A study based on a survey of familiarity ratings for multiword sequences among Japanese EFL learners

Paper presented at the 27th Annual Conference of the European Second Language Association (EuroSLA 27), University of Reading, England.

Isobe, Y., Kadota, S., Kanazawa, Y., Matsuda, N., & Morishita, M.

To explore how familiar EFL learners feel with highly frequent English multiword expressions, our research project conducted a major survey of familiarity ratings for whole word sequences on a seven-point scale using an online research tool, Survey Monkey. A tolal of 1,050 Japanese undergraduate students participated in this research. For the stimuli, 501 items were selected from the PRASE List (Martinez & Schmitt, 2012). The analysis of the Spearman rank correlation demonstrates that there is a mild correlation between the familiarity ratings by non-native speakers of English and BNC frequency for the multiword units (ρ= 0.47, p < .001, 95%CI [.40, .54]). No significant difference was observed between the familiarity ratings and the length of sequences, i.e., the number of words and syllables included, respectively (ρ= -.127, p = .004, 95% CI [-.212, -.04], ρ= -.163, p < .001, 95% CI [-.247, -.076]), which shows different trends from the word familiarity rating mentioned above. The overall results indicate that the familiarity rating for multiword sequences may be influenced by factors other than BNC frequency and the length of word strings.