From Data Handling to Evidence-Based Reasoning: Latent Profiles of Junior High School Learners’ Statistical Reasoning in Ghana.

Abstract:

This study examined how Junior High School learners’ statistical reasoning is configured across data representation, computation, interpretation, comparison, and evidence-based conclusion drawing. Drawing on a cross-sectional assessment of 437 JHS 2 learners from 13 schools in Bolgatanga Municipality, Ghana, the study used latent profile analysis to identify learner subgroups and multinomial logistic regression to examine predictors of profile membership. A three-profile solution was retained: emerging statistical reasoning, developing procedural reasoning, and advanced balanced reasoning. The largest profile showed relatively stronger computation and data representation but weaker interpretation, comparison, and evidence-based conclusion drawing, indicating that procedural data handling does not necessarily translate into statistical reasoning with evidence. Prior mathematics achievement, textbook access, private school attendance, data-task exposure, and school location were associated with profile membership. The study contributes to statistics education research by providing a construct-linked, person-centred account of how school-level statistical reasoning varies within an under-represented African basic education context and by showing how profile evidence can inform differentiated instruction in data-rich mathematics classrooms.