In defence of keyword experiments: A reply to Gruneberg's commentary

Gruneberg (1998) argues that our findings (e.g. Wang and Thomas, 1992; Thom as and Wang, 1996) that the keyword method of foreign language learning lea ds to poorer long-term retention than rote rehearsal when the delay interva l is manipulated between subjects is uninteresting because, in a natural se tting, students would normally be tested soon after study. In reply, we ass ert that our data are internally valid and inform us with respect to the na ture of the underlying causal relationship. Specifically, our experimental findings lead us to conclude that the imagery-based encodings of the keywor d method are very fragile and unlikely to be retrievable after a delay unle ss frequently rehearsed. Also, another consequence of the difficulty of ret rieving keyword images is that translation time is slower for keyword learn ers than for learners who rote rehearse. We argue that design and implement ation of language instruction programmes is most likely to be successful wh en educators understand the underlying causal factors. Consequently, inform ed debate on the effectiveness of a learning strategy can only proceed with evidence gathered from both, experimental and real-world settings. Copyrig ht (C) 1999 John Wiley & Sons, Ltd.