Résumé:
The present research aims at highlighting the efficiency of artificial intelligence (AI) and
Natural Language Processing tools (NLP) in text analysis and the assessment of students’
language proficiency. The study targets one dimension of language proficiency embodied in
the lexical richness of students’ writings. It represents a study focused on measuring the
density, sophistication, and diversity of students’ vocabulary. A descriptive-correlational
research design was adopted while employing a quantitative method for data analysis. The
participants involved 60 English as a Foreign Lnaguage (EFL) learners studying at the
University of Guelma - 8 Mai 1945. The selected samples embodied two groups: 30 first-year
and 30 third-year students from the English Departement. After assigning a writing task for
students, the output was transcribed to promote the process of automated data analysis. The
study employed the web-based Lexical Complexity Analyzer (LCA), introduced by Ai and Lu
in 2010 for the examination of the concerned corpus. The findings showed that extensive
exposure to English in academic settings does not necessarily increase the lexical density and
diversity of vocabulary in students’ writing. Yet, extensive exposure appears to be a
prominent factor in improving the lexical sophistication of students’ writings, which is
considered the most important proxy measure of lexical richness. These latter findings were
corroborated by the reported significant difference in the mean scores belonging to the two
evaluated groups, as third-year students scored higher averages (Av. LS1= 0.83; Av. LS2=
0.69) in comparison to first-year students (Av. LS1= 0.72; Av. LS2= 0.59). The discrepancy
in the mean values was statistically significant in relation to LS2 (p= .03) with a moderate to
large effect size (LS1, d= 0.5; LS2= d= 0.6) suggesting that such a variation has practical
meaningfulness, which in turn implies that the difference in the lexical sophistication of the
two groups is likely noticeable an ordinary reader. Further research in the academic context is
still needed to achieve a more comprehensive understanding of the trajectory of students’
vocabulary acquisition.
