Clarifying missing at random and related definitions, and implications when coupled with exchangeability

December 2015
Biometrika;Dec2015, Vol. 102 Issue 4, p995
Academic Journal
We clarify the key concept of missingness at random in incomplete data analysis. We first distinguish between data being missing at random and the missingness mechanism being a missing-at-random one, which we call missing always at random and which is more restrictive.We further discuss how, in general, neither of these conditions is a statement about conditional independence.We then consider the implication of the more restrictive missing-always-at-random assumption when coupled with full unit-exchangeability for the matrix of the variables of interest and the missingness indicators: the conditional distribution of the missingness indicators for any variable that can have a missing value can depend only on variables that are always fully observed. We discuss implications of this for modelling missingness mechanisms.


Related Articles

  • Sparsity and the Bayesian perspective. Starck, J.-L.; Donoho, D. L.; Fadili, M. J.; Rassat, A. // Astronomy & Astrophysics / Astronomie et Astrophysique;Apr2013, Vol. 552 Issue 1, p1 

    Sparsity has recently been introduced in cosmology for weak-lensing and cosmic microwave background (CMB) data analysis for different applications such as denoising, component separation, or inpainting (i.e., filling the missing data or the mask). Although it gives very nice numerical results,...

  • Bayesian imputation of non-chosen attribute values in revealed preference surveys. Washington, Simon; Ravulaparthy, Srinath; Rose, John M.; Hensher, David; Pendyala, Ram // Journal of Advanced Transportation;Jan2014, Vol. 48 Issue 1, p48 

    SUMMARY Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer...

  • Bayesian perspectives for epidemiologic research. III. Bias analysis via missing-data methods. Greenland, Sander // International Journal of Epidemiology;Aug2010, Vol. 39 Issue 4, p1116 

    A correction to the article "Bayesian perspectives for epidemiologic research. III. Bias analysis via missing-data methods" that was published in the 2009 issue is presented.

  • Practical Approaches to Principal Component Analysis in the Presence of Missing Values. Ilin, Alexander; Raiko, Tapani // Journal of Machine Learning Research;7/1/2010, Vol. 11 Issue 7, p1957 

    No abstract available.

  • Reality Check: Combining Choice Experiments with Market Data to Estimate the Importance of Product Attributes. Feit, Eleanor McDonnell; Beltramo, Mark A.; Feinberg, Fred M. // Management Science;May2010, Vol. 56 Issue 5, p785 

    Discrete choice models estimated using hypothetical choices made in a survey setting (i.e., choice experiments) are widely used to estimate the importance of product attributes in order to make product design and marketing mix decisions. Choice experiments allow the researcher to estimate...

  • Testing the null of a low dimensional growth model. Jensen, Peter Sandholt // Empirical Economics;Feb2010, Vol. 38 Issue 1, p193 

    Applied growth researchers face a sample with more regressors than observations when using cross-country data. In this paper, we examine simple procedures for testing if a true submodel has been found. Specifically, we examine three tests that can be used to test a low dimensional growth model...

  • Managing nonignorable missing data with clustered multinomial responses. Wang, Mu-Cyun; Lin, I-Feng // Statistical Modelling: An International Journal;Dec2015, Vol. 15 Issue 6, p548 

    Clustered multinomial responses are common in public health studies. In this situation, the baseline logit random effects model is usually suggested as a general modelling approach. When nonignorable missing outcomes exist, naïve methods such as complete case analysis or likelihood methods...

  • Introduction to the special issue on joint modelling techniques. Rizopoulos, Dimitris; Lesaffre, Emmanuel // Statistical Methods in Medical Research;Feb2014, Vol. 23 Issue 1, p3 

    Joint modelling techniques have seen great advances in the recent years, with several types of joint models having been developed in literature that can handle a wide range of applications. This special issue of Statistical Methods in Medical Research presents some recent developments from this...

  • Explicating the Conditions Under Which Multilevel Multiple Imputation Mitigates Bias Resulting from Random Coefficient-Dependent Missing Longitudinal Data. Gottfredson, Nisha; Sterba, Sonya; Jackson, Kristina; Gottfredson, Nisha C; Sterba, Sonya K; Jackson, Kristina M // Prevention Science;Jan2017, Vol. 18 Issue 1, p12 

    Random coefficient-dependent (RCD) missingness is a non-ignorable mechanism through which missing data can arise in longitudinal designs. RCD, for which we cannot test, is a problematic form of missingness that occurs if subject-specific random effects correlate with propensity for missingness...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics