Carnegie Mellon University

Eberly Center

Teaching Excellence & Educational Innovation

Assessment Of Student Learning And Misconception Identification In An
Introductory Statistics Course

Good assessments of student learning can help faculty quickly target student
misunderstandings, or evaluate their own instruction. However, students’
misconceptions may differ from faculty’s expectations. For example, students may select
incorrect answers for particular reasons unknown to the faculty, and consequently
assessment results may be misleading as a guide to improving instruction.
Our goal was to apply (an abbreviated version) of an assessment-design protocol from
Physics education (Adams & Wieman, 2010) to Reasoning with Data, a general
education introductory level statistics course at Carnegie Mellon. Phases include
interviewing current students to probe for misconceptions, devising multiple-choice
questions that distinguish correct answers from specific distractors, and interviewing
additional students to help validate the proposed assessments. We report findings on
two levels: (1) common misconceptions and the corresponding assessments, and (2)
lessons learned from our experience in adapting this protocol from Physics to a pilot
study in Statistics.

Nugent, Rebecca
Reinhart, Alex
Wieczorek, Jerzy
Elliott, Peter
Hyun, Sangwon
DC, Statistics & Data Science