We studied the effects of an electromagnetic field (EMF) as emitted by a 902 MHz mobile phone on human short term memory. This study was a replication with methodological improvements to our previous study. The improvements included multi-centre testing and a double blind design. A total of 64 subjects (32 men) in two independent laboratories performed a short term memory task (n-back) which poses a varying memory load (0-3 items) on the subjects' memory. They performed the task twice, once each under EMF and sham exposure. Reaction times (RTs) and accuracy of the responses were recorded. The order of exposure and memory load conditions were counterbalanced across subjects and gender. There were no statistically significant differences in performance between the two laboratories. We could not replicate our previous results: the EMF had no effect on RTs or on the accuracy of the subjects' answers. The inability to replicate previous findings could have been caused by lack of actual EMF effects or the magnitude of effects being at the sensitivity threshold of the test used.