|
|
On 30 Dec 2006 14:19:46 -0800, <ekkilu@xxxxxxxxx> wrote in
<news:1167517186.523501.225390@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
in sci.lang:
> Brian M. Scott wrote:
>> On 30 Dec 2006 12:52:09 -0800, <ekkilu@xxxxxxxxx> wrote in
>> <news:1167511929.531019.279330@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
>> in sci.lang:
>>> Brian M. Scott wrote:
>>>> Yes, but Aidan is correct in saying that a single example
>>>> does not disprove a *correlation*. [...]
>>> "Bayesian inference" are the keywords:
>> Not really. One needn't know anything about Bayesian
>> inference to know about correlation.
> It's one small step to climb, as proved by top political
> science departments. Two googlable keywords beats a
> non-googlable paragraph.
But you don't need the whole paragraph: 'correlation' beats
'Bayesian inference' by 87,900,000 to 1,110,000. (And
'cherry picking' gets only 1,220,000.)
> By the way, there are more cherry picking people than
> bayesian inference people, as shown by GoogleFight:
> http://www.googlefight.com/index.php?lang=en_GB&word1=cherry+picking&word2=bayesian+inference
>
<g> Hadn't seen that site before.
[...]
Brian
|
|