?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.relation=http%3A%2F%2Feprints.ncrm.ac.uk%2F1776%2F&rft.title=Strategy+for+modelling+non-random+missing+data+mechanisms%0D%0Ain+observational+studies+using+Bayesian+methods&rft.creator=Mason%2C+Alexina+&rft.creator=Best%2C+Nicky&rft.creator=Richardson%2C+Sylvia&rft.creator=PLEWIS%2C+IAN&rft.subject=3.6+Nonresponse&rft.subject=5.7+Longitudinal+Data+Analysis&rft.description=Observational+studies+are+notoriously+full+of+non-responses+and+missing+values.+Bayesian+full%0D%0Aprobability+modelling+provides+a+%C2%B0exible+approach+for+analysing+such+data%2C+allowing+a+plausible%0D%0Amodel+to+be+built+which+can+then+be+adapted+to+carry+out+a+range+of+sensitivity+analyses.+In%0D%0Athis+context%2C+we+propose+a+strategy+for+using+Bayesian+methods+for+a+%60statistically+principled'%0D%0Ainvestigation+of+data+which+contains+missing+covariates+and+missing+responses%2C+likely+to+be+non-%0D%0Arandom.%0D%0AThe+%C2%AFrst+part+of+this+strategy+entails+constructing+a+%60base+model'+by+selecting+a+model+of%0D%0Ainterest%2C+then+adding+a+sub-model+to+impute+the+missing+covariates+followed+by+a+sub-model%0D%0Ato+allow+informative+missingness+in+the+response.+The+second+part+involves+running+a+series+of%0D%0Asensitivity+analyses+to+check+the+robustness+of+the+conclusions.+We+implement+our+strategy+to%0D%0Ainvestigate+some+typical+research+questions+relating+to+the+prediction+of+income%2C+using+data+from%0D%0Athe+Millennium+Cohort+Study.&rft.publisher=Imperial+College+London&rft.date=2010&rft.type=Working+Paper&rft.type=NonPeerReviewed&rft.format=application%2Fpdf&rft.language=en&rft.identifier=http%3A%2F%2Feprints.ncrm.ac.uk%2F1776%2F1%2FStrategySubmitted.pdf&rft.identifier=++Mason%2C+Alexina++and+Best%2C+Nicky+and+Richardson%2C+Sylvia+and+PLEWIS%2C+IAN++(2010)+Strategy+for+modelling+non-random+missing+data+mechanisms+in+observational+studies+using+Bayesian+methods.++Technical+Report.+Imperial+College+London.++++(Submitted)++