Hi, everyone.

When I perform a t-test to compare two groups of a certain variable in a data set with multiple imputation of missing values using the statistical software SPSS, the analysis results of the pooled data show the mean and standard error, but not the standard deviation. Is it possible to calculate the standard deviation of the pooled data in the data set after multiple assignments? (However, I know that the SPSS manual states that the mean and standard error are calculated in the pooled data results when the data set is used after multiple assignments.)

How to deal with this problem?

I apreciate it and thanks in advance.

Best regards.

Hi Raim. The results shown as "Pooled" are not for pooled data, but for pooled results from analyziing separate imputed datasets. Multiple imputation does not result in a pooled dataset. The values shown for pooled results are estimates of population parameters based on pooling of results from the imputed datasets, not descriptive statistics. So there is no standard deviation available to compute, only a standard error based on the results for each of the imputed datasets.

Thank you so much for this explanation. But how do you get the most correct mean and SD for your dataset?