I'm doing a binary logistic regression with multiple imputation data. SPSS gives the regression output for each iteration, but no pooled results, which is what I'm looking for. It says "For at least one model, pooled estimates could not be computed because model parameters vary by imputation." Does anyone know how to address this? Alternatively, can the results of each iteration be averaged together to get the pooled result? Thank you!

Did you use a variable selection method that resulted in different models in different samples? You have to have the same model in each sample.