Role of examples in how students learn to categorize statistics word problems

In Experiment 1, students who studied example word problems that were grouped by t test, correlation, and chi-square were more likely to sort subsequent problems on the basis of structure and less likely to sort on the basis of surface characteristics than students who received no examples. In Experiment 2, this pattern was strongest when students studied structure-emphasizing rather than surface-emphasizing examples. In Experiment 3, students who studied and practiced 4 structure-emphasizing worked-out examples of t test and correlation problems were more likely to apply the appropriate statistical test correctly to subsequently presented statistics word problems than students who had studied surface-emphasizing examples. This pattern was strong for lower but not for higher ability students. Implications of a schema construction theory are discussed.