On confirmation bias

Doug Natelson has done an outstanding job at debunking a ridiculous charge of confirmation bias allegedly affecting a recent study of climate change. Such a charge is put forth in an article published in the popular press (on a very prominent venue). While ostensibly aimed at educating the general public about some aspects of how science works, the article sneakily rehashes one of the most common and dangerous misconceptions that exist out there about science, namely that in the end it is not as “objective” as its practitioners claim.

The argument goes as follows: a scientist is a human being, has her own axe to grind, her judgment can be clouded by her own preconceived idea of how a natural phenomenon should be interpreted. Consequently, what is peddled to the public as impartial, rigorous experimental observation, is nothing but someone’s personal, biased view of how the world works. The supposedly objective “data” upon which the scientist’s conclusions are based, are merely the restricted subset thereof that support her beliefs; her interpretation is just one of the very many possible, and the decision of making hers the “canonical” one, as opposed to any other one, is essentially ideological, there being no accepted criterion to acribe more credence to her theory rather than to a competing one.
Wildly off base and pernicious as it is, this contention is unfortunately (at least in my experience) popular among non-scientists, including otherwise highly educated individuals. And in a way, this is scarcely surprising.

Undermining the respect that people have for science, spreading the false belief that experimental data are themselves “subjective” [0], that picking one of the many competing theories as the correct one ultimately entails an act of faith — this rhetoric serves the interests of easily identifiable, very influential groups.
It helps set the stage for a political discourse and societal debate in which facts are irrelevant, any attempt to evaluate an argument based on its merit is vacuous, in which careful, objective scrutiny is forgotten and free thinking is replaced by ideological conformation, i.e., one’s stand on any issue must reflect a choice of camp, or party affiliation. In such an environment, it becomes easier for those who have the means and the resources to “scream the loudest”, to advance their own agenda.

Make no mistake, science is not perfect. Yes, its practitioners are human, we all have been guilty of sloppy thinking at times, and yes, on occasion one succumbs to the seduction of possible fame, success and recognition, failing to exercise the required thoroughness and skepticism, especially when a “breakthrough” seems at hand.
Or, as in the case of research fraught with broad potential consequences on society as a whole, such as climate change, a researcher may be so wedded to, so deeply convinced of the validity of a specific hypothesis, to be effectively unable to examine the data objectively, consciously or unconsciously giving greater emphasis to those that appear to confirm the person’s preconceptions.
That all of the above can happen, is not even debatable. But, going from that to positing that scientific research suffers from a built-in credibility gap due to confirmation bias, requires a great deal of ignorance (or, alternatively, a deliberate intent to mislead).

What is “confirmation bias” ?
Let us consider a very simple example. Say I am a theoretical condensed matter physicist, and develop my own model of the interaction between two identical molecules — such as water, for instance. Based on this model, I perform microscopic calculations (e.g., on a computer, using Molecular Dynamics), and compute the equation of state of water — you give me the temperature and the pressure at which water is, and I shall predict its density for you, at thermodynamic equilibrium, in a given range of values of these parameters.
Obviously, in order for my work to be useful and/or taken seriously, my prediction has to agree with experiment, i.e., my computed equilibrium density must reasonably close to that experimentally measured, to within an accepted degree of precision. So, I look up experimental data, and for a given temperature I find what is illustrated in the figure below:

It is clear what the problem is, is it not ? My theory (red line) seems to be doing a pretty good job most of the time, in that it goes through (agrees with) most of the experimentally measured Density-Pressure data points (green circles). In fact, it goes through all of them, except one (the one with a question mark by it), which falls noticeably off the curve [1].

Up to this point, there is no bias whatsoever on my part. It would be downright silly to accuse me of “bias” because I try to verify if my own calculated equation of state reproduces the one measured experimentally by others, as opposed to that computed by someone else, or some other curve picked at random out of the many possible that could go through the green points. This is nothing but validation of a scientific hypothesis.
If my curve badly missed the mark, if it did not go through any or most of the green points, clearly there would be something wrong with the model (i.e., my proposed inter-molecular interaction), which would have to be downright rejected as inaccurate (unless I made a mistake with the computer calculation). That would be progress — nothing to be excited about, to be sure, but some understanding would have been generated nonetheless.
If, on the other hand, the red curve did go through all of the points, that would still not necessarily mean that mine is the “best possible equation of state” — other models may be even more precise. It would mean that my model affords a certain degree of precision in reproducing the experimental equation of state (in the range of pressure and density to which points shown pertain). And that, of course, would also be progress.

OK, so, what do I do, now ? I am staring at the above figure, thinking:
“Darn, look at that… I almost have all the points right except for that one… funny, both the one above and below it are fine but… that one is not… how can that be… mmmm… that point seems odd though… could it be that maybe that one measurement is wrong ? Maybe I should contact the experimenters and see if this datum was not recorded properly… it must be a mistakemaybe I should just take that one point out of the figure… after all, it is just one point… why generate doubts among people, undermining my own work, when surely the problem must be with that one experimental datum.”

OK, see the part of the text in green ? That is what confirmation bias is. I am so convinced that my theory must be correct, that I am simply discarding experimental evidence that is not consistent with it [2]. By publishing my results without including that “strange” datum, I am at least in part misleading the community, because I am trying to convey the impression that my model is more accurate than it might be. If that experimental datum which I have arbitrarily discarded should turn out to be correct, then clearly there is something in the actual equation of state of water that my model does not reproduce.

What is important to remember, however, is that science is fairly robust against anyone’s ego, sloppiness, delusion, or even fraudulence. Experimental data are passed to the microscope, theoretical scenarios carefully scrutinized and calculations thoroughly checked by competitors eager to prove us wrong. It may take a while at times, but eventually the truth will emerge. If I were to do what described above, i.e., omit that puzzling datum [3], it is only a matter of time before someone else points it out and makes me look silly. It is simply not in my best interest.
So, while it exists, confirmation bias in the science is, in my opinion, not an issue.
Contending that, because of confirmation bias, there are many possible answers to a scientific question, and picking one over the other is a subjective proposition, is nonsensical. Trying to suggest that any scientific theory largely reflects the bias of its authors, and that there is no objective, accepted way of selecting the one that best fit experimental data; that no real consensus can therefore be achieved on any scientific question; that the prevailing explanation of a given phenomenon is simply that which happens to enjoy the support of the greater fraction of the community; all of that is nothing but hogwash. Only a crackpot or a con man can make such a contention, and only someone blissfully ignorant of how science operates can take it seriously.
If that were true, the technological advances that we owe exclusively to our scientific progress would not have taken (and be taking ) place.

Notes

[0] To be sure, this has also been the theme of a whole school of thought, very critical and skeptical of the scientific method, of which Paul Feyerabend was one of the most prestigious and radical exponents.

[1] All of this can be stated more precisely, but it is not necessary here. Suffice to say that experimental data come with a stated level of precision, and that the size of the green dots, in a plot like that shown, should be equal to the experimental uncertainty (“error bar“).

[2] That is very different from actually moving the one green dot up, in order to make it fall on the curve. That would be fraud, not confirmation bias. Totally different ball game.

[3] As opposed to publishing everything, including all available data in the plot. I can simply write explicitly on the article that there seems to be a problem with that particular datum, propose that maybe that measurements should be repeated, and let the community decide what to make of this. Who knows, maybe the referee(s) will be able to explain what the problem is. Maybe there is a problem with that measurement, and this is the proper way to bring it to the attention of interested researchers.

Tags: , , , ,

14 Responses to “On confirmation bias”

  1. Justin Caouette Says:

    Very good post! Being somewhat skeptical of scientific data interpretations myself I do have a question though.

    In Neuroscience and in Cognitive Psychology the interpretations of the data are subjective. I can’t make the same argument for Physics as my background is one high school class and 2 college courses, however, there seems to be an agenda in lots of the soft sciences (psych/econ, etc.). They interpret the data in a way that suits their subjective political agenda (not all the time but this does happen). Case and point is free will and neuroscience. Some neuroscientists (Sam Harris) contend that they have proved with their neuroscientific research that free will is an illusion. This interpretation of legitimate science is, as you say, “hogwash”. The experiments took place in the mid 80’s and have just come to light as wrong (a few days ago). For me, I think it’s academically and socially irresponsible to take a scientific study as if it it telling us something deeper about our own experiences, especially when the interpretation of the data can be handled differently and when the data is trying to debunk common intuitions about the human experience. This one took 30 years to debunk, it doesn’t seem to be as robust in finding errors as you’ve claimed (though, admittedly many errors are caught rather quickly).

    As Thomas Kuhn as pointed out, much of science is puzzle-solving. What happens to these puzzles when the paradigm shifts again? Do those that look at science as pragmatically important but who take it with a grain of salt get to laugh then? I don’t mean to be condescending, I’m just trying to shed some light on how dogmatic it is to see science with a realist lens. You’ve stated that anyone who sees (1)”that the prevailing explanation of a given phenomenon is simply that which happens to enjoy the support of the greater fraction of the community” is a “crackpot” or is “blissfully ignorant with regards to how science operates” was a bit harsh. I might not see science as you’ve explained in (1) but, I do think there is some truth to the claim. Science is funded by money, those who buy into a paradigm will not put money into experiments outside of that paradigm. I’m not a scientists but one need not be to see that last statement as a truism. Over half of the science being done in Universities (U.S) is funded by corporations with an agenda. Ghost writing is becoming increasingly common place in scientific journals across the world. The voice of those opposing certain experiments and interpretations do not have a voice. The editorial boards of academic journals, the bottom line by corporations looking to make money off the latest “safe” drug, ans the schools that are conducting the research, all have the fear of getting cut hanging over their heads. This all seems to point away from objectivity and more toward the unethical hand greed is playing a in the endeavor we call science.

    With that said, I truly enjoyed your post and your perspective. Thanks for sharing.

    • Massimo Says:

      Thank you for your comment. I am going to try and reply to the best of my ability, but please let me make it again crystal clear that this is nothing but my own personal opinion.

      In Neuroscience and in Cognitive Psychology the interpretations of the data are subjective… They interpret the data in a way that suits their subjective political agenda

      Same in physics or any other experimental science — no difference whatsoever.
      As long as experimental data are not sufficiently precise, or scarce, or in any case insufficient to rule out some of the competing theories, such theories cannot be discarded. If a long time is needed in order for data to be robust enough, so be it. Case in point, the recent discovery of the Higgs boson. The theory has existed for decades, but the data to accept it as valid were not there until a few months ago — until then it was just one of the (many) theories consistent with available experimental data.
      Now, on the other hand, theories that do not predict the existence of the Higgs boson are on shaky ground.
      I think that the only difference with Neurosciences and Cognitive Psychology is that collecting data and statistics is much more difficult, one must make do with limited samples and therefore it becomes more difficult to debunk dubious theories. But, are there “fundamental differences” ? I don’t believe it.

      I don’t mean to be condescending

      No worries, if anything I might find you “incomprehensible”. I don’t mean it as an offence, I have never really understood the meaning of questions such as:
      What happens to these puzzles when the paradigm shifts again?
      much like I have always found myself at a loss trying to make something, anything of what individuals like Thomas Kuhn, Ernest Mach or Paul Feyerabend were talking about.
      The one thing that has always been obvious to me, is that these people, who made a living out of writing about science, had really little or no direct experience with actual scientific research.

      Science is funded by money, those who buy into a paradigm will not put money into experiments outside of that paradigm.

      Even if I accepted the premise (and I don’t, not unless we agree on what “paradigm” means, at least) — So what ?
      What does that prove, exactly ? Are the individuals funding the research that serves their needs, also capable of making it yield the results that they want ? That there may be a grey area in some cases, where some tentative conclusions are oversold, or where some incorrect beliefs survive longer than they should, sure — but to say that science as an activity, is intrinsically biased by those who fund is sheer nonsense, I am sorry.

      There are people and corporations that fund research on cold fusion because they are convinced that there may be something to it. Does that make cold fusion a scientific fact ? Does it mean that they can push onto the pages of Physical Review Letters phoney experimental work supporting the notion of cold fusion, even if no one else can reproduce their results ? Let’s be serious, please.
      The tobacco companies poured money into research aimed at disproving claims that cigarette smoke was harmful. Were they successful ? Sure, as long as clinical data were not there. As the evidence started piling up on the deleterious effects of nicotine, the discussion ended. I could give you countless other examples.

      That scientific research is largely funded by people with some vested interests is true. What does that say about science itself ? Nothing. There is no “republican”, “democratic”, “communist”, “post-modernist”, “impressionist” science — there is science, period.

  2. Confused Says:

    I am confused. After the first paragraph, is your post responding to points made in the three Matt Ridley articles? If so, I can’t see the correspondence.

  3. Camilla Says:

    After reading Matt Ridley…..( this will be mostly an add on and rephrase of what Massimo was saying)

    So Matt Ridley argues that contrary to popular belief, science is not impartial and critical. Instead, scientists are guilty of confirmation bias; that is, scientists are naturally inclined to find supporting evidence for their established beliefs rather than to challenge it. As a result, scientists have the tendency to favor confirming evidence.

    However, as Massimo has successfully shown that there’s really not much room in science for confirmation bias. Any form of bias in science will just be a deliberate lie. This is because the nature of physics, as a quantitive science, allows the scientist to perceive their data clearly and distinctly. This point is illustrated by Massimo’s graph example where the scientist can clearly distinguish the points conforming to the curve representing theoretical prediction and dots representing experimental data. Unless one’s consciousness is pretty much blind… one would be fully aware of his or her adulteration of the experimental results.

    Ridley’s argument is self contradictory and hypocritical. Ridley argues that since scientists are only human, and it is in human nature to privilege confirming results instead of anomalies to their beliefs. So following his own reasoning, Ridley should agree that since he is also human, we should just discharge his argument for confirmation bias as itself a confirmation bias. The reason that Ridley charges scientists with confirmation bias is to confirm with his own bias that it is in our nature to have conformational bias….. I think we got the point, right? In fact, any argument for subjectivism undermines its own accountability.

    Ridley then attempts to reconcile the progression of science with his argument. He asserts that the progression of science is not from self critique, but the critique of others for the sake of advancing one’s own view. Ridley asserts that, “so long as there are competing scientific centers, someone will prick the bubbles of theory reinforcement in which other scientists live.” Like Massimo, I believe Ridley is confusing science with politics. In the case of science, nature serves as the ultimate check. All scientific theory must confirm to experimental observation, not one’s own bias.

    Just to be fair, let’s take a look at some of Ridley’s other arguments. He argues that the if-then modeling technique is often employed by scientists, and this is an example of confirmation bias. What Ridley means here is that suppose A is the theory/model to be tested. One test A by checking if it conforms to experimental observation B. So if A is true, then B must occur. If B does not occur, then A has made a false prediction. Ridley’s point is that even if B does occur, so showing that A predicts B, A is not the only theory that could predict the occurrence of B. So Ridley argues that the method is confirmation bias because it presuppose A as the only explanation for B until proven otherwise. However, as Douglas Natelson has shown, scientists will test many competing models to see which one best predicts the experimental result.

    Lastly, Ridley accuses science of detaching itself from reality by excessively relaying one models. Here Ridley is trying to undermine nature as the check for scientific theory.To respond to the last attack, let’s ask ourselves a philosophical question. Is there really a distinction between reality and models? Of course, we would all agree that nature is independently acting; that is, what it does is not controlled by the scientist. Although nature is itself independent from knowledge, as the that which all knowledge must conform to, knowledge/reality is dependent on us. This is not an assertion that reality is simply an illusion, rather I’m advancing the view that reality is itself a model. Reality(the way things are) is a function of your system of understanding. The validity of this system of understanding, scientific knowledge, is check by its conformity to the independently acting entity, nature.

  4. Me Says:

    Ah ah ah. An excerpt from Matt Ridley’s post:

    “I argued last week that the way to combat confirmation bias [..] is to avoid monopoly. ”

    Somebody has told him that nature has a monopoly on law of nature ?

  5. Douglas Natelson Says:

    Nice post, Massimo – you did a much more thorough job than I did. To be clear, Confused, Ridley seems to have a very weird (one might say biased) concept of confirmation bias. He seems to think that the very act of comparing data to a set of models and declaring that one is the best fit is inherently biased, because the person doing the comparing selected the models. That’s just incorrect. It is possible to make quantitative comparisons with models in an objective way. It is possible to ascribe “goodness of fit” metrics, and to compare these quantitatively, in an objective way.

  6. confused Says:

    It has been about a week and I guess I’ll bring up the points I disagree with.

    @Doug

    I think Ridley is actually saying this particular case of curve fitting (note his use of the word ‘such’) is an example of confirmation bias because of bad data and ‘unjustified adjustments’. Ridley even concedes that Muller may be right about CO2 being the best model to his data. Which suggests Ridley doesn’t doubt Muller’s metrics. In essence, Ridley doesn’t like the data processing that goes with the curve fitting in this case.

    • Massimo Says:

      bad data

      Uh ? Then why not call it “bad data”, instead of “confirmation bias” ? Why not point the readers to “better” data ? And why are data “bad” ? Are the uncertainties incorrect ? And what are the “unjustified adjustments” ?
      Come on….

      • confused Says:

        I didn’t say it was a good argument. I’m only saying that this more accurately portrays this aspect of his argument.

      • confused Says:

        To answer some of your questions: He pointed readers to Anthony Watts’ website, which contains the preprint showing Watts’ conclusions.

      • Massimo Says:

        To answer again your comments: the case for confirmation bias is absolutely untenable, no matter how much you are willing to stretch the notion of “confirmation” or “bias”.
        If you are ideologically aligned with Ridley and therefore feel that you should mount a defence of some kind of his argument (which is ultimately against climate change, let’s not kid ourselves), the best way is to provide better data, or a more convincing fit to the existing data using a different model. Trying to discredit science by building a non-existent case of confirmation bias is not the way to go, in my opinion.

  7. confused Says:

    I forgot to mention, in my last reply, that my problems are with the way people are interpreting Ridley’s arguments.

    @Camilla

    I see no problem with the first paragraph after your leading sentence, expect that it may be incomplete. For example, it is missing his argument why science progresses: competition.

    Your second paragraph, by the way, inappropriately equates the words “science” and “physics”.

    In general, data must be massaged and fitted to models, it is a rare case to have our data clear and distinct.

    “Unless one’s consciousness is pretty much blind…” In place of the ellipses I suggest you add ‘by confirmation bias’.

    In the third paragraph, I don’t think you have fairly presented his argument. He never asks the audience to completely discharge the arguments of scientists. Instead, he asks for skepticism.

    Regarding your fourth paragraph, science is very much a human endeavour and does have politics involved in it.

    Evidence must be judged by our faulty brains. For example, Dan Shechtman discovered quasicrystals in 1982 with some pretty solid evidence from electron microscopy. This was quickly followed up by more evidence by many others after their 1984 publication. Linus Pauling, a two time Nobel Laureate and arguably the Master of x-ray crystallography, never believed in quasicrystals because he erroneouly thought they went against a fundamental law of crystallography and they were obviously twinned crystals. Why did he have these misconceptions? Because right up to his death, in 1992, he never bothered to try to understand electron microscopy nor the models of quasicrystals.

    In your fifth paragraph, you talk about Ridley’s if-then example. I think that his example could be interpreted as if-then modelling that happens to be skewed by confirmation bias. However, he is not very clear on this point.

