> From: Savian
> Sent: Thursday, December 17, 2009 11:28 PM
> Subject: ClimateGate and us
> First off, this is not a political commentary: it really isn't.
> The issue that researchers
>>>---> may <---<<<
> have falsified data to prove a point
> shocks me. This is hitting at both the statistician level and the
> programmer level. The code appears to validate the suspicions and now
> data researchers are indicating that only selective data was used.
I will answer w/tongue in cheek, to make the point:
"How many PhDs does it take to construct a bad design?"
> My questions are 2-fold:
> 1. How does that impact us as programmers and statisticians?
what we do is often or mostly re-search;
how many of us are actually doing original search?
There are theories and attempts to duplicate the original-search results
or original-search with emphasis on other factors.
> 2. Has anyone been put in a position of being asked to change data
> or programs to alter a prediction?
and eventually everyone squinted carefully at the complexity being
began to grok what kind of work would be involved to explain the
and decided to unfix: return to original data.
Q: What is science?
A: an ongoing dialogue.
Take study results out of their context and they are sure to be
when presented w/extrapolated editorial spin.
Ron Fehd the data mechanic CDC Atlanta GA USA RJF2 at cdc dot gov
||12/18/2009 3:39:46 PM