1. Jane Oppenlander David D. Reh School of Business Clarkson University
How Adapting an M&M Experiment to the Online Classroom Yielded a Useful Messy Data Set
Jane Oppenlander
David D. Reh School of Business
Clarkson University

2. Goals
Illustrate the challenges encountered when adapting a classroom activity to the online environment.
Show how the activity modified for online yielded a messy data set that is used to teach practical data science skills applicable in both learning environments.
Motivate you to:
Adapt your classroom activities to the online environment
Incorporate data science skills into introductory statistics courses

3. Background: Candy Experiments
Candy experiments are used to teach
Central Limit Theorem and sampling distributions: (Rossman, 1996), (Dyck and Gee, 1998)
Probability: (Rossman, 1996), (Magel, 1998)
Chi-squared tests: Magel (1996)
Simple linear regression and correlation: (Froelich and Stephenson, 2013)
One-way ANOVA: (Schwartz, 2013)
Implements GAISE recommendations
Foster active learning
Integrate real data with a context and purpose
Teach statistical thinking
Teach statistics as an investigative process of problem- solving and decision making.

4. Setting: Introductory MBA Statistics Course
Required course, class sizes 5-30, 10 week quarter
Classroom: Evening class, once a week, 3 hours 20 minutes
Online: Fully asynchronous
Students
Full-time and part-time
Undergraduate, graduate students, working professionals,
Each class uses their data for a one-way ANOVA
Data file containing results from all previous classes used for other classroom activities (cumulative data set)

5. M&M Classroom Activity One-way ANOVA - 1
Brief lecture on basic principles of designing experiments
Instructor provides
Research Question: Is there a difference in the number of pieces of candy in packages of M&M’s for three different varieties?
Limited number of fun size packages of M&M’s: plain, peanut, pretzel
Blank spreadsheet
Required information to collect
Activity constraints: time, allergies

6. M&M Classroom Activity One-way ANOVA - 2
Class Discussion
How many bags should be opened?
How should students be assigned bags/varieties?
What other information should be collected?
How should the spreadsheet be constructed?
Activity
Students each count their M&M’s and record their data
Resolve issues
Request spreadsheet modification
What did you learn about collecting data?
Demonstration of One-Way ANOVA with a different data set
Class M&M data is used for their summative assessment
Data added to the Cumulative Data Set

7. Modifications for Online Class
Students plan data collection through a synchronous session a week before the ANOVA topic
Reach consensus on instructions for conducting the experiment and formatting the spreadsheet
Post instructions on course site
Single serving size instead of fun size
Students must acquire their own M&M’s
Online discussion forum for resolving issues

8. Challenges in Online Implementation
Based on their geographic location, some students were not able to find single serving size of one of the varieties (usually pretzel)
Communication takes longer in asynchronous online classes compared to synchronous online and classrooms.
Students participation varies
Most difficulties arise from student participation and the asynchronous nature of communication

9. Characteristics of Cumulative Data Set
Mix of package sizes (fun size and single serving)
Outliers
Online students who brought the wrong size
Classroom students sometimes use holiday themed packaged different distribution of colors
Missing data
Both complete records and attributes

10. Cumulative Data Set Pedagogical Applications
Offers many possibilities for students to apply statistical methods and data science skills
Data visualization to identify outliers
Creating derived data
Selecting appropriate records for analysis
Dispositioning missing data
Documenting units and data definitions
Examples
Has the proportion of red M&Ms changed?
Create a data visualization to show to distribution of the total number of M&Ms/package by variety
Use data visualization to identify sharing size

11. Lessons Learned Students engage with data they have created
Experiential learning about data collection and management
Good for discussion forums
Cumulative data set is very flexible for classroom or assessment use
Fosters growth in non-statistical skills
Teamwork, communication, decision making
Efficient way to integrate other concepts
Data science, quality control, experimental design
Several iterations were needed until the online activity ran smoothly
Be transparent with students and ask for their feedback

12. References
Dyck, Jennifer and Nancy Gee. “A Sweet Way to Teach Students about Sampling Distribution of the Mean.” Teaching of Psychology, 25 (1996):
Froelich, Amy and W. R. Stephenson. “How Much do M&M’s Weigh?” Teaching Statistics, 35.1 (2013):14-20.
GAISE College Report ASA Revision Committee, “Guidelines for Assessment and Instruction in Statistics Education College Report 2016,”
Magel, Rhonda. “Increasing Student Participation in Large Introductory Statistics Classes.” The American Statistician, 50 (1996).
Magel, Rhonda. “Using Cooperative Learning in Large Introductory Statistics Classes.” Journal of Statistics Education , 6.3 (1998).
Rossman, Allan, Workshop Statistics: Discovery with Data. New York: Springer-Verlag, 1996.
Schwartz, Todd. “Teaching Principles of One-Way Analysis of Variance Using M&M’s Candy.” Journal of Statistics Education 21.1 (2013).