Phoneme discrimination tests based on presenting minimal pairs (MP) are frequently used to assess auditory processing. However, test results may be influenced by other factors beside correct acoustic representation. Here we studied if working memory capacity (WMC) correlates with reproducing sets of minimal pairs.Participants were 161 preschool children whereby WMC was evaluated by memory span for digits and MPs were either real words or non-words (logatoms). Both real word MPs and logatom MPs were presented in doublets or triplets. Children were asked to repeat what they had heard. Correlation analysis was performed using Spearman rank correlation.All four test item blocks showed a significant positive correlation to WMC (group values). Data from individuals indicated that some children with poor WMC performed sufficiently on MP discrimination and vice versa.Besides confirming the phonological similarity effect our data indicate a significant interdependence between different processes and phonological representations (according to the Logogen model). This must be taken into account when using MP tests for evaluating "auditory processing" in the clinical routine. However, poor WMC is neither necessary nor obligatory for poor phoneme discrimination.