Do you know what you like?

Imagine you are in a study and you are asked to pick which of two faces you find more attractive. After you make your pick, you explain why you liked the one you did. Maybe you like blonds. Perhaps you prefer strong, bold features. Everyone can give reasons for why they found one person more attractive than another, but those reasons might be entirely after-the-fact rationalizations rather than the real reasons. My favorite study of the past few years did just that. Lars Hall and Petter Johansson asked people to pick which of two portraits they found most attractive. They then handed people their selected picture and asked them to justify their choice. On some critical trials, though, they used sleight of hand to reveal a sleight of mind: They handed people the portrait that they had NOT picked — the one that they had found less attractive. Not only did people fail to spot the switch (change blindness), they gave the same sort of justification that they would have given for the picture they did select. They looked at the picture they had been handed and came up with reasons why they preferred it. Had you picked a brunette and been handed a blond, you would justify your selection by saying you prefer blonds!

Watch this short YouTube video of their study, filmed by the BBC:

Hall and Johansson call this effect “choice blindness,” and they also show that people have the wrong intuitions about what they will notice: Most people are convinced that they would notice when handed the wrong picture even though they don’t. After all, we must have chosen that face for a reason! Their discover of choice blindness was published in Science in 2005. It was also their first scientific publication — not a bad start to their careers!

Now they are back with a new paper in-press in Cognition entitled “Magic at the marketplace: Choice blindness for the taste of jam and the smell of tea.” In the new study, they set up a taste test at a local market and asked shoppers to sample tea or jam.  Even with radically different flavors (cinnamon-apple replaced by bitter grapefruit), more than half missed the change and justified their selection of the wrong item!  As for judgments of attractiveness, such emotional taste preferences do not lend themselves to rational, deliberative explanation.  We have preferences,  but we don’t necessarily know the reasons for them, and when forced to explain, we can justify liking things that we actually don’t like.  As for many judgments about how our own minds work, we get it wrong.  Our intuitions might get it right for emotional preferences, but our meta-intuition (intuitions about our intuitions) get it wrong.

Their new paper is available as a pdf

Sources Cited

Johansson P, Hall L, Sikström S, & Olsson A (2005). Failure to detect mismatches between intention and outcome in a simple decision task. Science (New York, N.Y.), 310 (5745), 116-9 PMID: 16210542

Hall L, Johansson P, Tärning B, Sikström S, & Deutgen T (2010). Magic at the marketplace: Choice blindness for the taste of jam and the smell of tea. Cognition PMID: 20637455

Toyotas, media hysteria, and the illusion of cause

Independent investigative journalist Mike Fumento has been looking into the recent media uproar over Toyotas that suddenly accelerate out of control. Recall the tearful testimony before Congress of cars going out of control, media coverage worldwide documenting all of the deaths from sudden acceleration problems, and Toyota’s unprecedented recall. His surprising conclusion: There is little if any evidence for a Toyota sudden acceleration problem. Almost all of the documented cases can be attributed to driver error (e.g., using unsecured floor mats that slid against the accelerator, hitting the accelerator rather than the brake) or to actual hoaxes (see earlier reporting here, here, and here).

Today he has a new column in Forbes Magazine (print edition), reprinted here, that looks at complaints filed in the National Highway Traffic Safety Administration database. The NHTSA database includes 93 deaths from 73 crashes that drivers and the media attributed to sudden uncontrolled acceleration. The media has taken these numbers as indisputable facts, but Fumento looked into the database and found that anyone can submit an incident report and that nobody verifies the claims. He writes: “Anybody can enter anything. An entry filed by someone named “Damnable Liar” claimed his car accelerated to the moon because of a child seat problem. That was mine.” Yet, media reports regularly cite the NHTSA database deaths as evidence for the sudden acceleration problem. The commentary is a fascinating look at how the media — and all of us — tend to be overly trusting of anecdotes and fail to look for the simpler, alternative explanations.

