[utstat.general] Statistics Seminar: Effect on Data Analysis of the Use of Imputed Values

ruth@utstat.uucp (Ruth Croxford) (02/06/90)

Topic:    Effect on Secondary Data Analysis of the Use of Imputed Values
Speaker:  J. Sedransk
	  National Center for Health Statistics and University of Iowa
Date:     Thursday, Feb. 15, 1990     4:00 - 5:00 p.m.
Place:    Room 1085, Sidney Smith Hall, 100 St George Street, U of T.

Abstract:
	Considering several imputation methods, we investigate their properties
from the perspective of the secondary data analyst who assumes that the data set
to be analyzed has only observed responses.  We first review our findings for the
case when the missing data can be regarded as missing at random.  Then we 
describe a general specification that defines respondent and nonrespondent 
subpopulations.  This specification is the basis for determiing the properties
of statistics such as the sample mean and variance when the latter contain both
observed and imputed values.  We consider the use of standard imputation methods
together with alternatives that make use of complete or partial knowledge of the
missing data process.  The methods are compared, and implications for the practice
of imputation are discussed.
------------

Coffee and tea will be served in the De Lury Lounge (SS6006) at 3:30 p.m.