Quantitative Analysis in SoTL

1
Quantitative Analysis in SoTL

• Chris Cooper
• Department of Political Science Public Affairs
• Western Carolina University

2
Todays Agenda

• A Puzzle
• The Logic of Quantitative v. Qualitative Analysis
• Questions you can answer
• 10 Rules for Quantitative Analysis
• Stat packages

3
The Puzzle

• Why is there so much less quantitative work in
  SoTL journals than in traditional disciplinary
  journals?

4
The Basics

• Quantitativenumbers qualitativeno numbers.
• Quantitative analysis should not be viewed as
  different than qualitative analysisrelies on the
  same logic.

5
Questions you can answer

• Is there a difference?
• If yes
• Which direction?
• How much difference?

6
10 Rules for Quantitative Analysis

• Keep it simple
• Know your units
• Know thy dependent variable
• Dont throw away information
• Think about operationalization
• Know the difference between statistical and
  substantive significance
• Use randomization to your favor
• Think about how you present data
• Graphs are more powerful than tables
• Choose the right tool

7
Keep it simple

• Does A affect B?

8
Know your units

• Think about your unit of analysis
• Specify your unit of analysis
• Dont mix your units
• Corollary Choose the right number of units and
  decimal places

9
Know thy dependent variable

• Describe it, graph it, know it.
• If its not interesting, then youve got a boring
  topic.

10
Dont throw away information

• Dont dichotomize continuous variables

11
Think about operationalization
12
Statistical v. substantive significance

• Theyre not the same thing know the difference.
• If youre running regressions, interpret the
  coefficient.
• If youre running logistic regressions, compute
  predicted probabilities.

13
Whats the substantive significance?
14
Better
15
Example

• Morality and age are correlated.
• OR
• Among the elderly, mortality roughly doubles for
  each successive five year group.

16
Use randomization to your favor

• If you dont, learn fancier stats.

17
Think about presenting data
18
Lots wrong here
19
Sort by Something Meaningful
20
Sort by Something Meaningful
21
Graphs are more powerful than tables
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
But What About Regression Models?
26
But not all graphs are created equal

• Maximize your dataink ratio
• Dont use 3-D
• Dont use pie charts (theyre evil)
• I dont use pie charts, and I strongly recommend
  that you abandon them as well (Few, Show me the
  Numbers).

27
Choose the right tool
28
Review

• Keep it simple
• Know your units
• Know thy dependent variable
• Dont throw away information
• Think about operationalization
• Know the difference between statistical and
  substantive significance
• Use randomization to your favor
• Think about how you present data
• Graphs are more powerful than tables
• Choose the right tool

29
Stat packages

• Excel
• Spss
• SAS
• Stata
• Stat transferthe one program you cant live
  without.

30
Resources

• Anything by Edward Tufte (on graphs)
• Jane Miller, Chicago Guide to Writing About
  Numbers.
• Jane Miller, Chicago Guide to Writing About
  Multivariate Statistics.
• Neil Salkind, Statistics for People Who (Think
  They) Hate Statistics.