This is a preliminary report on a perception experiment in which English stress shift was the object of study. Spanish speaking learners of English (FELs) and native English listeners (NELs) took part in the test. They were asked to identify auditorily the position and relative prominence of stresses in (i) English double-stressed single words and (ii) English compound words with an end-stress pattern, both when these structures exhibited their default end-stress pattern and when stresses appeared shifted to the front. Results indicate that both listener groups recognise the differing stress patterns in the two conditions -unshifted and shifted- very accurately. Looking at the relative strength and number of stresses perceived, it was seen that NELs showed a stronger tendency towards prominence shift perception in single words than FELs, who were more likely to hear single words as displaying one stress only. As far as compound words were concerned, differences between the listener groups were not significant. It is concluded that, in the experimental tasks posed, native competence did not prove to be an indicator of possible advantage for stress identification. It is suggested that such a result may be partly due to differences in metalinguistic knowledge between English and Spanish listeners.