B1328 - Developing and disseminating robust methods for handling missing data in epidemiological studies - 30/03/2012
B number:
B1328
Principal applicant name:
Prof Kate Tilling (Univeristy of Bristol, UK)
Co-applicants:
Prof Jonathan Sterne (Not used 0, Not used 0), James Carpenter (Not used 0, Not used 0), Prof Mike Kenward (Not used 0, Not used 0)
Title of project:
Developing and disseminating robust methods for handling missing data in epidemiological studies
Proposal summary:
Aims:
To examine the impact of newly developed methods for missing data.
Methods:
In particular, new methods have been developed to maximise efficiency when covariate data are MNAR (and therefore complete case unbiased and MI biased). In the first instance, we would like to analyse a sample of 2000 women from ALSPAC, modelling birthweight (outcome) on weight before pregnancy (self-report), adjusting for maternal age, parity, education and smoking and sex of offspring. Data on estimated pre-pregnancy weight (from GWG models) will be used to show that data on pre-pregnancy weight are MNAR.
Date proposal received:
Thursday, 29 March, 2012
Date proposal approved:
Friday, 30 March, 2012
Keywords:
Bias, Methods
Primary keyword: