Confirmation bias, confirmatory bias, statistics, perception,

See Saw
See Saw | Source

Selective thinking

It seems that ordinary folk are doomed to misinterpret information and evidence. Without specialist training and strict control only dumb luck seems to prevail when making decisions about evidence.

What has a violin, a glass of water and a sonic insect repellent have in common?

The violin

People cannot tell the difference between a good modern violin and a Stradivarius.

The glass of water

A pure glass of water is identical to a homeopathic remedy.

Sonic insect repellent

Shouting at the mosquitoes is just as effective as a sonic insect repellent device.

The Stradivarius

A Stradivarius can cost millions of dollars and command a full-time body-guard. Yet a study in 2011/12 showed using blind experimental techniques that there is generally no perceptible difference in sound quality between one of these old instruments and a quality modern instrument.

The glass of water.

Homoeopathic remedies are diluted in water beyond any conceivable ability to detect a single molecule of the original additive. A substance is mixed with pure water, then diluted, shaken, diluted again, shaken and so on many times until the original substance in the water is essentially zero. Therefore, there is no difference between a glass of water and a homoeopathic remedy.

Sonic insect repellent

These devices simply don't work. Every scientific study performed using them as test objects show no statistically significant advantage as an insect repellent. Once particular manufacture claims the device mimics the ultrasonic chirp of a bat and therefore scares away mosquitoes. However, the mosquitoes are not sophisticated enough to make the connection and this is why controlled studies show they don't work.

Yet people say otherwise. Why?

I seriously doubt that anyone who has spent $2,000,000 on a violin will even glance in the direction of a scientific study claiming they were overcharged by a factor of 10,000. You might find musicians who would roll on the floor in agony upon hearing a sub-standard G note from a modern instrument.

Homoeopathic patients will swear they feel better after treatment with nothing more than pure water. Practitioners honestly believe they have a real effect. Yet there is not only no scientific reason to suggest there could be any effect at all, blind studies show it to be no different to the well known (and real) placebo effect.

People who have bought sonic insect repellent devices will often claim they work, perhaps because they are used in conjunction with other techniques like being cleaner in the house or using chemical barriers; or perhaps because they don't want to be seen to have wasted money.

Whatever the reason, it seems there is a very strong and almost unavoidable psychological twist to human reasoning that inevitably causes the brain and evidence to logically clash.

In the case of homoeopathic activity, a respected scientist called Jacques Beneviste published a paper in the prestigious journal Nature which claimed that water had some kind if imprint left from an initial dissolved substance, long after it was diluted away to nothing. To cut a long and interesting story short, his career was terminated after independent auditing showed that his counting procedure was exposed to human counting bias. The researchers peering into microscopes were aware of the sample origin, and which was the control. Despite their best efforts, they were humanly incapable of avoiding a counting bias. The bias in fact was the only statistical effect to be seen. When the experiments were repeated blind, the statistical significance vanished, and with it, Jacques' career and the memory of water.

N-rays

Jacques Beneviste is not the only victim. Pons and Fleishman's Cold Fusion seems to fall into a similar category, and there is a famous historical case again originating in France known as N-rays.

The N-rays case is particularly interesting. René Prosper Blondlot (1849-1930) claimed to have discovered a new form of radiation. He followed the discovery of X-rays, and named his discovery N-rays because he lived in the town of Nancy, with a university of the same name.

These so called N-rays were confirmed by dozens of other scientists but each one was deluded by their own unconscious bias. They saw what they wanted to see. The experiment relied on a prism inside a metal tube, and the effect was apparently seen external to this tube. The flaw was revealed when an auditor secretly removed the prism and scientists still claimed to see the effect. When the scientists subsequently thought the prism was removed, but it was actually in place, the effect magically vanished. In other words, this was a hallucination caused entirely within the brain of the researchers.

More examples

I'd like to collect more examples. If you know of them please comment.

Here are some candidates:

  1. Dowsing
  2. Sickness from electrical overhead wires
  3. Benefits of magnets in the bedsheets
  4. The difference between Coke and Pepsi
  5. Beer judging
  6. Wine tasting
  7. Bargain shopping
  8. Speaker cables
  9. Digital HDMI cables
  10. Pseudo Science
  11. Having rolled a few sixes on a fair die, what comes next?
  12. Choice of breakfast cereal
  13. Whiter-than-white clothes washing
  14. Political voting
  15. Investor portfolios
  16. The role of colour in advertising
  17. Proof-reading
  18. Trademarks that are too similar but come from different companies
  19. Large pills work better than small pills, and coloured pills work better than white ones.
  20. Clothing worn by accused during a trial
  21. Owners of particular computer operating systems

Conclusion

Confirmation bias is when you first take a position and then only accept evidence that supports it. It seems easy to identify and therefore easy to avoid, but in practice, even skilled researchers fall into the trap, and everyday, billions of decisions are made by ordinary people that are biased. We can't help it.

