Measuring lexical development in/through writing

1 downloads 0 Views 674KB Size Report
Daller, Michael et al. 2013. “Vocabulary acquisition and the learning curve.” In: Jarvis, Scott and Michael Daller. Vocabulary Knowledge. Human ratings and ...
Measuring lexical development in/through writing Paul Pauwels

Past research • LFP can be used to differentiate between proficiency levels + stable for similar types of text (Laufer and Nation, 1995) • Mixed results in showing lexical development o Laufer (1998): 2 groups in same programme – year 1 and year 2 – no significant progress o Laufer and Paribakht (1998:384) – low, int, adv EFL and ESL - “…implying that a moderate increase in [passive vocabulary] within a proficiency level would not necessarily result in a more sophisticated free active vocabulary.” o Daller et al. (2013) – beginners – 10 compositions during 80wk programme – clear progress

2015 pilot study – set up o o o

o o

o

18 year 1 Ba students, Flanders, classroom based; 250 word pros and cons essays +/- 10 words 3 topics (changing the clock in summer and winter; jobdependent retirement age; entrance exams for HE) 3 times (before christmas, before easter, after easter) 3 LFP-versions using ‘Range’: GSL/AWL (Laufer and Nation 1995), COCA/BNC (Nation and Webb 2011), Common Core (Gardner 2013) Use of TYPE % (not wordfamily)

2015 pilot study – results 1 • Correlations between measures not stable across essays • ANOVA/Bonferroni o

o o o

o

significant progress from time 1 to time 2 and 3 (condensed profile (Laufer 1998), BNC/COCA 1000); from time 1 to time 2 only (AWL, BNC/COCA 2000); from time 1 to time 3 (BNC/COCA 3000+4000); no progress, contrary to previous investigations: function words; no measures show progress from time 2 to time 3;

2015 pilot study – results 2 • Different growth paths for students scoring above vs students below average on essay 1 o Below average writers make more progress o More advanced writers do not progress on function words, progress less spectacularly on condensed profile, but do use increasingly fewer BNC1000 words (spoken, informal) • impression: essay topic 2 may have elicited more advanced vocabulary (or topic 1 less advanced) • Only 1 significant correlation with external proficiency measures: Vocablab levels test (taken at time 1) – essay 1 Common Core A+B lists: -.531 (p 5%; NS speaker use of function words