<![CDATA[This study analyzes the representation of higher-order thinking skills (HOTS) in elementary school end-of-semester assessment items and examines the design characteristics of items classified as higher-order tasks. A quantitative descriptive design supported by qualitative document analysis was employed. Data consisted of 240 assessment items collected from 12 elementary school teachers in Jakarta across four core subjects: Indonesian Language, Mathematics, Science, and Social Studies. Each item was coded using the revised Bloom's taxonomy and grouped into lower-order thinking skills (C1–C3) and higher-order thinking skills (C4–C6). The classification rubric was validated by assessment specialists, and coding was conducted by independent raters through calibration, independent coding, and consensus procedures. The findings reveal that assessment items were strongly dominated by lower-order thinking skills, representing 72.5% of total items, while only 27.5% were classified as higher-order thinking items. Within the HOTS category, analyzing (C4) was the most frequent cognitive level (45.5%), followed by evaluating (C5, 36.4%), whereas creating (C6) appeared only marginally (18.1%). The distribution also varied across subjects, with Indonesian Language and Science showing higher proportions of HOTS items than Mathematics and Social Studies. Qualitative analysis indicates that most HOTS items were presented as open-ended or short-answer tasks embedded in everyday contexts, but the use of visual, numerical, graph-based, and data-rich stimuli remained limited. The study concludes that elementary assessment practices have not fully translated the higher-order thinking orientation of contemporary curricula into summative test design. Implications emphasize that teacher professional development should move beyond general curriculum socialization toward sustained assessment-literacy training, collaborative item review, use of diverse stimuli, and school-based supervision. ]]>