Hopefully, having seen some examples, and how powerful is this bias, it will help you read critically, make better decisions, and provide a useful tool for life.

More by this Author


Comments 15 comments

Pcunix profile image

Pcunix 4 years ago from SE MA

Computer operating systems are different. Depending on your needs, your choice can matter a lot.


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

... have you seen them lining up in the Apple store? ;-)


Larry Fields profile image

Larry Fields 4 years ago from Northern California

Hi Manna,

Long time, no talk! Voted up.

A non-expert's take on the violin question. I don't know if I could distinguish between the big-name classic and a high-end modern one. It's almost beside the point. The dollar-value of the classical one is determined by a number of other factors, as well.

•Snob appeal.

•Historical value as a museum piece.

•The difficulty and expense of crafting a credible fake. If you hire a world-class craftsman to do the job and to keep his mouth shut for 1 million USD, then that places an upper limit on the value of your fancy new/old violin, doesn't it?

•The collectibles bubble in general, which is perceived as a hedge against inflation. But if the economics profession finds a reasonable way to keep inflation at low levels everywhere, then the famous violin, Elvis' guitar, rare postage stamps, and other stupid collectibles will fetch bottom dollar if and when their owners fall on hard times.

All things considered, is the big-name violin over-priced at the moment? I don't know, and neither do you. Rare Earth metals, anyone?


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

Hi Larry. You make great points about economics. I am not sure if it alters the point about objectivity. I've often thought that after a massive disaster the richest guy would be the one with the can opener, not the penny-black!


Pcunix profile image

Pcunix 4 years ago from SE MA

Yes, I have seen the lines at Apple. A neighbor had to wait 45 minutes to be allowed to buy an iMac this week!

In fact, when I first saw that in 2003, I started buying Apple stock :-)

But if you think that's some sort of "reality distortion field" as Apple haters sometimes say, you are very wrong. There are reasons why Apple is a better choice for some and why Linux is a better choice for others. There are reasons why Microsoft is increasingly becoming a poor choice, too, though all of that could be very different later.

If you are just someone who browses the web and sends email, you are right: it doesn't matter. If that's you, you don't even need a computer: a tablet is actually a far better choice. But for those of us with other needs, operating systems do matter.


Pcunix profile image

Pcunix 4 years ago from SE MA

Let's talk about distinguishing too.

My wife and I will be buying an HDTV eventually, but right now we are just looking around. One of the difficulties we have is that she discerns detail and colors much more sharply than I do. I can barely identify an HDTV from a non-HD and absolutely cannot see why she says the $1,500 set is "better" than the $700.00 model. I just don't see it; she does.

Might not it be the same with the violin? Just because you and I can't tell the difference doesn't mean that it isn't there. Did the study do double blind tests with real experts? Did they fool Yo-Yo Ma? If they did, I'll believe it. If it's just a computer saying the wave forms are similar, I do not.


Larry Fields profile image

Larry Fields 4 years ago from Northern California

Pcunix wrote:

"Might not it be the same with the violin? Just because you and I can't tell the difference doesn't mean that it isn't there. Did the study do double blind tests with real experts? Did they fool Yo-Yo Ma? If they did, I'll believe it. If it's just a computer saying the wave forms are similar, I do not."

Excellent point. A friend who died several years ago, was the world's top designer of high-end audio amplifiers, until that industry rode off into the sunset.

One of Steve's beefs with traditional engineers in his field was that they had convinced themselves that frequency response was everything. That meme stuck, in part because of confirmation bias.

Once you have achieved acceptable frequency response in your design, you need to be aware of phases-shifting of harmonics. And he was the only audio engineer to ever address that issue.

A world-class musician could probably distinguish between the amplifier that Steve made for his home use, and the world's second-best design. As Pcunix suggests, there may be a similar issue with violins.

Some maven claims that there's precious little difference in quality between a big-name violin of the past, and a modern high-end one. BFD. Does he really know what he's talking about. As is the case with ordinary mortals, smug wanna-be skeptics with tin ears can play the confirmation bias game too, without even knowing it.


Pcunix profile image

Pcunix 4 years ago from SE MA

Of course it all can swing back on itself, too. Two things can be different without one being better - the "better" can just be our bias toward what we are accustomed to.

That's why I sometimes get angry with people who make silly statements like "Apple sucks!". I have almost 40 years experience working with everything from 1967 to today and have used, programmed and done troubleshooting in both home and business environments with . If you can match my experience, you have an opinion I can respect. If all you have ever done is use Windows at work or at home. your opinion is of no value to me and shouldn't be to anyone else.


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

Here is the original article on the violin double-blind tests: http://www.pnas.org/content/early/2012/01/02/11149...

Considering the enormous influence the musician has over the sound of the instrument, I am not surprised that is the first-order effect, leaving age and value buried in the noise.

