1932

Abstract

I review assumptions about the missing-data mechanisms that underlie methods for the statistical analysis of data with missing values. I describe Rubin's original definition of missing at random (MAR), its motivation and criticisms, and his sufficient conditions for ignoring the missingness mechanism for likelihood-based, Bayesian, and frequentist inference. Related definitions, including missing completely at random, always MAR, always missing completely at random, and partially MAR, are also covered. I present a formal argument for weakening Rubin's sufficient conditions for frequentist maximum likelihood inference with precision based on the observed information. Some simple examples of MAR are described, together with an example where the missingness mechanism can be ignored even though MAR does not hold. Alternative approaches to statistical inference based on the likelihood function are reviewed, along with non-likelihood frequentist approaches, including weighted generalized estimating equations. Connections with the causal inference literature are also discussed. Finally, alternatives to Rubin's MAR definition are discussed, including informative missingness, informative censoring, and coarsening at random. The intent is to provide a relatively nontechnical discussion, although some of the underlying issues are challenging and touch on fundamental questions of statistical inference.

Loading

Article metrics loading...

/content/journals/10.1146/annurev-statistics-040720-031104
2021-03-07
2024-06-22
Loading full text...

Full text loading...

/deliver/fulltext/statistics/8/1/annurev-statistics-040720-031104.html?itemId=/content/journals/10.1146/annurev-statistics-040720-031104&mimeType=html&fmt=ahah

