Likelihood Approach for Bayesian Logistic Weighted Model

Missing Completely at Random Case

Keywords: Likelihood, Logistic weighting, Missing data, Preterm infats


Increasing the response rate and minimizing non-response rates represent the primary challenges to researchers in performing longitudinal and cohort research. This is most obvious in the area of paediatric medicine. When there are missing data, complete case analysis makes findings biased. Inverse Probability Weighting (IPW) is one of many available approaches for reducing the bias using a complete case analysis. Here, a complete case is weighted by probability inverse of complete cases. The data of this work is collected from the neonatal intensive care unit at Erbil maternity hospital for the years 2012 to 2017. In total, 570 babies (288 male and 282 females) were born very preterm. The aim of this paper is to use inverse probability weighting on the Bayesian logistic model developmental outcome. The Mental Development Index (MDI) approach is used for assessing the cognitive development of those born very preterm. Almost half of the information for the babies was missing, meaning that we do not know whether they have cognitive development issues or they have not. We obtained greater precision in results and standard deviation of parameter estimates which are less in the posterior weighted model in comparison with frequent analysis.


Download data is not yet available.


D. Wolke, B. Sohne, B. Ohrt and K. Riegel. Follow-up of preterm children: Important to document dropouts. Lancet, vol. 345, no. 8947, p. 447, 1995.

S. Johnson, S. E. Seaton, B. N. Manktelow, L. K. Smith, D. Field, E. S. Draper, N. Marlow and E. M. Boyle. Telephone interviews and online questionnaires can be used to improve neurodevelopmental follow-up rates. BMC Research Notes, vol. 7, p. 219, 2014.

D. B. Rubin. Inference and missing data. Biometrika,vol. 63, no. 3, pp. 581-592, 1976.

M. Hofler, H. Pfister, R. Lieb and H. U. Wittchen. The use of weights to account for non-response and drop-out. Social Psychiatry and Psychiatric Epidemiology, vol. 40, no. 4, pp. 291-299, 2005.

L. Lazzeroni, N. Schenker and J. Taylor. Robustness of Multiple-imputation Techniques to Model Misspecification. United States: Proceedings of the Survey Research Methods Section, American Statistical Association, pp. 260-265, 1990.

B. L. Carlson and S. Williams. A Comparison of Two Methods to Adjust Weights for Non-response: Propensity Modeling and Weighting Class Adjustments. United States: Proceedings of the Annual Meeting of the American Statistical Association, 2001.

C. Agostinelli and L. Greco. Weighted Likelihood in Bayesian Inference. New York: Proceedings of the 46th Scientific Meeting of the Italian Statistical Society, 2012.

S. R. Seaman and I. R. White. Review of inverse probability weighting for dealing with missing data. Statistical Methods in Medical Research, vol. 22, no. 3, pp. 278-295, 2013.

S. R. Seaman, I. R. White, A. J. Copas and L. Li. Combining multiple imputation and inverse‐probability weighting. Biometrics, vol. 68, no. 1, pp. 129-137, 2012.

D. Clayton, D. Spiegelhalter, G. Dunn and A. Pickles. Analysis of longitudinal binary data from multiphase sampling. Journal of the Royal Statistical Society, vol. 60, no. 1, pp. 71-87, 1998.

R. J. Little and D. B. Rubin. Statistical Analysis with Missing Data. Chichester: Wiley, p. 5, 1987.

L. H. Curtis, B. G. Hammill, E. L. Eisenstein, J. M. Kramer and K. J. Anstrom. Using inverse probability-weighted estimators in comparative effectiveness analyses with observational databases. Medical Care, vol. 45, no. 10, pp. S103-S107, 2007.

How to Cite
Kadir D. Likelihood Approach for Bayesian Logistic Weighted Model. cuesj [Internet]. 13Aug.2020 [cited 24Sep.2020];4(2):9-2. Available from:
Research Article