Senior High School Students’ Statistical Discourse in Interpreting JAMOVI-Generated Outputs: A Commognitive Case Study

Abstract

Statistical software can broaden students’ access to data analysis while also shaping what they notice, say, and treat as evidence. This qualitative case study examined how Senior High School students constructed statistical meaning while interpreting outputs generated in JAMOVI. Guided by Sfard’s commognitive theory, the analysis focused on four features of discourse: word use, visual mediators, narratives, and routines. Three purposively selected student groups completed four structured tasks involving descriptive statistics, histograms and boxplots, correlation, and a paired-samples t test. The data corpus comprised completed worksheets, audio-recorded group discussions, and the corresponding JAMOVI outputs. Data were coded iteratively in NVivo and analyzed through commognitive discourse analysis. The students used increasingly formal terms—including mean, standard deviation, correlation coefficient, and p-value—but their explanations remained predominantly procedural. Across tasks, JAMOVI outputs frequently operated as epistemic anchors: students read values directly from tables or graphs, named the statistic, and moved quickly to a conclusion. Visual mediators supported participation but also encouraged surface-level interpretations based on salience, shape, or threshold rules. Four recurring explanatory patterns were identified: procedural repetition, definition-by-repetition, negotiated meaning, and value-based interpretation. Reasoning routines were stable, typically following a read–name–interpret–conclude sequence. The findings suggest an emerging but incomplete shift from ritualized toward exploratory statistical discourse. Because the study involved three groups in one bounded context and did not compare JAMOVI with another platform, the findings are analytically transferable rather than statistically generalizable and cannot isolate software effects from conceptual difficulty. The study highlights the need for teacher-led prompts that position software outputs as objects for explanation, comparison, and critique rather than as final answers.