Upload
phungliem
View
217
Download
3
Embed Size (px)
Citation preview
The Half-Life of a Social Statistician
Andrew Gelman
15 Nov 2012
Andrew Gelman The Half-Life of a Social Statistician
control
treatment
"before" measurement, x
"afte
r" m
easu
rem
ent,
y
Andrew Gelman The Half-Life of a Social Statistician
Estimated partisan bias in previous election
Est
imat
ed p
artis
an b
ias
(adj
uste
d fo
r st
ate)
-0.05 0.0 0.05
-0.0
50.
00.
05
no redistricting
bipartisan redistrict
Dem. redistrict
Rep. redistrict.
. .. .
.
..
.....
...
.
.
.
...
.
.
..
..
.
.
.. .
. .
.
.
.
.
.
..
..
.
. .
.
.
. .
..
..
.
.. ..
.
.
.
..
.
..
..
.
.
. .
.
.
.
..
..
.
.
.
.
.
.. .
.
.
..
.
.
.
. .
..
..
.
..
.
. .
.
.. o
o
o
o
o o
ox
x
x
x
x
xx
x
x
x
•
•• •
•
•
•
•
•
•
•••
•
•
(favors Democrats)
(favors Republicans)
Andrew Gelman The Half-Life of a Social Statistician
grade
corr
elat
ion
1 2 3 4
0.8
0.9
1.0
controls
treated
Andrew Gelman The Half-Life of a Social Statistician
Hierarchy of stories:
1. Joke, or fiction such as Harry Potter
2. Urban legend
3. Dubious story (the lost soldiers in the Alps)
4. Plausible but undocumented story (e.g., Iraqs weapons ofmass destruction)
5. Documented anecdote
Andrew Gelman The Half-Life of a Social Statistician
Some reasons to believe:
I Mathematical theory (e.g., asymptotic convergence)
I Computer simulations (e.g., approximate coverage of intervals)
I Solutions to toy problems (e.g., 8 schools)
I Improved performance on benchmark problems
I Cross-validation and external validation
I Success as recognized in a field of application
I Success in the marketplace
Andrew Gelman The Half-Life of a Social Statistician