×

Protective estimation of longitudinal categorical data with nonrandom dropout. (English) Zbl 0964.62514

Summary: Partially observed longitudinal categorical data, where the partial classification arises due to monotone dropout, are analyzed using a protective estimator, which was first suggested by C. H. Brown [Biometrics 46, 143-155 (1990; Zbl 0718.62215)] for normally distributed data. It is appropriate when dropout depends on the unobserved outcomes only, a particular type of nonignorable nonresponse. Estimation of measurement parameters is possible, without explicitly modelling the dropout process. Necessary and sufficient conditions are derived in order to have a unique solution in the interior of the parameter space. It is shown that precision estimates can be based on the delta method, the EM algorithm, and on multiple imputation. The relative merits of these techniques are discussed and they are contrasted with direct likelihood estimation and with pseudo-likelihood estimation. The method is illustrated using data taken from a psychiatric study.

MSC:

62P10 Applications of statistics to biology and medical sciences; meta analysis
62F10 Point estimation
62H17 Contingency tables

Citations:

Zbl 0718.62215
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Agresti A., Categorical Data Analysis (1990) · Zbl 0716.62001
[2] Arnold B.C. Strauss D. Pseudolikelihood estimation Department of Statistics, University of California Riverside, CA 1988
[3] Baker S.G., J. Amer. Statist. Assoc 83 pp 62– (1988) · doi:10.1080/01621459.1988.10478565
[4] Brown CH., Biometrics 46 pp 143– (1990) · Zbl 0718.62215 · doi:10.2307/2531637
[5] Conaway M.R., Statistics in Medicine 11 pp 799– (1992) · doi:10.1002/sim.4780110610
[6] Cressie N., Statistics for spatial data (1991) · Zbl 0799.62002
[7] Diggle P.J., Appl. Statist 43 pp 49– (1994) · Zbl 0825.62010 · doi:10.2307/2986113
[8] Fay R.E., J. Amer. Statist. Assoc 81 pp 354– (1986) · doi:10.1080/01621459.1986.10478279
[9] Fuchs C., J. Amer. Statist. Assoc 77 pp 270– (1982) · doi:10.1080/01621459.1982.10477795
[10] Geys H., Pseudo-likelihood inference for clustered binary data (1996) · Zbl 0954.62530
[11] Laird N.M., Statistics in Medicine 7 pp 305– (1988) · doi:10.1002/sim.4780070131
[12] Le Cessie S., Applied Statistics 43 pp 95– (1994) · Zbl 0825.62509 · doi:10.2307/2986114
[13] Lehnen R.G., Public. Opinion Quarterly 38 pp 40– (1974) · doi:10.1086/268133
[14] Little R.J.A., J. Amer. Statist. Assoc 88 pp 125– (1993)
[15] Little R.J.A., J. Amer. Statist. Assoc. 90 pp 1112– (1995) · doi:10.1080/01621459.1995.10476615
[16] Little R.J.A., Statistical Analysis with Missing Data (1987) · Zbl 0665.62004
[17] Meilijson I., J. ft. Statist. Soc. B 51 pp 127– (1989)
[18] Molenberghs G. Kenward M.G. Lesaffre E. The analysis of longitudinal ordinal data with informative dropout 1994 · Zbl 0825.62797
[19] Molenberghs G., J. Amer. Statist. Assoc 89 pp 633– (1994) · doi:10.1080/01621459.1994.10476788
[20] Park T., J. Amer. Statist. Assoc 89 pp 44– (1994) · doi:10.1080/01621459.1994.10476444
[21] Rubin D.B., Multiple Imputation for Nonresponse in Surveys (1987) · Zbl 1070.62007 · doi:10.1002/9780470316696
[22] Rubin D.B., J. Amer. Statist. Assoc 81 pp 366– (1986) · doi:10.1080/01621459.1986.10478280
[23] Stasny E.A., J. Amer. Statist. Assoc 81 pp 42– (1986) · doi:10.1080/01621459.1986.10478236
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.