Fraction of Missing Information (γ) at Different Missing Data Fractions in the 2012 NAMCS Physician Workflow Mail Survey*

Details

• Alternative Title:
  Appl Math (Irvine)
• Personal Author:
  Pan, Qiyuan ; Wei, Rong
• Description:
  In his 1987 classic book on multiple imputation (MI), Rubin used the fraction of missing information, |, to define the relative efficiency (RE) of MI as RE = (1 + |/|)|, where | is the number of imputations, leading to the conclusion that a small | (≤5) would be sufficient for MI. However, evidence has been accumulating that many more imputations are needed. Why would the apparently sufficient | deduced from the RE be actually too small? The answer may lie with |. In this research, | was determined at the fractions of missing data (|) of 4%, 10%, 20%, and 29% using the 2012 Physician Workflow Mail Survey of the National Ambulatory Medical Care Survey (NAMCS). The | values were strikingly small, ranging in the order of 10| to 0.01. As | increased, | usually increased but sometimes decreased. How the data were analysed had the dominating effects on |, overshadowing the effect of |. The results suggest that it is impossible to predict | using | and that it may not be appropriate to use the |-based RE to determine sufficient |.
• Subjects:
  Article Fraction Of Missing Information (γ) Missing Data Multiple Imputation NAMCS Sufficient Number Of Imputations
• Source:
  Appl Math (Irvine). 7(10):1057-1067
• Pubmed ID:
  27398259
• Pubmed Central ID:
  PMC4934387
• Document Type:
  Journal Article
• Funding:
  CC999999/Intramural CDC HHS/United States
• Volume:
  7
• Issue:
  10
• Collection(s):
  CDC Public Access
• Main Document Checksum:
  urn:sha256:a400981442b428327b46675c350f7c530d0ec91dd033964be28de32d8b5b7f28
• Download URL:
  https://stacks.cdc.gov/view/cdc/40334/cdc_40334_DS1.pdf
• File Type:
  [PDF - 930.41 KB ]