Solving Normal-Distribution Probability Problems with and without Diagrams

James CorterTeachers College, Columbia University
Chenmu XingTeachers College, Columbia University

Abstract

Two studies are reported that explore the usefulness of diagrams for solving normal distribution probability problems. Study 1 analyzed performance on normal probability problems from a midterm exam in a graduate level introductory statistics course. For each problem, we coded whether or not students used a diagram in their work, and whether the diagram was correct, and complete. We also coded correctness of the answer and "procedural correctness" - whether the answer used a correct solution strategy (even if computational errors resulted in the wrong answer). Use of a diagram was associated with procedural correctness of the solution. Study 2 gave normal probability problems to an online sample of participants screened to have had at least one statistics course. Participants provided with a diagram "hint" did better than participants given no hint. The results add to the body of research showing facilitative effects of diagrams in problem solving.

Files

Solving Normal-Distribution Probability Problems with and without Diagrams (1 KB)



Back to Table of Contents