Fraction of Missing Information γ at Different Missing Data Fractions in the 2012 NAMCS Physician Workflow Mail SurveyReportar como inadecuado




Fraction of Missing Information γ at Different Missing Data Fractions in the 2012 NAMCS Physician Workflow Mail Survey - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

In his 1987 classic book on multipleimputation MI, Rubin used the fraction of missing information, γ, to definethe relative efficiency RE of MI as RE = 1 + γ-m?1-2, where m is the numberof imputations, leading to the conclusion that a small m ≤5 would besufficient for MI. However, evidence has been accumulating that many moreimputations are needed. Why would the apparently sufficient m deduced from theRE be actually too small? The answer may lie with γ. In this research, γ wasdetermined at the fractions of missing data δ of 4%, 10%, 20%, and 29% usingthe 2012 Physician Workflow Mail Survey of the National Ambulatory Medical CareSurvey NAMCS. The γ values were strikingly small, ranging in the order of10?6 to 0.01. As δ increased, γ usually increased but sometimes decreased. Howthe data were analysed had the dominating effects on γ, overshadowing theeffect of δ. The results suggest that it is impossible to predict γ using δ andthat it may not be appropriate to use the γ-based RE to determine sufficient m.

KEYWORDS

Multiple Imputation, Fraction of Missing Information γ, Sufficient Number of Imputations, Missing Data, NAMCS

Cite this paper

Pan, Q. and Wei, R. 2016 Fraction of Missing Information γ at Different Missing Data Fractions in the 2012 NAMCS Physician Workflow Mail Survey. Applied Mathematics, 7, 1057-1067. doi: 10.4236-am.2016.710093.





Autor: Qiyuan Pan, Rong Wei

Fuente: http://www.scirp.org/



DESCARGAR PDF




Documentos relacionados