Teaching Statistical Concepts with Computational Software

1
Teaching Statistical Concepts with Computational
Software

• Robert H. Carver
• Stonehill College
• Decision Sciences Institute
• Washington DC November 24 2003

2
Intro Course Context

• Objective(s)
• Emphasis on reasoning
• Service course vs. intro to major
• First course?
• Last course?
• Student preparation to read statistics texts

• Constraints
• Loosening
• Student computing
• Classroom technology
• Tightening/Fixed
• 13 weeks
• Budgets for software
• Cognitive budget stats vs software

3
Strengths of Computational Software

• In the classroom workplace
• Permits emphasis on interpretation statistical
  reasoning rather than computation
• Vehicle for constructing concepts
• Ease of replication and simulation permitting
  illustrative comparison

4
Strengths of Computational Software

• Individualized results
• Integration with texts technologies
• Ad hoc interactivityteachable moments
• Realtime in-class prediction falsification

5
Some introductory concepts

• Variation
• Shape
• Center
• Spread
• Randomness
• Distributions
• Inference

• Density
• Random Sampling
• Bias
• Null hypothesis
• Error
• Linear ( non-linear relationships

6
An exercise about SHAPE

• Variable Airline delays
• Class discussion
• Common assignable causes
• Departures vs. arrivals
• Sources of data
• Class predictions about variation
• Class predictions about SHAPE

7
Data Source BTS Tool
8
BTS Data Available
9
Exercise

• Analysis of departure delays at one airport
• Discussion of process dynamic
• Speculation about likely shape
• Sketch predictions
• Each student selects an airline arbitrary set
  of dates
• Download CSV file ? Transfer to Minitab
• Construct histogram

10
Typical result
11
Results for 5 students
12
Modified Graph
13
Arrival Delays
Lets talk about CENTER and SPREAD now
14
DIY

• Which concepts are most elusive?
• Whats your best approach?
• Which of these can we better approach
  experientially/interactively?
• Role of illustrative data (raw/ cooked/
  synthetic/ simulated)
• How can a given approach exploit software at hand
  data student background?

15
Important connections

• BTS website
• http//www.bts.gov/ntda/oai/DetailedStatistics/
• Author website
• http//faculty.stonehill.edu/rcarver/carverindex.h
  tm
• Email author
• rcarver_at_stonehill.edu
• Thanks to Jennifer Karp, Stonehill 2005