (1989) Applied Logistic Regression.
A more elementary account is provided by:
Kleinbaum, D.G., Kupper, L.L. and Muller, K.E. (1997) Applied
Regression Analysis and Multivariable Methods, 3rd Ed., Wadsworth.
A 87 page introduction to event history analysis, which covers both
discrete-time and continuous-time methods, is provided by:
Allison, P.D. (1984) Event History Analysis. Sage.
Please indicate on the registration form if you would like to be sent
details of local accommodation. Participants are left to book their
own accommodation according to individual needs.
Deadlines and Refunds
Course places are limited and early completion of this form is
recommended. Payment must be made by 7 days before the course. No
refunds can be made for cancellations made later than 7 days before
the start of the course.
Event History Analysis: Discrete-time
14-16 September 1998
Please complete this form and return to:
Department of Social Statistics
University of Southampton
Southampton SO17 1BJ
(Fax: 01703 593846 Tel: 01703 593048 E-mail:
[log in to unmask])
Address for Correspondence:
Please indicate the nature of any data which you would like to bring
to the course:
(1) Standard fee
stlg240 ____________ OR
(2) Reduced fee with ESRC bursary
You must be a registered
university student, UK academic
staff or ESRC funded researcher stlg65 ____________
Please provide details:
Please indicate either:
Cheque enclosed payable to University of Southampton
Send invoice to: _______________________________________________
Any special dietary or other requirements: