대학수학능력시험 영어영역에 대한 사회적 담론 분석: 자연어 처리 기반 텍스트 마이닝과 소셜 네트워크 분석을 활용한 2021-2024년 온라인 미디어 분석

A social discourse analysis on the English section of the Korean college scholastic ability test: Analysis of online media from 2021 to 2024 using natural language processing-based text mining and social network analysis

This study analyzes social discourse on the English section of the Korean College Scholastic Ability Test (K-CSAT) in online media from 2021 to 2024, using Natural Language Processing (NLP)-based text mining and Social Network Analysis (SNA). Scrutiny of 88,561 online documents employed Term Frequency (TF), Term Frequency-Inverse Document Frequency (TF-IDF), Bigram, CONvergence of iteration CORrelation (CONCOR), and sentiment analysis. The findings reveal three distinctive patterns. First, despite the introduction of absolute evaluation in 2018, grade-centered hierarchical culture remains dominant, leading to increased dependence on private education, as evidenced by ‘cram school (or private institute)’ showing the highest importance in TF-IDF analysis and eight differentiated learning strategy clusters emerging in CONCOR analysis. Second, early preparation trends are prominent, with the high importance of the ‘middle school’ keyword indicating that K-CSAT preparation begins at the middle school level. Third, sentiment analysis reveals that while positive sentiment (67.11%) predominates, negative sentiment (32.89%) is also significant, with learning burden and psychological pressure as major concerns. This study provides empirical data for policy improvement through comprehensive analysis of social perceptions regarding the K-CSAT English section.