I have conducted analyses to compare dependent correlations (Steiger, 1980) using multiply imputed data. The statistics follow a t distribution.

Question: What is the correct way to pool these statistics, with the goal of generating a p-value?

Here is what the results look like (from 20 imputations): -2.938462 -2.627696 -2.562013 -2.354825 -2.542520 -2.442978 -2.583879 -2.421124 -2.746131 -2.936789 -2.328653 -2.884968 -2.423605 -2.567752 -1.787200 -2.675324 -2.815900 -2.307922 -2.968579 -2.299145

  • 1
    $\begingroup$ What is your question and what is the design of the experiment? Also, can you post a sample of your data? $\endgroup$
    – André.B
    Commented Sep 12, 2019 at 22:29

1 Answer 1


Take a look at the following paper, it talks about conducting inference when you have multiply imputed data.

Schafer, J. L. (1999). Multiple imputation: a primer. Statistical Methods in Medical Research, 8(1):3-15.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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