As we discuss in The Invisible Gorilla, people seek causes in the events immediately preceding a potential consequence. In this case, if someone has an accident, happens to be driving a Toyota, and has no simple explanation for why they crashed, they will now be far more likely to attribute the crash to a sudden acceleration. Tearful congressional testimony about the cause of a crash is compelling even if it might not have been accurate. Fumento does an excellent job of documenting the media hysteria and investigating other possible causes of some of the well-publicized crashes. Check out his reporting.

Invisible Gorilla Tweets on 2010-07-10

  • If you listen to Wisconsin Public Radio, you can hear Chris and Dan right now on the Veronica Rueckert Show: http://www.wpr.org/rueckert/ #
  • Original study got it right and only claimed an association http://bit.ly/b9NOpE. Headline writers getting it wrong. #
  • Inaccurate headline alert http://bit.ly/aMxaPB. (good site, bad headline). No evidence 4 causal link in study… #cause #
  • Media reporting (again) that games & tv cause attention probs (re: http://bit.ly/b9NOpE). Hint: correlation ≠ causation #

Powered by Twitter Tools

Invisible Gorilla Tweets on 2010-07-09

Powered by Twitter Tools

compensating for risk when driving

Tom Vanderbilt re-posted an interesting discussion of risk compensation, the idea that people take greater risks if they think they are somehow protected. For example, they might ride bicycles more aggressively when wearing a helmet. Risk compensation can go the other direction as well. When drivers are distracted, they do, sometimes, compensate somewhat for that distraction. For example, when drivers are just following at some distance from another car and don’t have to pass, brake, or engage in any other tactical driving maneuvers, they tend to drop back a little farther when they are talking on a phone. Although the effect could just be due to general slowing when distracted, it looks like they might be compensating somewhat for their impaired response time when on a phone. Unfortunately, compensation of this safer sort has its limits too. In a study I did with Bill Horrey a few years ago, we put drivers on a divided highway with traffic in a high-end simulator and simply asked them to drive as they normally would. In some cases, drivers were distracted with a task designed to simulate the cognitive distraction of talking on a cell phone and in other cases they weren’t. When drivers happened to be following another car in the same lane for an extended time (known as steady-state following), they compensated for the distraction, maintaining greater headway. However, whenever they were performing a tactical maneuver such as passing another car, they not only failed to compensate for distraction, they actually drove more dangerously when distracted.

This study suggests that the main impact of distracted driving might be on those situations in which we have to make driving decisions. When deciding when it’s safe to pass or change lanes, distractions make us drive even riskier than we would otherwise. In other words, we don’t compensate for distraction when it actually matters.

The idea of compensation is particularly worrisome given that most regulations on the use of phones while driving ban only hand-held phones. The evidence suggests that hands-free phones are no safer than hand-held ones, but if people think hands-free phones make them safe, they might be tempted to “compensate” by making calls in even more dangerous situations.

Horrey WJ, & Simons DJ (2007). Examining cognitive interference and adaptive safety behaviours in tactical vehicle control. Ergonomics, 50 (8), 1340-50 PMID: 17558673

Invisible Gorilla Tweets on 2010-07-08

  • In the August issue of Kiplinger's Personal Finance magazine, Bob Frick writes about the perils for investors of… http://fb.me/CuxZgAYh #

Powered by Twitter Tools

Invisible Gorilla Tweets on 2010-07-06

  • Chris and Dan have a new essay at PBS Need to Know on why Obama's comparison of the BP spill and 9/11 fails to… http://fb.me/sqtyLGiy #

Powered by Twitter Tools

Invisible Gorilla Tweets on 2010-07-05

  • Updated tweet plan: @invisgorilla -themes related to The Invisible Gorilla. @profsimons -Dan Simons on science, perception, etc. #

Powered by Twitter Tools

Invisible Gorilla Tweets on 2010-07-03

Powered by Twitter Tools

Invisible Gorilla Tweets on 2010-07-01

Powered by Twitter Tools