Imputation procedures

Imputation of MOS Score

Each of the original numeric values from the MOS Sleep Problem Index will be converted to a 0-100 ange score with questions 2-5 being reverse scored. Then the converted values will be averaged. The resulting average will be the MSS score if at least three of the six questions are answered (nonmissing); otherwise, the MSS score will be missing. Hence this procedure will result in effectively imputing values if 3, 4 or 5 questions are answered. If at least 20% of the subjects have at least one imputed MSS score then the statistical analyses will be repeated based on non-imputed scores. (Matt's e-mail from Tue, 2 Aug 2005 13:10:55 -0400)

Imputation of CPSI Score

Imputation will be done for the pain subscale as follows. If no more than two binary items are missing from questions 1a-1d, 2a, 2b, and no other pain items are missing then the pain subscale will be imputed by: (sum of all non-missing items*(21)/(21-# missing items) If only question 3 is missing then the pain subscale will be imputed by: (sum of all non-missing items)*(21/16) No imputation will be done for the urinary and quality of life subscales. If any question is not answered on either the urinary or the quality of life subscales then the corresponding subscale score and the total CPSI score will be missing. If more than 20% of the subjects have missing data on the CPSI then the sensitivity of the analysis results to missing data will be examined using alternative imputation schemes. (Matt's e-mail from Mon, 16 May 2005 15:36:25 -0400)
Topic revision: r1 - 02 Aug 2005, SvetlanaEden

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback