2
$\begingroup$

We have missing data which we want to impute in order to provide an imputed value to some business users. However, we will not be providing any other information other than the point estimate.

In addition, we don't have the capacity to provide a range of possible values derived from multiple imputations. The only thing we can provide is one single value for each missing data point.

In this situation, is it still advised to use multiple imputation (and take the average over the imputations) over single imputation?

$\endgroup$
  • $\begingroup$ are giving your users a dataset or model results? $\endgroup$ – Maarten Buis Nov 5 '14 at 14:49
  • 2
    $\begingroup$ Literally just performing (multiple) imputation to ascertain estimates for missing values and passing those values through to the users. They will, in many respects, be none the wiser as to how they are generated. $\endgroup$ – NickB2014 Nov 5 '14 at 15:25
1
$\begingroup$

As explained in this answer, multiple imputation is used to generate multiple datasets, perform statistical analysis on them, and average the results. Basically, multiple imputation takes a simple imputation and adds to it a random value to try to restore randomness lost in the imputation process. Therefore, averaging multiple imputations before doing any statistical analysis on them just removes most of that restored randomness (by averaging) and gives a result close to simple imputation plus an small random error.

Therefore, there is no advantage in using multiple imputations to report average of imputations over just using and reporting simple imputation.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.