Determining the Constructs Measured by the High School Entrance Exam (LGS): A Study of Construct Validity

In many countries, various examinations are administered to assess the readiness of middle school students for high school admission. These exams aim to select students with the highest levels of knowledge and skills among a large pool of applicants for a limited number of elite schools. Although the implementation of such exams varies across countries, since 2018, Turkey has adopted a centralized examination system known as the High School Entrance Exam (LGS). The sustainability of these exams in line with their intended purposes depends on factors such as the examination process, test items, and instructions. However, it is not always possible to claim that the items used in these exams adequately reflect the intended constructs. Therefore, the present study aims to examine the construct validity of the LGS exams administered in Turkey in 2022, 2023, and 2024. To this end, a research sample of 32,400 students was randomly selected by the Ministry of National Education. Data from 10,800 students for each year (2022, 2023, and 2024) were analyzed. Due to the dichotomous (0–1) nature of the data, Tetrachoric Exploratory Factor Analysis (TEFA) was employed instead of traditional Exploratory Factor Analysis. The findings indicated that the explained variance rates for the 2022, 2023, and 2024 LGS exams were 42.8%, 47%, and 42.2%, respectively. Furthermore, the test items were found to load onto two distinct factors. Upon examining the content of these factors, it was determined that the first factor primarily consisted of items assessing reading comprehension skills, whereas the second factor included items that required the application of knowledge. The results suggest that the LGS exams administered in 2022, 2023, and 2024 predominantly consist of items measuring reading comprehension skills, and that these items account for the majority of the explained variance.