OK my title is a bit of media sensationalism but so what, you are now reading this thread as a result aren’t you? I’ll admit though that it should really say ‘Is Google biased towards the man-made global warming/climate change hypothesis?’
My blogging on this topic has been prompted by Willis E’s recent ‘Open Letter to Google‘ thread on WUWT. Willis was in turn prompted to blog on this topic because of Google’s recent announcement/decision to fund 21 (so called) climate research ‘fellows’ (http://solveclimatenews.com/news/20110318/google-climate-change-fellows-science-new-media) who will ‘improve the way the science of global warming is communicated using new media’.
Has Willis E ‘jumped the shark’ on this one? Probably not but he certainty has over-reacted IMO. This announcement/decision hasn’t been made directly by what I’ll call ‘corporate Big G’, but rather by its ‘philanthropic arm’ known as Google.org (not Google.com). Is it possible for corporate Big G to completely divorce itself from its ‘philanthropic arm’ Google.org? I doubt it, but that’s a completely different question.
The question I’m about to ask and attempt to answer is ‘Is Google (corporate Big G) biased (prejudiced) towards the man-made global warming/climate change hypothesis?’ Well how does one go about answering that question? In my case, I’m going to use a simple search for the phrase ‘climate change’ on both Google UK Search and the UK version of Bing.
First of all then, over to Google.co.uk search and type in ‘climate change’ in the search box and click Search or press enter.
Note that at any given time when you click on the above link, the ‘Top 10’ non-sponsored links appearing on Page 1 won’t necessarily be the same each time. The results returned (and their relative order) will most likely vary from day to day and possibly even hour to hour.
So at this moment in time (12:40 pm on 213/2011) who do Google think are the moment relevant authorities on Climate Change? Here are the ‘top 10’ ranked URLs in descending order of their ‘relevance’ as far as Google are concerned.
With the exception of Wikipedia, notice anything about all those web sites? Are they all fully funded or in part funded by the UK Government? Answer – Yes!. Is it much the same for the URLs that appear on page 2,3, 4 etc? Yes! What proportion of them are ‘institutions’/NGOs etc? Answer – nearly all of them?
Well it pretty clearly shows that as far as Google is concerned that the opinions (on climate change) of the UK Government and the institutions/organisations it chooses to fund are far more relevant as far as Google UK is concerned than the opinions of other non-government funded sources it chooses not to fund. Does this mean that Google UK is biased and can justifiably be accused of only presenting one side of the climate change debate? Well that depends.
One could argue that the data needs to be ‘adjusted’ and (a)no(r)malised to allow for the fact that there are probably far more web sites on the internet that present the case ‘for’ rather than ‘against’ man-caused global warming/climate change. Obviously we also need to adjust for the no. of ‘institutional’ versus ‘non-institutional’ web sites that present information on and discuss climate change on the internet. This is of course just the UK and different levels of adjustment will most likely be needed for other countries ‘raw’ Google Search data. Where there is insufficient data available for a given country will will also need to ‘in-fill’ the missing data with appropriately extrapolated data (particularly in far northern latitudes) as I’m sure searches carried out with Google within a 1200 km radius are well correlated and its not our fault that not that many people do Google searches above the Arctic circle.
If the level of adjustment exceeds the the amount of bias (we already know is there), then we’ll need to be even more carefully. We may need to use novel statistical techniques to tease out the ‘bias signal’ in the data if it is not immediately apparent post our required adjustments. I’ve heard that using de-centred PCA, padding the end points and smoothing the data when looking for the bias trend always comes in handy. I sent an email to RealClimate for some advice on this point.
If we are looking for a more longer term trend in bias, we might even decide to forget the actually data altogether and opt for a ‘proxy’ instead. It doesn’t really matter if the proxy data needs to be adjusted as well, as it’s the bias trend that we are interested in isn’t it and (a)no(r)malising the data will make full allowance for that I’m 99% certain.
OK time for a bit of ‘Binging’. What are the ‘Top 10’ search results for for ‘climate change’ on Bing UK (http://www.bing.com/?cc=uk) at 13:54 on 21/3/2011?
Here’s Bing’s Top 10
Now hang on that list is remarkably similar to the earlier Google ‘Top 10’ list. Greenpeace and FOE have sneaked in at the expense of DEFRA and DECC so maybe the UK government spends more on advertising with Google than it does with Microsoft? Or maybe Microsoft ‘Does evil’ as well? It’s a similar situation on pages 2,3,4 5 etc of Bing as well. Lots and lots of UK government funded institutions/NGOs. Despite once more adjusting and (a)no(r)maling this ‘raw’ data and applying the same novel statistical technique to it as I did for the Google data I’m struggling to find any statistically significant difference between the ‘climate change bias’ I’ve found (OK was already looking for because I knew it was there!) in the Google data and the Bing data.
I’m tempted to try Yahoo and AltaVista as well to see if they also ‘Do evil’ but I’ve ran out of time.
As a result of Willis E’s publishing his ‘Open letter to Google’ on WUWT have I changed my default search engine from Google to Bing? No! That’s the wrong question again. Have I changed my default search engine so that it is no longer Google? Yes! Why? Because, like I suspect a lot of other folks, I’ve got fed up with being re-directed to some irrelevant web site (I’m not interested in horny housewives Google – honest ;-)!) at the whim of Google whenever I first click on a link in their returned search engine results. Oh hang on! Is Google really the ‘keeper of all our secrets’ as Willis E seems to think? For the sake of the continued longevity of my happy marriage, God I hope not! My wife might be able to find out just how much time I spend on the internet ‘researching’ (climate change of course!) when I’m suppose to be ‘keeping her in the life she has become accustomed to’.