Data missing not at random (MNAR) is a major challenge in survey sampling. We proposean approach based on registry data to deal with non-ignorable missingness in healthexamination surveys. The approach relies on follow-up data available from administrativeregisters several years after the survey. For illustration we use data on smoking prevalencein Finnish National FINRISK study conducted in 1972-1997. The data consist ofmeasured survey information including missingness indicators, register-based backgroundinformation and register-based time-to-disease survival data. The parameters of missingnessmechanism are estimable with these data although the original survey data areMNAR. The underlying data generation process is modelled by a Bayesian model. The resultsindicate that the estimated smoking prevalence rates in Finland may be significantlyaffected by missing data.
展开▼