Literature Cited

  1. Afifi AA, Elashoff RM. 1966. Missing observations in multivariate statistics 1: review of the literature. J. Am. Stat. Assoc. 61:595–604
    [Google Scholar]
  2. Allen AS, Rathouz PJ, Satten GA 2003. Informative missingness in genetic association studies: case-parent designs. Am. J. Hum. Genet. 72:671–80
    [Google Scholar]
  3. Anderson TW. 1957. Maximum likelihood estimates for the multivariate normal distribution when some observations are missing. J. Am. Stat. Assoc. 52:200–3
    [Google Scholar]
  4. Barnard J, Frangakis CE, Hill J, Rubin DB 2002. School choices in NY city: a Bayesian analysis of an imperfect randomized experiment (with discussion). In Case Studies in Bayesian Statistics, Vol. 5 R Kass, B Carlin, A Carriquiry, A Gelman, I Verdinelli, M West 33–97 New York: Springer
    [Google Scholar]
  5. Beesley LJ, Taylor JMG, Little RJ 2019. Sequential imputation for models with latent variables assuming latent ignorability. Aust. N. Z. J. Stat. 61:2213–33
    [Google Scholar]
  6. Cox DR. 1975. Partial likelihood. Biometrika 62:2269–76
    [Google Scholar]
  7. Diggle P, Kenward MG. 1994. Informative drop-out in longitudinal data analysis. J. R. Stat. Soc. C 43:49–73
    [Google Scholar]
  8. De Groot MH, Goel PK 1980. Estimation of the correlation coefficient from a broken random sample. Ann. Stat. 8:2264–78
    [Google Scholar]
  9. Follman D, Wu M. 1995. An approximate generalized linear model with random effects for informative missing data. Biometrics 51:1151–68
    [Google Scholar]
  10. Harel O, Schafer JL. 2009. Partial and latent ignorability in missing data problems. Biometrika 96:137–50
    [Google Scholar]
  11. Hartley HO. 1958. Maximum likelihood estimation from incomplete data. Biometrics 14:174–94
    [Google Scholar]
  12. Heitjan DF. 1994. Ignorability in general incomplete-data models. Biometrika 81:4701–8
    [Google Scholar]
  13. Heitjan DF. 2004. Estimation with missing data (correspondence). Biometrics 50:580
    [Google Scholar]
  14. Heitjan DF, Rubin DB. 1990. Inference from coarse data via multiple imputation with application to age heaping. J. Am. Stat. Assoc. 85:410304–14
    [Google Scholar]
  15. Higgins JPT, White IR, Wood AM 2008. Imputation methods for missing outcome data in meta-analysis of clinical trials. Clin. Trials 5:225–39
    [Google Scholar]
  16. Imbens GW, Rubin DB. 2015. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction Cambridge, UK: Cambridge Univ. Press
    [Google Scholar]
  17. Jacobsen M, Keiding N. 1995. Coarsening at random in general sample spaces and random censoring in continuous time. Ann. Stat. 23:3774–86
    [Google Scholar]
  18. Kenward MG, Molenberghs G. 1998. Likelihood based frequentist inference when data are missing at random. Stat. Sci. 3:3236–47
    [Google Scholar]
  19. Lipsitz SR, Ibrahim JG, Zhao LP 1999. A weighted estimating equation for missing covariate data with properties similar to maximum likelihood. J. Am. Stat. Assoc. 94:1147–60
    [Google Scholar]
  20. Little RJ. 1976. Discussion of “Inference and Missing Data” by D.B. Rubin. Biometrika 63:590–92
    [Google Scholar]
  21. Little RJ. 1992. Regression with missing X's: a review. J. Am. Stat. Assoc. 87:1227–37
    [Google Scholar]
  22. Little RJ. 1995. Modeling the drop-out mechanism in longitudinal studies. J. Am. Stat. Assoc. 90:1112–21
    [Google Scholar]
  23. Little RJ. 2020. On algorithmic and modeling approaches to imputation in large data sets. Stat. Sin. 30:1685–96
    [Google Scholar]
  24. Little RJ, David M. 1988. Weighting adjustments for non-response in panel surveys Tech. Rep., US Dep. Commerce, Bur. Census Washington, DC:
    [Google Scholar]
  25. Little RJ, Rubin DB. 2019. Statistical Analysis with Missing Data New York: Wiley. , 3rd ed..
    [Google Scholar]
  26. Little RJ, Rubin DB, Zanganeh SZ 2016. Conditions for ignoring the missing-data mechanism in likelihood inferences for parameter subsets. J. Am. Stat. Assoc. 112:314–20
    [Google Scholar]
  27. Little RJ, Zhang N. 2011. Subsample ignorable likelihood for regression analysis with missing data. J. R. Stat. Soc. C 60:4591–605
    [Google Scholar]
  28. Marini MM, Olsen AR, Rubin DB 1980. Maximum-likelihood estimation in panel studies with missing data. Sociol. Methodol. 11:314–57
    [Google Scholar]
  29. Mealli F, Rubin DB. 2015. Clarifying missing at random and related definitions and implications when coupled with exchangeability. Biometrika 102:4995–1000
    [Google Scholar]
  30. Mealli F, Rubin DB. 2016. Amendments and corrections. Biometrika 103:2491
    [Google Scholar]
  31. journal 2010. The Prevention and Treatment of Missing Data in Clinical Trials Washington, DC: Natl. Acad. Press
    [Google Scholar]
  32. Park S, Palta M, Shao J, Shen L 2002. Bias adjustment in analysing longitudinal data with informative missingness. Stat. Med. 21:277–91
    [Google Scholar]
  33. Pauli F, Racugno W, Ventura L 2011. Bayesian composite marginal likelihoods. Stat. Sin. 21:149–64
    [Google Scholar]
  34. Pearl J, Glymour M, Jewell NP 2016. Causal Inference in Statistics: A Primer New York: Wiley
    [Google Scholar]
  35. Peng Y, Little RJ, Raghunathan TE 2004. An extended general location model for causal inferences from data subject to noncompliance and missing values. Biometrics 60:598–607
    [Google Scholar]
  36. Robins JM, Gill RD. 1997. Non-response models for the analysis of non-monotone ignorable missing data. Stat. Med. 16:39–56
    [Google Scholar]
  37. Robins JM, Rotnitzky A. 1995. Semiparametric efficiency in multivariate regression models with missing data. J. Am. Stat. Assoc. 90:122–29
    [Google Scholar]
  38. Robins JM, Rotnitzky A, Zhao LP 1995. Analysis of semiparametric regression models for repeated outcomes in the presence of missing data. J. Am. Stat. Assoc. 90:106–21
    [Google Scholar]
  39. Rubin DB. 1976. Inference and missing data. Biometrika 63:581–92
    [Google Scholar]
  40. Rubin DB. 1978. Bayesian inference for causal effects: the role of randomization. Ann. Stat. 6:134–58
    [Google Scholar]
  41. Rubin DB. 1987. Multiple Imputation for Nonresponse in Surveys New York: Wiley
    [Google Scholar]
  42. Scharfstein D, Rotnitsky A, Robins J 1999. Adjusting for nonignorable dropout using semiparametric models (with discussion). J. Am. Stat. Assoc. 94:1096–146
    [Google Scholar]
  43. Seaman S, Galati J, Jackson D, Carlin J 2013. What is meant by “missing at random?. Stat. Sci. 28:2257–68
    [Google Scholar]
  44. Sinha D, Ibrahim JG. 2003. A Bayesian justification of Cox's partial likelihood. Biometrika 90:3629–41
    [Google Scholar]
  45. journal 2017. SAS® Visual Data Mining and Machine Learning 8.1 Statistical Procedures Cary, NC: SAS Inst. Inc.
    [Google Scholar]
  46. Trawinski IM, Bargmann RW. 1964. Maximum likelihood with incomplete multivariate data. Ann. Math. Stat. 35:647–57
    [Google Scholar]
  47. Ventura L, Cabras S, Racugno W 2009. Prior distributions from pseudo-likelihoods in the presence of nuisance parameters. J. Am. Stat. Assoc. 104:768–74
    [Google Scholar]
  48. Wu MC, Carroll RJ. 1988. Estimation and comparison of changes in the presence of informative right censoring by modeling the censoring process. Biometrics 44:175–88
    [Google Scholar]
  49. Wu MC, Follman DA. 1998. Use of summary measures to adjust for informative missingness in repeated measures data with random effects. Biometrics 55:175–84
    [Google Scholar]
/content/journals/10.1146/annurev-statistics-040720-031104
Loading
/content/journals/10.1146/annurev-statistics-040720-031104
Loading

Data & Media loading...

  • Article Type: Review Article
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error