Answers are grounded in this paper and cite their sources.
It is well known that students have difficulties drawing adequate conclusions from data. This study supports this result: about one quarter of 1.200 high school students in Germany were unable to correctly assign more than three out of six hypotheses to the data presented. What our study also found and which was not investigated before, is that the reverse procedure also leads to difficulties: almost one third of the same students were unable to correctly predict the expected data based on hypotheses given. Further, the number of correct answers increases from grade level 7 to 10. Ex post we also found that the level of instruction is heavily associated with the results: While around 80% of the students in more advanced classes made correct data predictions, the number dropped to around 40% in less advanced classes of the same age group. Finally, students with low data prediction abilities are less likely to determine a correct hypothesis for given data. This uncovers an important influence that impacts data-driven reasoning: If students are not able to predict their data expectations for a given hypothesis, they struggle to compare their hypothesis with obtained data and, thus, may come to wrong conclusions. For teaching practice, this means that teachers must expect that students at different levels of instruction have very different prior knowledge of data-based reasoning.
No takes yet. Share an insight, caveat, or question.
Chroszczinsky et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: