Supporting Fourth-Grade Students’ Statistical Reasoning Through a Flipped Classroom Intervention: A Mixed-Methods Study of Two Intact Classes

Developing statistical reasoning in primary education requires opportunities to connect data, representations, and context. This mixed-methods intervention study examined two intact fourth-grade classes in Türkiye (intervention n = 39; comparison n = 40). Six sessions combined pre-class videos, collaborative data investigations and digital representations; the comparison class continued regular teacher-led instruction. Recomputed mean total scores increased by 7.179 and 1.225 points, respectively, on a 0–64 scale. Exploratory baseline-adjusted regression yielded an intervention-minus-comparison contrast of 5.333 points (95% CI 2.274 to 8.393; p < 0.001). Pooled total-score alpha was 0.853 at pretest and 0.924 at posttest, with weaker consistency in some process subsets. Representation increased in both classes, while description showed little change. Interviews, diaries and classroom records illustrated valued preparation and access difficulties. One class per condition, nonrandom allocation and teacher confounding prevent causal attribution. The study contributes a documented implementation and a cautious integration of score patterns with student experiences.