Google’s Global Warming Search Bias – Part 2

Posted on Sat 02/17/2018 by


By David Wojick, Ph.D. ~

In my previous article I found that Google search results for prominent skeptics of climate change alarmism included a lot of negative material. Four or five negative items out of ten or twelve on the first page of results were typical. This raises the question of whether Google’s search algorithm is tuned to find and present negative items on skeptics.

Another possibility is that these search results merely reflect the fact that there is a lot of negative material on both sides. The climate change debate is indeed acrimonious and feelings run high.

To test this alternative I did Google searches on a number of prominent alarmists. What I found was that there were very few negative items, in fact in most cases there were none at all. This clearly suggests that Google’s search algorithm likes climate alarmists and dislikes climate skeptics, which has been rumored for some time.

Here are the first page results that I got for Google searches for prominent climate alarmists. A lot of negative material has been written about every one of them.

James Hansen
12 items, zero negative

Michael Mann
10 items, one negative

Ben Santer
10 items, zero negative

Susan Solomon
10 items, zero negative

Gavin Schmidt
10 items, 2 negative

Bill McKibben
12 items, zero negative

Naomi Oreskes
10 items, one negative

In contrast, our search results for skeptics typically contained 4 or 5 negative items on the first page. Thus there is a dramatic difference in how the Google search algorithm treats climate alarmists (positively) versus skeptics of alarmism (negatively).

There are content differences as well. Google search results often include laudatory articles about the leading alarmists, as well as some of their major writings. This almost never happens with the skeptics that we analyzed.

I have myself developed or worked on several search algorithms for the US Energy Department. Tuning an algorithm to target prominent skeptics of alarmism with negative material is not that hard.

For example, for every skeptic we analyzed, the search results include a link to DeSmogBlog. This site’s leading feature is a large database of negative dossiers on prominent skeptics. It would be easy to tell the algorithm to treat everyone listed in the DeSmogBlog database negatively.

Then too, in the natural language processing community there is a lot of research on how to recognize negative content. This is used to analyze everything from message traffic to news items, especially in the intelligence and marketing communities. In the science community there is a whole new field called “altmetrics” that seeks to use things like the volume of tweets and news stories to evaluate the impact of breaking new scientific findings. Here being able to tell the difference between positive and negative items is crucial.

The point is that the technology clearly exists to tune an algorithm into being either positive or negative in compiling and presenting search results. If I can design a search algorithm that negatively targets skeptics of climate change alarmism, then Google can certainly build one. After all, their corporate research and development budget is several billion dollars a year. A good chunk of that probably goes to fine tuning their search algorithm, since it is their claim to fame and fortune.

If Google is in fact using its search algorithm to target skeptics of climate alarmism, by systematically presenting a lot of negative results, then it is time to speak out. This activity may well be illegal, because it is potentially damaging to those being targeted.

David Wojick contributes Posts at the CFACT site. He is a journalist and policy analyst, and he holds a doctorate in epistemology, specializing in the field of Mathematical Logic and Conceptual Analysis.

Read more excellent articles at CFACT