Monday, March 28, 2011

Confirmation Bias & the Availability Error

[caption id="" align="alignright" width="300" caption="Image via Wikipedia"]A set-theoretic illustration of hypothesis tes...[/caption]

There are many ways by which we can deceive and delude ourselves, and two of the better known mechanisms for this are the well-established phenomena known as Confirmation Bias and, its partner in confusion, the Availability Error.

Confirmation Bias... a universally human tendency to seek out and pay more heed to information which confirms our views and opinions, and is one of the major reasons why much scientific investigation involves attempts to falsify its hypotheses, at least in principle, when doing so would be more decisive than verifying them in the testing of an idea.

Verifying an idea may occasionally be useful, like finding even one actual example of an alien civilization in the search for extraterrestrial intelligence, but in many cases verification alone is so ambiguous, so trivial, that it tells us nothing new and interesting.

Confirmation bias is often evident in our daily lives, and is often seen in our tendency to read, watch, or listen to media outlets which support our ideologies and associate mostly with those who agree with us.

It can be offset by deliberately seeking out information which disagrees with us, and this can be entertaining as well. Personally, I find going on fringe-proponents' websites interesting, amusing, and educational, and I highly recommend it, because it not only allows exposure to information contrary with my views, but also allows fascinating insights into the thinking and worldviews of proponents, and anything that helps me understand 'the opposition' better rather than remaining ignorant of what they are really saying and arguing is a good thing.

It's far too easy to find data to validate any claim one wants to make if one were to ignore or avoid contrary data and cherry-pick from the unruly mass of information that which seems to support it, especially when the claim is framed in a way that makes it unfalsifiable even in principle, and it is primarily for this reason that this can be so misleading in any systematic search for the truth of a claim.

The Availability Error...

...only makes the above more likely, and is an equally natural tendency to pay more heed, to more easily notice and remember, or that otherwise gets our attention, that data which is more easy to think about and imagine, more vivid, more memorable, more salient to us, and to pay less attention to, even forget, that which doesn't stand out to us and 'stick' in our heads.

This is evident when we derive assessments from information that really grabs our attention and interest rather than being accurate or reliable, and this leads to errors in reasoning such as the hasty generalization and the appeal to ignorance, as well as faulty anecdotal reasoning, such as our propensity to pay attention to testimonials, develop and hold superstitious beliefs, and make sweeping claims about an entire set or class of things from far too few examples of it, which is one way stereotypes are formed and perpetuated.

To make a valid, reliable assessment, it's a good idea to make sure that the sample used to make a generalization has enough members, and that all members of the population that the sample is taken from are equally likely to be represented.

Anecdotes are useful in science only if they are both true, and representative of the class they are drawn from, and even then they are best suited for hypothesis formulation rather than testing.

Misjudgments of probability are common when psychologically available information is used in assessing something's likelihood, such as errors in determining the relative chance of accidents in traffic, airplane crashes, shark attacks, and injury or deaths from malfunctioning amusement park rides, all of which tend to be meaningful, vivid, and lend themselves well to visualization.

The availability error is sometimes oversimplified in pop psychology as 'remembering the hits and forgetting the misses,' but disconfirming evidence can be memorable as well if it too is more psychologically available, such as losing a costly wager when playing in a casino or when betting on a horse race.

Neither of these tendencies should be taken as a slur against anyone's intelligence, sobriety, sanity, or integrity. We all do these things, including yours truly. It's just an unfortunate side-effect of how our brains perform their normal functions. But we can offset this error in otherwise reliable thought processes if not avoid it, like confirmation bias, and I'll offer a couple of suggestions below:

Try to come up with many explanations for a claim, while looking for alternate views, even opposing ones, and attempting to avoid hasty judgments, for these are often greatly misleading. Try then to think of and test ways to show each of these hypotheses in turn wrong rather than right, which leads to a more fertile understanding, and a more decisive though provisional evaluation.

Try to seek out all of the relevant evidence, time and energy allowing, rather than just that which stands out, or that which most easily comes to mind, since even seemingly unimportant or counterintuitive data can be vital for testing the truth-value of a claim and its attendant hypotheses.

Reference (How to Think About Weird Things: Critical Thinking for a New Age, 4th edition, by Theodore Schick Jr. & Lewis Vaughn - pp. 67-73.)

No comments:

Post a Comment

Comment below. No awards please. Your comments are my award for this blog. Word verification has been turned on because of spam.