Improving Inferential Statistics Teaching Methods to Accommodate Millennial Learning Styles Dr. Buddy Bilbrey
Purpose of the work Create a method that reaches the students in a way that enforces learning Quantify changes to drive process improvement in teaching inferential statistics
Motivations from last year Session last year two striking methods for instruction were proposed 1) skip early inferential statistics 2) write/use software to provide tools for students Both methods were driven by the lack of time to cover the undergraduate topics inferential statistics AACSB constraint – cover topics in detail reasonable to industry expectations
Difficulties from THIS year Variables hard to control over multiple semesters Technology Communication with Math Department Lack of Excel Skills OFAT for this experiment
Traditional Method Assign homework relative to the lectures Homework will be done on students’ time Lecture with examples and notes Homework is designed to test knowledge, not teach the material All work graded after submission
What was happening? Students were bored/uninterested in class Life is MUCH more interactive for them today Communications are instant Immediate access is the “norm” Examples: Facebook, texting, Snapchat, etc. Preconceived notion that “Statistics is hard”
New Approach Students are asked questions for the problems during class – answers are unknown and problems are algorithmically generated Forces students to follow and keep up with the class Use end lecture time for hands-on problems (Short extra credit problems to encourage effort) Can’t lean on neighbors for answers
Preliminary Results variance reduced
Preliminary Results Total Points – Normalized by Max Pts
Current Problems – Future Work Noticed that the students won’t draw the normal curve Homework is broken down into very small problem sets vs. the traditional (Spring 2018) Trying to guess the solutions instead of working the problems Which data to remove? Dropped students? Online lectures?
Questions?