Penerapan Pembelajaran Berdiferensiasi Produk pada Keterampilan Menulis Puisi Siswa Kelas X-5 SMAN 3 Jember

Poetry writing requires precise diction, figurative language, and control of sound and form, yet students often struggle to produce original work under time constraints. Evidence on product-based differentiated instruction remains largely descriptive. This classroom action research examined changes in poetry-writing achievement among 28 Grade X students at a public senior high school across two cycles of planning, action, observation, and reflection. Students selected one of three products: independent written poems (Product A), poetry posters (Product B), or digital poetry (Product C). Pre-test and post-test writing tasks were assessed by one rater using an eight-aspect rubric. Additional data came from classroom observations, product assessments, student worksheets, and teacher reflections. Quantitative analysis used mean scores, mastery percentages based on the minimum mastery criterion (KKM 75), and normalized gains calculated from class means; qualitative data were analyzed using the interactive model. In Cycle I, the mean declined from 68.4 to 62.5 (N-Gain = −0.19, negative), accompanied by limited supervision, incomplete worksheets, and concerns about originality. In Cycle II, spaced seating, an originality statement, diction cards, and supervised digital references were introduced during product development. The mean increased from 51.1 to 79.6 (N-Gain = 0.58, moderate), while mastery rose from 0% to 64.3% (18 of 28 students). However, the Cycle II baseline was 17.3 points lower than the Cycle I baseline, making N-Gain sensitive to initial scores rather than an effect size. The findings highlight the importance of structured supervision, including for independently completed assessment tasks. The study provides descriptive two-cycle evidence and explicitly documents supervision, originality, and time-management challenges not reported in the reviewed studies. Because multiple adjustments occurred simultaneously, assessment prompts were not equivalent, scoring relied on one rater, and no comparison group was included, the results indicate descriptive changes, not causal effects of product differentiation.