Title: A comparative study of authentic assessment design across science and humanities subjects under Competency Based Curriculum
Authors: Dr. Arinaitwe Julius, Ahumuza Audrey, Asiimwe Isaac Kazaara
Volume: 10
Issue: 7
Pages: 168-177
Publication Date: 2026/07/28
Abstract:
This study undertook a comparative, school-clustered examination of how authentic assessment was designed and implemented across Science and Humanities subjects by secondary school teachers operating under Uganda's Competency Based Curriculum (CBC). Grounded in the recognition that authentic assessment - tasks that mirror real-world application of competencies - sits at the core of CBC's shift away from rote, examination-driven pedagogy, the study asked whether subject area, teacher training in CBC-aligned assessment, teaching experience, and school type systematically predicted the quality of authentic assessment design, and whether this quality clustered meaningfully within schools. A cross-sectional survey of 240 teachers (120 Science, 120 Humanities) drawn from 24 purposively sampled secondary schools was analysed using a composite Authentic Assessment Design Quality Score (ADQS) built from three validated subdomains: Task Authenticity, Real-World Relevance, and Rubric/Competency Alignment. Univariate description characterised the sample and outcome distribution; independent-samples t-tests and a chi-square test of association compared Science and Humanities teachers on each subdomain and on the use of portfolio- or performance-based assessment; Pearson correlations examined interrelationships among subdomains; and a two-level linear mixed-effects model with teachers nested within schools quantified fixed effects of subject area, training, experience, and school type on ADQS while partitioning variance into school-level and teacher-level components. Humanities teachers recorded significantly higher overall ADQS than Science teachers (mean difference ? 2.2 points, p = 0.008), driven primarily by Real-World Relevance (p < 0.001), while Rubric/Competency Alignment did not differ significantly between the two groups. CBC assessment training emerged as the strongest fixed-effect predictor of design quality (? = 4.13, p < 0.001), exceeding the effect of subject area itself. The mixed-effects model further revealed that 65.4% of the total variance in ADQS resided at the school level (ICC = 0.654), indicating that institutional context - rather than individual teacher characteristics alone - substantially shaped assessment design practice. The study concluded that authentic assessment design quality under CBC was jointly a function of subject-specific pedagogical tradition and school-level implementation culture, and that targeted, school-based training investment was likely to yield larger gains than individual-level interventions alone. Three recommendations were advanced concerning differentiated subject-specific training, school-level capacity strengthening, and closer alignment of Science practical assessment with real-world application.