Abstract
This article is devoted to the theoretical substantiation and experimental validation of a methodological model for developing lexical skills through the use of linguistic corpora in the process of teaching Kazakh as a second language. The aim of the study is to analyze the theoretical foundations of corpus-based approaches to lexical skill development in Kazakh language instruction and to examine the effectiveness of an empirically tested instructional model. The research is grounded in the principles of applied linguistics, corpus linguistics, and Data-Driven Learning (DDL), on the basis of which a structured framework of lexical skills was established. The instructional materials were developed using data from the Kazakh National Corpus and were selected through frequency analysis, collocational analysis (MI-score ≥ 3.0), concordance analysis, and the KWIC method. The experimental teaching was conducted over a six-week period with both experimental and control groups. The research data were processed using SPSS 26.0, employing paired samples t-tests, independent samples t-tests, ANOVA, and Spearman correlation analysis.The results of the experiment demonstrate that the corpus-based methodology has a statistically significant impact on learners’ lexical skill development. In addition, the lexical richness of learners’ written texts was measured using TTR, MTLD, and collocational density indices. The findings provide empirical evidence supporting the effectiveness of systematic corpus use in teaching Kazakh and offer a solid scientific basis for updating curricula, instructional materials, and digital educational resources. The study confirms the research hypothesis that corpus-based instructional methods significantly enhance learners’ lexical development and contribute to the improvement of educational programs for teaching Kazakh as a second language.