The Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.


Math Forum » Discussions » sci.math.* » sci.stat.math

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: "z" correlation coefficient (?)
Replies: 6   Last Post: Sep 24, 2012 7:05 AM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]
Ray Koopman

Posts: 3,383
Registered: 12/7/04
Re: "z" correlation coefficient (?)
Posted: Sep 23, 2012 6:04 PM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

On Sep 23, 4:45 am, Art Kendall <A...@DrKendall.org> wrote:
> or based on my early training as an Aristotelian philosopher,
> I suggest slipping between the horns of the dilemma of p vs CI,
> and point vs interval estimates.
>
> "In the group of nonsurvivors (n = xxx), a statistically significant
> correlation was found between ICP and GCS (tau = xxx, p <= xxx,
> 95%CI = (xxx, yyy), and CPP and GCS (tau = xxx, p = xxx, 95%CI =
> (xxx, yyy), while no statistically significant correlation was observed
> between MAP and GCS (tau = xxx, p <= xxx, 95%CI = (xxx, yyy))."


If you must. But please, "p = xxx", not "p <= xxx". xxx is the
probability of getting a sample tau as extreme as or more extreme
than the observed tau, given that the true tau is 0. It is not an
upper bound for that probability.

>
> If I were the quality assurance reviewer, I would also ask the writer
> if (s)he had double checked the data entry overall and triple checked
> the one point that is visually unusual.
>
> Art Kendall
> Social Research Consultants
>
> On 9/23/2012 4:17 AM, Ray Koopman wrote:

>> On Sep 22, 12:28 pm, Vlad <vldsc...@gmail.com> wrote:
>>> Please help to interpret the correlation coefficient specified
>>> in this article http://hrcak.srce.hr/file/22665.
>>> They suggest they used Kendall Tau nonparametric correlation,
>>> but I can't figure out why it's noted as "z", not "É?" ...
>>>
>>> "In the group of nonsurvivors, a significant
>>> correlation was found between ICP and GCS (z=-2.351),
>>> and CPP and GCS (z=3.231), while no correlation was
>>> observed between MAP and GCS (z=1.287). "
>>>
>>> Best Regards,
>>> Vlad.

>>
>> That's a good example of bad reporting.
>> They should have said something like
>>
>> "In the group of nonsurvivors (n = xxx), a significant
>> correlation was found between ICP and GCS (tau = xxx, p = xxx),
>> and CPP and GCS (tau = xxx, p = xxx), while no significant
>> correlation was observed between MAP and GCS (tau = xxx, p = xxx)."
>>
>> or (better)
>>
>> "In the group of nonsurvivors (n = xxx), a significant
>> tau was found between ICP and GCS (95%CI = (xxx, yyy)),
>> and CPP and GCS (95%CI = (xxx, yyy)), while no significant
>> tau was observed between MAP and GCS (95%CI = (xxx, yyy))."




Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© The Math Forum at NCTM 1994-2018. All Rights Reserved.