When listening in noisy conditions, word recognition seems to be much harder in a non-native language than in one's native language. Native listeners use both word-initial and word-final information for word recognition in clean listening conditions, where word-initial information is the most important. When listening in noise, however, word-final information becomes relatively more important. This study investigates whether non-native listeners are able to use word-initial and word-final information when recognizing words in noise, and whether these information sources are equally important when listening conditions become increasingly harder. Forty-seven Dutch students participated in an English word recognition experiment, where either a word's onset or offset was masked by speech-shaped noise with different signal-to-noise ratios. The results showed that non-native listeners are able to use both word-initial and word-final information for word recognition, but fewer words were recognized with increasing difficulty of the listening conditions when the onset of words was masked. Thus, word-initial information is more important than word-final information for word recognition when listening conditions become harder. This increasing effect occurred independently from the proficiency level in the non-native language of the participants, although proficiency level was correlated to test performance in general.