the Air Vent

Because the world needs another opinion

P Value Correlation, Who wants a backward hockeystick anyway?

Posted by Jeff Id on September 5, 2008

I found the following paragraph from the latest Mann paper interesting.

Reconstructions were performed based on both the ‘‘full’’ proxy
data network and on a ‘‘screened’’ network (Table S1) consisting of
only those proxies that pass a screening process for a local surfacetemperature
signal. The screening process requires a statistically
significant (P  0.10) correlation with local instrumental surfacetemperature
data during the calibration interval. Where the sign of
the correlation could a priori be specified (positive for tree-ring
data, ice-core oxygen isotopes, lake sediments, and historical documents,
and negative for coral oxygen-isotope records), a one-sided
significance criterion was used. Otherwise, a two-sided significance
criterion was used. Further details of the screening procedure are
provided in SI Text.

It’s a bit more scientific than many of the posts here but what he says is a one sided p value calculation is used. Here is a quote from a statistics group.

The only situation in which you should use a one sided P value is when a large change in an unexpected direction would have absolutely no relevance to your study.

This is quite relevent to the situation in the latest Mann (hockey stick) paper. Why would you increase the rejection of data based on a reversed trend?

This won’t appeal to many of my readers, but without a single sided P value correlation, the end of the hockey stick graph would be further muted. If every individual decision by the researcher is designed to increase the temperature peak at the end of the graph what are we to think?

I flatly reject the possibility that a reversal in the data should invalidate the entire trend. It doesn’t make any sense, and is certainly not proof of a superior accuracy data series when it passes this test.

I have skimmed a couple of other papers and found that this idiotic (yup I said it) rational happens fairly often.

Deeper I go, into the abyss.
A note for those who are less science oriented, this type of criteria is used to reject many many noisy datasets in paleoclimatology (past climate reconstruction). If you have thousands of temperature graphs which are very noisy (i.e. looks like a two year old drew a straight line) and you sort them for only those with a positive slope at the end and take an average – you make a hockey stick.

I have several expert readers who come by my site from time to time. Would someone please help me understand if this is actually the accepted practice through most paleoclimatology papers these days, or is it just some of them? Maybe I got (un) lucky in my quick search earlier today.

Time for some sleep I think.

2 Responses to “P Value Correlation, Who wants a backward hockeystick anyway?”

  1. sandra407 said

    Hi! I was surfing and found your blog post… nice! I love your blog. 🙂 Cheers! Sandra. R.

  2. megan fox said

    Sign: umsun Hello!!! rcuwwymhyw and 4103ssgfhphzye and 7646I love your site. 🙂 Love design!!! I just came across your blog and wanted to say that Ive really enjoyed browsing your blog posts.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: