Testing models of the irrelevant speech effect on working memory with event-related potentials

The impairment of short-term recall when irrelevant auditory material is presented together with the items to be memorized - the irrelevant speech (IS) effect - has been explained by two competing theories. Baddeley assumes that the IS effect is based on the phonological similarity between the material to be memorized and the irrelevant material. In contrast, Jones conceives the IS effect to be due to the sound variability within the irrelevant stream. In the present study we tested these theories by means of recording event-related potentials (ERPs) to visually presented letters alternating with more or less disturbing irrelevant auditory material selected as far as possible according to the critical conditions formulated by each model. Although Baddeley's IS condition was more detrimental to performance than the one according to Jones. a significant IS effect in ERPs was found only for the latter. Therefore it appears that the IS effect may have two different sources, possibly in line with each of both theories.