Many concepts in undergraduate statistics are hard to learn, and the misconceptions harbored by students are difficult for teachers to identify. In interviews with instructors, we learned that after traditional instruction, students still struggle with pervasive misconceptions. To address this challenge, we have connected undergraduate statistics education with promising research methodologies from cognitive and learning science.
Our goal is to develop outlines of the cognitive steps required to solve statistical inference problems, using cognitive task analysis (CTA). Think-aloud interviews of experts validate the steps; interviews of students reveal which cognitive skills students lack and highlight misconceptions, helping instructors target interventions to specific missing cognitive skills.
This study will provide useful data to create new assessments and target instruction to specific cognitive skills students struggle with in statistical inference. More generally, we hope this model can advance learning science by providing a new domain-specific model that can be used in an Intelligent Tutoring System (ITS).
|