Understanding Scientific Bias (Part 2)


In Part 1 I discussed several cases of intentional and deliberate scientific fraud.  In those cases the researchers involved fabricated data to give specific results.  That is the top of the pyramid in issues of scientific fraud.  They get the most attention, but they are also the rarest problems involved in scientific research, at least I hope that is the case.  The second part of the problem is a more subtle problem.  It involves the selection of data, avoiding contrary data, or just misinterpreting the data.  These examples are not really examples of clear scientific fraud as much as they are of scientific bias.

Unfortunately there are many examples of these problems.   The simplest one is the choice of data to be used.  This is easy to find in the topic of global warming.  It is not uniformly followed, so I will show lots of examples.  One of the simplest ways to tell if someone is a warmist or a skeptic is the choice of what global temperature set they choose to use.  Skeptics tend to use the UAH or RSS satellite data.  A simple Google search of “UAH Temperature data” brings results that are almost all skeptic websites or articles.  RSS tends to be more neutral in it’s use.  GISS and CRU are almost all websites supporting global warming or articles bashing those sets of data.  Certainly there are exceptions to this rule and many of them, but in general this is how it works.

For example I looked at the Top 10 arguments used on the Skeptical Science website and the typical result looks like this.  In a discussion that warming stopped in 1998 it uses the GISS, GHCN (NCDC), and CRU temperature sets.  It never once mentions satellite data.  That is typical for SkS.  In almost all cases that site focuses on one of those three data sets for temperature.  The main exception is when they are intentionally trying to point out how skeptics mislead by using the satellite data such as this.

The Inconvenient Skeptic

Skeptical Science Data Choice favors warming.

Of course skeptics are equally guilty of such choices.  Since Dr. Roy Spencer is associated with the UAH project it goes without saying that is what he uses.  But it doesn’t stop there.  The Hockey Schtick has an RSS data set displayed at the top of its website.  Most times that the GISS, CRU or GHCN data is used on Watts Up With That, it is when they are being discredited.  I may get savaged to saying this, but bias exists on both sides of the argument.  It is almost impossible to avoid this.  I know that I prefer the UAH data because it provides the most comprehensive coverage of the Earth, but I also understand that it is a form of bias that I use myself.  That is why I make myself use multiple sets of data for global analysis.

The Inconvenient Skeptic

Four of the main temperature sets I use to create the blended set.

One of the more famous examples of avoiding contrary data was made famous by Climategate.  One of the more famous quotes from that entire debacle was “Hide the Decline.”  While that has been beat to death, the most comprehensive discussion of that was made on Climate Audit by Steve McIntyre.  In this discussion it is clearly pointed out that the data past 1960 was truncated due to a ‘divergence problem.’  Steve clearly shows from the sequence of the data and the Climategate emails that the chart generated by Briffa was truncated and adjusted to fit with the end result of providing a tidy story of Anthropogenic Global Warming.  I try to avoid the Climategate fiasco (mostly because others involved have covered it so well), but in this case the example is too clear to avoid.

Climate Audit

Climategate and IPCC versions of the data

Finally there is the issue of poor interpretation of the data.  Since as far back as 2006 there have been stories widely distributed in the media that global warming is harming the penguin population in Antarctica.  Since 2006 when Camille Parmesan of the University of Texas authored a report indicating that the penguins were already dying as a result of global warming.  This has been reported in many different places.  The key fact was that this report used a the method of placing a band on the penguin.  A recent report by Claire Saraux studied the impact of the bands on the penguins and found that the act of placing a band on a penguin dropped the survival rate by 20% and reduced the number of chicks by 41%.  In this case the researchers were causing the drop in penguin survival rate and attributing the cause to global warming.

Save the penguins

More disturbing is that even though there is now strong evidence that banding penguins that banding them is harmful, many professors refuse to stop using them because they “cannot believe” that this practice hurts penguins.  More on this story later, but you can be sure I will be following up on this.

I have indirectly tackled bias and interpretation of results before when I discussed the results of the Arctic Report Card.  That entire discussion is a perfect example of ignoring the data and showing only the portion that conforms to the theory much like Briffa did with the tree ring data used by the IPCC.  Keeping science un-biased can be very difficult.  The more politically the results are used, the harder it is to keep the science un-biased.  That is why the issue of global warming is so conflicted.

Posted in Bad Science by inconvenientskeptic on January 23rd, 2011 at 1:05 am.

4 comments

This post has 4 comments

  1. intrepid_wanders Jan 23rd 2011

    Certain fields of science seem to have too much funding to be wrong. Another sad period in science is the anthropogenic depletion of the polar ozone layers. As the UNEP celebrated their success, http://ozone.unep.org/Publications/MP_A_Success_in_the_making-E.pdf, others were noting the serious flaws in the theory.

    Appinsys (http://www.appinsys.com/GlobalWarming/Ozone.htm) has come up with a very good layout of information concerning this. A 2007 study reported in Nature [http://www.nature.com/news/2007/070924/full/449382a.html] “As the world marks 20 years since the introduction of the Montreal Protocol to protect the ozone layer, Nature has learned of experimental data that threaten to shatter established theories of ozone chemistry. If the data are right, scientists will have to rethink their understanding of how ozone holes are formed and how that relates to climate change. … The rapid photolysis of Cl2O2 is a key reaction in the chemical model of ozone destruction developed 20 years ago. If the rate is substantially lower than previously thought, then it would not be possible to create enough aggressive chlorine radicals to explain the observed ozone losses at high latitudes, says Rex. The extent of the discrepancy became apparent only when he incorporated the new photolysis rate into a chemical model of ozone depletion. The result was a shock: at least 60% of ozone destruction at the poles seems to be due to an unknown mechanism, Rex told a meeting of stratosphere researchers in Bremen, Germany, last week.”.

    While I am a fan of “Keep It Simple Stupid”, I am noting a rash of researchers that take a simple physical effects and apply them to a chaotic system and refuse to test the theory (Because their scientist?). Then, when new data comes in, ignore it. Strange brave new world, I guess.

  2. This is just insane (banding penguins). How can so called scientists continue this practice against helpless animals while ignoring obvious empirical evidence that the practice of banding harms these beautiful animals? And to what end?

  3. GregO, THANK you. I thought I had lost my grasp on logic when I read about this. Let’s claim concern for the penguins while increasing their odds of biting the dust. Fantastic. When did our world turn into a Dali painting?

Web Design & Dev by

Mazal Simantov Digital Creativity