@pcunix: I can match your computer experience but won't say "Apple sucks" because that's not true. OSX is a proper pre-emptive operating system like almost all modern multi-tasking systems and the code-quality is good. My point is about having a pre-conceived conclusion, and then getting biased about subjective measures after that and I am sure there is a lot of that going on in that line-up at the Apple-store. To be fair, the same applies to people who are hooked into Windows, Linux, Amiga etc.

On discerning colour, you are right, women can discriminate colour better. Can she pick out a more expensive TV when it's shown in isolation without branding? Perhaps for extremes of the market-end, but I am sure that branded weasel-speak about how fine details are incredibly superior plays on people's confirmation bias. Marketing does this all the time.

@Larry The human ear-brain system is not very sensitive to phase-shift. I don't mean the timing differences between left and right from source - but from the signal itself where the signal starts - e.g. with a compression or not. At high frequencies where the harmonics are, the period is short, and the brain will have more trouble. So I think high-frequency harmonic phase shifts are going to be inaudible or massively drowned by all the imperfections in the listening space of the room, then to the speakers, then to frequency response, then to audible-freqency distortion.


Pcunix profile image

Pcunix 4 years ago from SE MA

No, she knows before we see the price. Then she laughs, saying "of course I picked the big price".


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

That's cool. Is she one of those who can see more colours than men?


Pcunix profile image

Pcunix 4 years ago from SE MA

I don't know. Is there a place you know of we can test that?

She might just be normal. I have 20/400 (yes 400) uncorrected vision and it was NOT corrected until well after I started school - i don't think my brain ever really learned to process details as well as normal folks can.


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

Here is something about it: http://www.post-gazette.com/pg/06256/721190-114.st...

You would need the services of a geneticist.


Larry Fields profile image

Larry Fields 4 years ago from Northern California

"The human ear-brain system is not very sensitive to phase-shift."

Does a cornet playing a sustained Concert B-Flat sound different to you than a clarinet playing the same note? Suppose further that both musicians are making a concerted effort (yes, bad pun) to minimize vibrato.

If the two instruments playing the same note sound different, why is that? I can think of three possible reasons: the differing harmonic profiles for the respective instruments; or the differing phase relationships with respect to the fundamental frequency; or both.

If you think that the former is the only thing that matters, here's a 2-step experiment. Part 1 is easy. Listen to an old lo-fi Louis Armstrong recording from the 1930s. Can you make out the different instruments. So far, so good.

Part 2 is either an expensive time-consuming real experiment, or an el cheapo gedanken experiment. Suppose that getting rid of the scratches from the old vinyl record is not a problem. Using modern audio equipment, can you also get rid of the tinny sound?

For each note in the piece, tweak the harmonics, in order to compensate for the inherently poor frequency response of the original recording equipment. Here's how.

For each note, make an educated guess about the relative contributions of each instrument to the total sound of that note. Then look up the harmonic composition of that note for that specific instrument. Then use an appropriate linear combination of harmonic enhancements to make that particular recorded note sound more like the live sound of the original combined instruments.

And no, harmony would not be a problem.

I predict that the music will still sound tinny, no matter how much or how little you fiddle with the harmonic profiles of each individual note. If I'm right, the old Sherlock Holmes saying would apply:

Once we have eliminated the impossible, whatever remains, however implausible, must be the case.

On the other hand, if I'm mistaken, then there's a great business opportunity for some enterprising audio engineer. Any takers out there?


Manna in the wild profile image

Manna in the wild 4 years ago from Australia Author

"Does a cornet playing a sustained Concert B-Flat sound different to you than a clarinet playing the same note?"

Yes of course because a 'note' is not really a single note. It's a wave-form that can be de-constructed using Fourier analysis. When viewed in the frequency domain, these two 'notes' will look very different. The common point is dominance about the fundamental frequency.

A tinny sound is a result of frequency response - lacking in amplitude at the low-end. You could possibly boost the lower frequencies, but due to information lost in the original recording, and the inevitable boosting of noise in the same operation, then the result won't be crash-hot.

Clicks and scratches are easily removed by using a time-delayed output. The advanced (in time) processing detects a click and inserts a blank-space (or interpolates) where the click is in the delayed output.

I still say amplitude and harmonic content is more important than phase.

I admit, mathematically, the frequency spectrum of a phase-modulated signal is identical to frequency modulation. In each case, the harmonics are theoretically infinite. But phase modulation is when the carrier phase modulation is the time integral of the FM signal.

Here are two equations. One suffering a phase shift compared the other:

w_1 = A_0 sin( bwt ) + A_1 sin( nbwt ) , n more than 1

and

w_1 = A_0 sin( bwt ) + A_1 sin( nbwt +p ) , n more than 1

The +p is a phase shift on the harmonic. The time-dependent view of these two waves will look different. The frequency dependent view will look the same.

The human ear has a very hard time discriminating p.

    Sign in or sign up and post using a HubPages Network account.

    0 of 8192 characters used
    Post Comment

    No HTML is allowed in comments, but URLs will be hyperlinked. Comments are not for promoting your articles or other sites.


    Click to Rate This Article
    working