Latent Class Analysis: Pemetaan Eksploratif Kompetensi Literasi Siswa SMA

Abstract

High school students' literacy skills in Indonesian still show various problems, especially in understanding implicit meanings, logical relationships between ideas, evaluating information, and interpreting context. Learning evaluations generally focus only on total scores, thus failing to map the characteristics of students' literacy skills in depth. Therefore, this study aims to identify students' literacy skill profiles through a Latent Class Analysis (LCA) approach based on the results of the Indonesian Language Academic Ability Test (TKA). This study was chosen because literacy skills are multidimensional and require an analytical approach that can reveal the heterogeneity of student abilities more specifically. The study used a quantitative descriptive method with an exploratory approach. The research subjects consisted of 21 students of SMAN 1 Banyumas, Pringsewu Regency, in the even semester of the 2025/2026 academic year. The research instrument consisted of 20 TKA questions covering aspects of pragmatics, discourse, stylistics, semantics, and syntax. The instrument has passed validity and reliability tests with a KR-20 value of 0.95. Data analysis was conducted descriptively and continued using LCA assisted by R software. The results showed that the average percentage of students' correct answers only reached 36.67%, with the main weaknesses in the syntax, semantics, and discourse aspects. The LCA results showed that the two-class model was the best model with class proportions of 61.29% and 38.71%. These findings indicate that students' literacy competencies are heterogeneous and divided into different ability profiles. This study contributes to the development of ability profile-based literacy evaluation and provides a basis for the development of adaptive Indonesian language learning at the high school level.