    The last paragraph is quite confusing, but you did make a clear statement of what you think Ridley’s says. Unfortunately, you provide no evidence of this.

    • Massimo Says:

      In general, data must be massaged and fitted to models, it is a rare case to have our data clear and distinct.

      Excuse me ?
      I am sorry, I try to let everyone write whatever they like and for the most part I am good at ignoring nonsense, but if I don’t say something here people may think that I agree with it.
      I am not sure where you got this from, much less what your research experience is, but “massaging the data” (in physics or any other science) is a colloquial expression that is usually utilized in reference to someone’s less than transparent methodology and/or otherwise dodgy scientific practices. Me, I was taught in graduate school and have always believed (and still do) that it is the model that has to fit the data, not the other way around. If someone tells me “I had to massage the data to fit that model”, unless they burst into laughter and add “just kidding !” shortly thereafter, I have the tendency not to believe a word that they say.

    • Camilla Says:

      1)
      “For example, it is missing his argument why science progresses: competition.”

      I have quoted Ridley, so I’m not sure what you mean that I excluded his argument. The point Ridley was making is that if confirmation bias exist, how can scientific progression be possible. His explaination is through the prevention of monopoly. Monopoly prevents fair competition. Here the problem is about power. If one view dominates, it will prevent other views from being valid by virtue of its authority. I disagree with this point Ridley made because he is confusing science with politics. Science is not about who has the authority, but whether the theory conform to experimental findings. Nature(experimental findings) does not play much of a role in the world of economic and politics, so power and competition plays a much greater role. We can encourage competiton as much as we want, but this will not make nature want to conform to your theory. It is not clear how the encouragement of competition will lead to scientific progression. In fact, I believe it’s the collaboration amongst scientists, who work together as a team, that make science exceedingly successful at solving problems in comparison to other deciplines.

      2)
      “Your second paragraph, by the way, inappropriately equates the words “science” and “physics”.”

      Ok, I did reduce the word science to physics. You caught me. Although you’re commenting on a physicist’s blog, so we prefer to keep the discussion within our scope of experience. However, even for sciences that are more qualitative, is there not clarity and distinctness? Will our faculty of reason fail us in those sciences?

      Clarity and distinctness are qualities attributed to reason. I have borrowed this concept from the philosopher, Rene Descartes. Descartes argues that what I perceive clearly and distinctly is true.

      Suppose that you have just proved the Pythagorean theorem. Does the truth of this theorem depend on your personal bias? You can clearly and distinctly see how the pythagorean theorem is true by following your proof. Likewise, when a biologist conducts ther experiment, many of her conclusions are drawn based on logic and reasoning. She makes her conclusions only when she clearly and distinctly perceive that they’re true, in the face of many experimental findings.

      So how exactly does comfirmation bias contort reason? If I don’t believe in the Pythagorean theorem will my bias allow me to prove it wrong? How exactly does truth depend on me?

      So why do we still make errors? Oh, I don’t know God doesn’t exist, and we’re not perfect? On a more serious note, just because it’s an error it doesn’t mean that it’s a confirmation bias. Science is a very complex and tricky subject, and our faculty of reason is not omnipotent. However, we’re able to correct ourselves from errors. The very fact that we’re able to correct ourselves shows that our reasoning isn’t inherently biased and flawed. This fact is another reason why science is successful because scientists are able to correct themselves and to correct each other.

      3)
      “I think that his example could be interpreted as if-then modelling that happens to be skewed by confirmation bias. However, he is not very clear on this point.”

      If he is not clear on this point, on what grounds do you believe in his arguments?

      4) ” The last paragraph is quite confusing, but you did make a clear statement of what you think Ridley’s says. Unfortunately, you provide no evidence of this.”

      The idea of reality is a very complex issue. My goal was only to entertain Ridley’s arguments, rather than solving the problems once for all.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s


Follow

Get every new post delivered to your Inbox.

Join 28 other followers

%d bloggers like this: