Implementing Mystery Box Media in Grade 4 Science: Classroom Action Evidence on Participation and Achievement

Abstract

This classroom action study examined the implementation of a low-cost Mystery Box medium and the accompanying changes in participation and achievement in Grade 4 Natural and Social Sciences (IPAS) at a public primary school in Indonesia. The participants were 29 pupils, comprising 10 boys and 19 girls. Conducted from November to December 2025, the study followed a plan-act-observe-reflect sequence across a pre-action stage, two Cycle I meetings, and Cycle II. The intervention combined teacher explanation, concealed instructional prompts, pupil selection and oral response, guided feedback, individual worksheets, and written assessment. Individual score records were reanalyzed using means, standard deviations, medians, ranges, and the proportion reaching the school mastery threshold of 75. Observation records were summarized as 1-4 affective and psychomotor indices, while a 10-item teacher-activity checklist described implementation. Mean scores were 53.79 (SD = 22.74) at pretest, 70.93 (SD = 15.99) at Cycle I Meeting 1, 68.97 (SD = 28.58) at Cycle I Meeting 2, and 88.97 (SD = 13.72) at Cycle II. Mastery increased monotonically from 24.14% to 41.38%, 58.62%, and 89.66%, exceeding the 80% classroom success criterion. Teacher-activity completion rose from 5 of 10 to 8 of 10 and then 10 of 10 items, while both observation indices increased. From baseline to Cycle II, 24 pupils improved their recorded scores and 19 newly crossed the mastery threshold. The evidence supports local classroom improvement when the Mystery Box is used as a structured prompt-response-feedback routine, not merely as a novel object. Because there was no comparison group, the assessments were not documented as psychometrically equivalent, and no delayed test was administered, the findings do not establish causal efficacy or durable learning.