cc: Gavin Schmidt date: Thu Jan 3 15:15:19 2008 from: Phil Jones subject: A thought and GDP and Education? to: gschmidt@giss.nasa.gov, Rasmus Benestad Gavin, Hope you do something. The other Dutch paper is much the same. MM try to use it as support for their work, but it appears to work for the same invalid reasons. The significance testing assuming that all the points are independent just seems wrong to me. The 440 points can't be independent of each other. One way of guessing what the right number is to put the data through a PCA and then see how many PCs are significant - using a conservative rule. I'd not taken in the significance of their e term (for education) before. As g (GDP) and e must be fairly strongly correlated positively (more GDP more literacy/education), then why are their coefficients the opposite sign? I'm assuming here higher e means higher literacy/better education? They divide GDP by area, but even then it should still be correlated with e. So why should the e relationship be -ve and the g relationship +ve with the temperature residuals. What if you try running the regressions with g or e but not both? It is easier to cope with multicollinearity by using PC regression. Do the regression after doing a PCA on the RHS variables. This is done in most paleo reconstruction work. If you've run with RSS, then useful to show these results alongside UAH. By the way, hope to send you the Wengen paper by the end of the month ! Cheers Phil At 14:11 03/01/2008, Gavin Schmidt wrote: Hi Phil, I've been looking at this a little. The correlations to GDP and education in M&M are robust to changes to RSS or updates of the CRU or GISTEMP data, but the other economic 'significant' variables are not. More interestingly, the net effect of their correlations is to focus on very restricted spatial areas and as you know, the smaller the area the less powerful any attribution can be. Thus in climate model results you see a wide range of correlations - some easily as signficant as those shown in M&M. The conclusions are unsupportable, but I'm still thinking about the best way to demonstrate this. I anticipate this will be a new paper that takes in the de Laat and Maurellis (2006) paper as well (also a statistical fluke). Gavin *--------------------------------------------------------------------* | Gavin Schmidt NASA/Goddard Institute for Space Studies | | 2880 Broadway | | Tel: (212) 678 5627 New York, NY 10025 | | | | gschmidt@giss.nasa.gov [1]http://www.giss.nasa.gov/~gavin | *--------------------------------------------------------------------* On Thu, 3 Jan 2008, Rasmus Benestad wrote: Dear Phil, I wish you a happy new year, although the death of Bert Bolin was a really sad start. Funny you mention this, as we have been thinking about this, and I have recived the same request by Ross McCitrick himself (!) - see below. Gavin has a very clever take on this (repeating the analysis on GCM results - not affected by the economic factors - and getting similar 'answers'...), and we are thinking about drafting a comment about it. The Douglas et al paper is already commented in another post on RealClimate: [2]http://www.realclimate.org/index.php/archives/2007/12/tropical-troposphere-trends/lan gswitch_lang/sw Perhaps we were a bit naughty posting it before the release of the paper? All the best! Rasmus Phil Jones wrote: Rasmus, Have had an email from Susan Solomon about this paper. She wonders if I'd considered doing a response. I replied saying I'd see if you were doing anything - more than what you've put onto Real Climate? I reckon if the analysis was redone with the RSS satellite data, any significance for the supposed economic growth indices would go away. And then there are the myriad of things you mentioned earlier. There is a group working on another awful paper by Douglass et al. which is in press in IJC. Sad day - hearing of the death of Bert Bolin. All the best for the New Year Phil Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ---------------------------------------------------------------------------- Dear Rasmus I hope you have had a Merry Christmas. Here in Ontario it was a very white Christmas, the way I prefer it, and we have been enjoying the snow and the skating rink outside. It seems to me that your main technical concerns about my paper with Pat was the possible spatial autocorrelation. I have posted a paper about this issue at my web site. [3]http://www.uoguelph.ca/~rmckitri/research/jgr07/jgr07.html (scroll down a bit) I submitted it to JGR, however, they said that it is not sufficient as a stand-alone paper. The editor said it is a response to critical comments, but since the comments have not been submitted they cannot proceed with it. Occasionally, an author will submit a Comment on a paper published, challenging the use of a certain technique by the authors of the original paper. We then facilitate correspondence between the scientists involved, publishing that correspondence in the form of a Comment and Response. Your manuscript has the flavour of the 'Response' but there are no scientists that have prepared a 'Comment' to challenge your original paper. Therefore, I don't see how I can publish the current manuscript. So, I was wondering if you would be willing to write your Real Climate posting up as a comment and submit it to JGR, and then we might be able to get an exchange published, which would probably be of interest to our readers. In the meantime, Happy New Year! Cheers, Ross Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ----------------------------------------------------------------------------