Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • cutt-off missing data

    Dear statlisters,
    i have a data set with 1500 observation and 60 variables. the highest percentage of missing data is 7.8% (see attached). do i need to impute the missing data in this case. some literature suggests it is not necessary bu i am not sure.

    stata 15.1 on mac




    Attached Files
    Last edited by abdelilah arredouani; 14 Feb 2019, 02:27.

  • #2
    Abdelilah:
    please see the FAQ about attachments.
    That said, (multiple) imputation of missing data depends on the underlying missing mechanism (that you do not mention in your post).
    Unfortunately, there's no a percentage cut-off (if this is what you mean in your post) that you can invoke to justify your leaving missing data untreated.
    In general, discarding observations with missing values is not a good
    Last edited by Carlo Lazzaro; 14 Feb 2019, 02:57.
    Kind regards,
    Carlo
    (Stata 15.1 SE)

    Comment


    • #3
      Just as a side note, you can do a sensitivity analysis and compare the complete-case estimations with the imputed ones.
      Best regards,

      Marcos

      Comment

      Working...
      X