This study investigated how the adoption of a constructivist model of teaching and learning and simple item analysis techniques can be used to explore the instructor's pedagogical content knowledge in teaching elementary statistics. Descriptive data (percent of students responding to multiple-choice test options) are provided that support the case for specific student statistical learning problems on the following topics: calculation and interpretation of measures of central tendency and variability, understanding of reliability and validity, interpretation of correlation coefficients, estimation of correlation coefficients from graphic scatterplots, and the selection of the best test-retest reliability scenarios. It is suggested that item analysis findings from multiple-choice examinations can be used to discover student conceptual misunderstandings, improve classroom instruction, and refine test-item writing. An attached table graphically displays the 10 findings. (Author/ND)