BANISHING THE THEORY-APPLICATIONS
DICHOTOMY FROM STATISTICS EDUCATION
Larry Weldon
Department of Statistics and Actuarial
Science
Simon Fraser University, Burnaby, CANADA
?
“Issue” Questions
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
Implications for Stats Course Taxonomy
?
?
Some Questions
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
?
Basic Theory: More than Math?
•
•
•
•
•
•
Obs Study vs Experiment
Distributions: Averages and Variability
Random Sampling, Estimation
Independence (and dependence)
Time Series
Statistical Significance
Example: Dependence
When does a portfolio of stocks have enough
independence to provide stability of return?
One needs to understand the
dependence-independence concept
A & B independent -> P(A&B)=P(A)*P(B)
is not enough
Basic Theory: More than Math?
•
•
•
•
•
•
Obs Study vs Experiment
Distributions: Averages and Variability
Random Sampling, Estimation
Independence (and dependence)
Time Series
Statistical Significance
Theory =
Generally Applicable Concepts
(Much more than Mathematics)
?
Question Answered?
• Is Mathematical Statistics = Theory of
Statistics?
• No! Theory is Generally Applicable Concepts.
More Questions ->
?
?
Some Questions
?
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
Levels of Expertise
• Generalist
– requires stats appreciation
• Practitioner
– requires stats appreciation
– requires stats methods & hazards
– requires exposure to expert capability
• Expert
– all the above and much more
Cumulation Model of Statistics Education
Do Practitioners need
“Appreciation” Course?
• Overview for when-to-consult
• Motivation to integrate with applied focus
• Awareness of naïve user (hazards)
Experts need “stats appreciation”?
• Yes, because they need informed choice of
career
• Real expert statisticians are generalists as well
as specialists, so they can absorb context
• Need to explain to naïve user
Experts need “Practitioner” training?
• of course!
• early exposure helps education
• no need to learn everything the hard way
Proposed Course Sequence:
Appreciation -> Practitioner -> Expert
Questions ->
?
Some Questions
?
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
Motivation Clusters?
• Does “auto engine size” or “golf participation”
interest biologists?
• Does “potato pest resistance” or “threatened
species of birds” interest social scientists?
Contextual Interest is Important for Seeking Data-Based Information
Stats Streams for Major Groups?
Context Material Matters!
Because Context-Major Students chose context!
Minimal Context Segregation for Courses …
• General (Wide Focus)
(segregation by context …
• Life Science
not by methods introduced)
• Social Science
Important for early courses,
perhaps not feasible for higher level ones.
Questions ->
?
Some Questions
?
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
Undergrad Course Structure?
•
•
•
•
•
•
Statistics 1 (life) Statistics 1 (social) Statistics 1 (general)
(Appreciation courses)
Statistics 2 (life) Statistics 2 (social) Statistics 2 (general)
Statistics 3 (life) Statistics 3 (social) Statistics 3 (general)
(Practitioner Courses)
Statistics 4 (general)
Statistics 5 (general)
Statistics 6 (general)
More courses where numbers permit.
(Expert courses)
Note:
1. No specialized technique courses like Nonparametrics, Time Series,
Experimental Design, Quality Control, Bayesian Analysis
2. No “service” stream
3. No “baby” stat courses
Experts need “MORE” not “DIFFERENT”
Experiential Learning&Teaching
• Sequence of Projects
– data collection
– data analysis
– data summary
• Techniques as Required
• Concepts as they Arise
Example ->
Experiential Learning Examples
• Sports Leagues
– probability
– measures of variability
– simulation
• Daily Delivery Schedules
– censored data (demand exceeds sales)
– parametric variability, prediction
– optimization
Many concepts and techniques will be introduced
Questions ->
Some Questions
• Is Mathematical Statistics = Theory of Statistics?
• Expert vs Practitioner vs Generalist
different stats education?
• Motivation for practitioner grps?
• What undergrad course sequences?
– for practitioners
– for experts
• Motivation for Stats Instructors?
Motivation for Stats Instructors?
•
•
•
•
•
Case Studies/Projects – experiential learning
Discussion & Presentations
Novelty and Creativity encouraged
Active engagement of students and instructors
Better Use of Instructor Expertise & Experience
Motivation for Stats Instructors?
•
•
•
•
•
Case Studies/Projects – experiential learning
Discussion & Presentations
Novelty and Creativity encouraged
Active engagement of students and instructors
Better Use of Instructor Expertise & Experience
Motivation for Stats Instructors?
•
•
•
•
•
Case Studies/Projects – experiential learning
Discussion & Presentations
Novelty and Creativity encouraged
Active engagement of students and instructors
Better Use of Instructor Expertise & Experience
Motivation for Stats Instructors?
•
•
•
•
•
Case Studies/Projects – experiential learning
Discussion & Presentations
Novelty and Creativity encouraged
Active engagement of students and instructors
Better Use of Instructor Expertise & Experience
Motivation for Stats Instructors?
•
•
•
•
•
Case Studies/Projects – experiential learning
Discussion & Presentations
Novelty and Creativity encouraged
Active engagement of students and instructors
Better Use of Instructor Expertise & Experience
Summary
• Experiential Learning is Authentic Learning
• It can be motivating for most students and
instructors
• It can be efficient in reducing the number of
courses offered
• Levels of expertise correspond to number of
courses completed (not math level)
• Downside? Requires instructors with an interest
in, and experience with, using statistical theory.
Thanks for attending this session. Comments? weldon@sfu.ca