Re: Pearson-r in SQL

"Alan" <not.me_at_rcn.com> wrote:

>>
>> It is not valuable information at all.
>>
>
>Not necessarily. It may be that knowing that something does not correlate is
>important. It is certainly information. Is it important information? I don't

     But we do not know that it correlates or not.

>know. Sometimes it is and sometimes it isn't. To some degree, it depends on
>what "null" means in the given context. That is why I suggested reporting on

     Which is why using nulls in a database is Bad Stuff.

>the uncorrelated points separately (only if this information is important).
>The point is, a "no correlation" condition may or may not be important. We
>just don't know.

     That is the point: we do not know. Not only that, but we can not know.

>This still does not answer Joe's question, which was, in essence, "How is
>this problem normally handled in reference to the particular function
>Pearson's -R?"

     Thee is no data to correlate. There would have to be at least one data pair.

Sincerely,

Gene Wirchenko

Computerese Irregular Verb Conjugation:

     I have preferences.
     You have biases.
     He/She has prejudices.