Title of Workshop: Multilevel Modelling for Missing Data in Longitudinal Studies
Date:: Thursday 11th November, 2010
Location: Postgraduate Statistics Centre, Lancaster University.
Presenter: Professor Don Hedeker, Professor of Biostatistics at the University of Illinois
Sponsor: ESRC National Centre for Research Methods Lancaster-Warwick-Stirling node.
Description
In this 1-day workshop, attendees will learn about the use of multilevel or mixed models for analysis of missing data in longitudinal studies. The focus will be on data analysis, with direct application illustrated using standard statistical software (i.e., SAS). Some of the topics to be covered include:
(a) missing data mechanisms in longitudinal studies,
(b) pattern-mixture and selection models for non-ignorable missing data; also simple multiple imputation methods for non-ignorable binary outcomes will be described.
1. Missing data in longitudinal studies are common, and so methods for dealing with missing data will be discussed. First, the various missing data mechanisms (missing completely at random, missing at random, not missing at random) proposed by Rubin in the 1970s will be defined. Methods for testing the missing completely at random assumption will be illustrated for longitudinal data.
2, For situations in which the missing at random assumption does not seem tenable, sensitivity analysis will be detailed. In particular, non-ignorable pattern-mixture and selection models will be described and illustrated. Though these models can be complex, it will be shown how they can be estimated using SAS PROC NLMIXED. Finally, for dichotomous outcomes, a simple multiple imputation approach will be described for non-ignorable missing data.
Cost: £30 postgraduate students
£60 staff/researchers in academic institutions
£220 other individuals.
The cost includes notes, lunches, tea and coffee breaks but not accommodation. Attendees needing overnight accommodation can see a range of accommodation choices at:
http://shortcourses.maths.lancs.ac.uk/accommodation
Timetable:
09.30 - 10.00 Registration
10.00 - 11.15 Lecture 1
11.15 - 11.45 Break
11.45 - 13.00 Lecture 2
13.00 - 14.00 Lunch
14.00 - 15.15 Lecture 3/Practical
15.15 - 15.45 Break
15.45 - 17.00 Lecture 4/Practical
17.00 - 17.30 Discussion & concluding remarks
17.30 Departure
For further information and registration please see : http://socialstatistics.org/workshops/longitudinalstudies2010
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|