statistician drowning in river of average depth 25 cm
DESCRIPTION
Statistician drowning in river of average depth 25 cm. A.N. Whitehead. “The things directly observed are, almost always, only samples. We want to conclude that the abstract conditions… also hold for all the other entities which… appear to us to be of the same sort.”. - PowerPoint PPT PresentationTRANSCRIPT
March 29, 2007 KNAW Lecture 1
March 29, 2007 KNAW Lecture 2
Statistician drowning in river of average depth 25 cm
March 29, 2007 KNAW Lecture 3
A.N. Whitehead
“The things directly observed are, almost always, only samples. We want to conclude that the abstract conditions… also hold for all the other entities which… appear to us to be of the same sort.”
March 29, 2007 KNAW Lecture 4
Observation in the face of variation
Video ergo lego
(I see therefore I select)
March 29, 2007 KNAW Lecture 5
Effect of selection on observation
March 29, 2007 KNAW Lecture 6
March 29, 2007 KNAW Lecture 7
Vulnerability Analysis of Spitfires (sample: 15/400)
March 29, 2007 KNAW Lecture 8
Composite of hits
*
* ***
*
*
***
**
*
*
Abraham Wald
Advice:
March 29, 2007 KNAW Lecture 9
Numbers speak for themselves. Not.
March 29, 2007 KNAW Lecture 10
…as we know,
there are known knowns;
there are things we know we know.
We also know there are known unknowns;
that is to say,
we know there are some things
we do not know.
But there are also unknown unknowns—
the ones we don’t know we don’t know.Donald Rumsfeld
Another anatomy of missingness
(set to music, see NPR website)
March 29, 2007 KNAW Lecture 11
…as we know,
there are known knowns;
there are things we know we know.
We also know there are known unknowns;
that is to say,
we know there are some things
we do not know.
But there are also unknown unknowns—
the ones we don’t know we don’t know.Donald Rumsfeld
Non-missing
MCAR/MAR
Non-ignorable
Translation into modern statistics
March 29, 2007 KNAW Lecture 12
Summary of Part A
1. Variation in everyday life2. Observation is selection3. Random selection gold standard 4. Lack of randomness is challenge
to valid inference
March 29, 2007 KNAW Lecture 13
Causal assertion. Is it testable?
March 29, 2007 KNAW Lecture 14
Question Testable version
March 29, 2007 KNAW Lecture 15
Example from Science (February 23, 2007)
Title Redefining the age of ClovisFront page Flints (pictures)Page 1045 Summary paragraphPage 1067 News storyPage 1122—1126 Article (numbers)
Very different “flavor” for each section
March 29, 2007 KNAW Lecture 16
Observational vs experimental studies
Characteristic Observational ExperimentOrientation Retrospective
ProspectiveResearcher control Less MoreSelection bias Big problem Less Confounding Present AbsentRealism More LessCausal plausibility Weaker StrongerAnalysis More complicated LessEthical issues Fewer MoreInferenceWeaker Stronger
March 29, 2007 KNAW Lecture 17
Usual state of nature
Effect?
March 29, 2007 KNAW Lecture 18
APHA Journal, May, 2004
March 29, 2007 KNAW Lecture 19
Conclusion: we always need to ask:
1. What is the question?
2. Is it measurable or testable?
3. Where will I get the data?
4. What do I think the data are telling me?
March 29, 2007 KNAW Lecture 20
1. Variation is fact of life
2. “Population” as model
3. Regression to the mean
4. Random selection
5. What is the question?
6. Testable question?
7. Association or causation
8. Causation through randomization
Experience can benefit from everyday statistics
Variation
Causation