A Cognitive Measurement Model for Pre-Service Teachers’ Higher Order of Thinking Skills in Digital Literacy.Begimbetova GuldanaAtymtaevnaauthorHeriRetnawatiauthorThis study aims to; (1) Develop a HOTS digital literacy skills measurement model for prospective teachers at KazNPU, hereinafter referred to as the DLMCI; (2) Evaluate the quality of the digital literacy skills measurement instrument; and (3) Provide empirical evidence regarding the feasibility of measuring pre-service teachers' digital literacy skills, and (4) Analyse the pre-service teachers’ digital literacy HOTS profiles. This study employed a quantitative research method in the context of R&D. The subjects were 151 prospective teachers (n = 151) at KazNPU. The data collection instrument consisted of a two-part test: a Tier-Two Item Test and a Situational Judgment (SJ) in the form of an anti-cyberbullying video. Data analysis was conducted using RStudio version R 4.4.3 to examine the parameters of the developed items. The results showed that (1) the Cognitive Measurement Model, the DLMCI, was successfully developed to assess pre-service teachers’ HOTS digital literacy. The (2) DLMCI has very high content validity, with an Aiken V value of 0.94 and a composite reliability (McDonald's Omega) of α = 0.803. In classical test theory (CTT), Two-tier choice items showed an average a = 0.510, indicating moderate discrimination, and an average difficulty level of b = 0.643. In IRT, the instrument is multidimensional: MDIFF ranges from -1.27 to +1.29, and MDISC between 0.78 and 3.92. The residual correlation r_res. ≈0 supports local independence; the odd￾even theta coefficient (r = 0.925) confirms parameter invariance. Most of the distractors in the items functioned well, indicated by a selection rate above 5%, except for one case with a percentage of 1.3%. The measurement model that best fits the data is the Generalised Partial Credit Model (GPCM). In a logistic scale [-3 to +3], Theta estimates for measuring higher-order thinking skills (HOTS) of digital literacy of prospective teachers at KazNPU are in the range of -3.01 ≤ θ ≤ +2.4. The (3) DMCI demonstrated feasibility when implemented in a real-world setting without hindrances; it is both efficient and environmentally friendly. The (4) HOTS profile of digital literacy among prospective teachers at KazNPU is divided into four categories: Low (-1.93 ≤ θ < -0.85), Medium (-0.85 ≤ θ < 0.23), High (0.23 ≤ θ < 1.31), and Very High (1.31 ≤ θ ≤ 2.4). The minimum and maximum Theta values are -3.00 and +2.4, respectively, with an average Theta of -0.3, which falls into the “Low” category. The Low, Medium, High, and Very High categories are represented by 30 (19.86%), 67 (44.37%), 40 (26.50%), and 14 (9.27%) pre-service teachers, respectively. Therefore, this study recommends that the Ministry of Education in Kazakhstan utilise the DLMCI for data-driven decision-making.Teknologi PendidikanEvaluasi Pendidikan2026-01-07Sekolah Pascasarjana;Penelitian dan Evaluasi PendidikanThesis