Introducing Statistical Inference: Design and Evaluation of a Learning Trajectory

Abstract

The increasing amount of data in media over the last year—think of COVID—illustrates the necessity for students to become statistically literate—including interpreting inferences. Drawing inferences involves making data-based claims under uncertainty when only partial data are available. However, inferences are challenging for students in Grade 10 and higher. This thesis focused on the question: How can a theoretically and empirically based learning trajectory introduce 9th-grade students to statistical inference? To answer this question, we used a design-based research approach, complemented with a case study into learning statistics from and with technology. The design of the trajectory was informed by theories on repeated sampling and statistical modeling using a black box paradigmatic context. The learning trajectory was implemented in teaching practice during three interventions. A pre- and posttest were designed to evaluate the trajectory’s effects in the large-scale final cycle. A national and international comparison of student results showed that students who took part in the learning trajectory (N = 267) scored significantly higher on statistical literacy than the comparison group that followed the regular curriculum (N = 217), in particular, on the domain of statistical inference. We also observed positive effects on other domains of statistical literacy. These findings suggest that current statistics curricula for grades 6–9 can be enriched with an inferential focus. The benefit of this early introduction is that students learn more about inference and not less about the other domains of statistical literacy, to anticipate for subsequent steps in students’ statistics education.