Thursday, October 29, 2009

“The Smell of Virtue”: Drive-by Research Trivializes Both Science and Morality

A smell study made a big splash this week. It was an item on Instapundit, the British press ran with it, as did Time, and ScienceDaily obligingly published an edited version of a university press release.

What generated such boffo attention? A research paper slated for Psychological Science. Among its chief attractions for the media are a confident, catchy title (“The Smell of Virtue: Clean Scents Promote Reciprocity and Charity”), brevity (at 1,500 words it’s the length of a large-ish blog post), and breezy quotability (“there is some truth to the claim that cleanliness is next to godliness”). If there’s such a thing as science by sound bite, this is it.

The authors are a trio of B-school academics: Katie Liljenquist is an assistant professor of Organizational Relationships and Strategy at BYU, Chen-Bo Zhong teaches Organizational Behaviour and Human Resources at the University of Toronto, and Adam Galinsky is a professor of Ethics and Decision Management at Northwestern’s Kellogg School. Liljenquist and Zhong are his recent doctoral students.

The first of their two experiments is in the contrived style of the “prisoner’s dilemma” genre. The test subject is led into a room and told that a stranger in another room was given $4 and promised that any portion he chose to send the test subject would be tripled. The stranger supposedly sent the full $4 and it is now up to the test subject to decided how much, if any, of the money to return to the sender. While the economically rational thing to do is to keep the entire $12, most people return a portion of the money. This delights economists and game theorists who take it as evidence of human irrationality, while social scientists see it as evidence of a deeply ingrained social impulse to altruism and reciprocity.

The experiment was conducted in two rooms identical but for a spritz of citrus-scented Windex. The Windex scent pushed the amount of money returned to $5.33 from $2.81. Put another way, the Windex sniffers pocketed an average of $6.67. These guys are unlikely to be taking a monastic vow of poverty anytime soon.

In the second experiment, test subjects were put in the Windex-scented or the non-scented room and “asked to work on a packet of unrelated tasks.” (Yes, that’s the entire methods description provided by Liljenquist et al. Sort of the scientific equivalent of yadda-yadda-yadda.) The packet contained a flier soliciting volunteers for Habitat for Humanity. The test subjects were asked to rate their willingness to volunteer. The Windex treatment increased interest in volunteering by nine-tenths of a point on a seven-point rating scale, just enough to tip the needle over the 4.0 point of indifference.

(How many interest-in-volunteering points does it take to actually get a Habitat for Humanity house built? Silly question. Lab-based social science is a form of Kabuki: you can marvel at its stylized beauty but you’d never mistake it for real life.)

What do the authors conclude from their data?
The current findings suggest there is some truth to the claim that cleanliness is next to godliness; clean scents summon virtue, helping reciprocity prevail over greed, and charity over apathy.
Gag. The editors at Psychological Science (to which I subscribe and in which I’ve published) should be ashamed of themselves for letting this sentence stand: it is sophomorically overwritten press bait. (Yet it worked!) Note too the inaccurate use of “clean scents” in the plural: only one scent was tested.

The use of only a single scent ought to have been a red flag to the paper’s reviewers: it severely limits the conclusions that can be drawn. Years ago Robert Baron found that people in a mall were more likely to help a stranger (e.g., pick up a dropped pen) when the scent from the Cinnabon store was in the air. Suppose the Liljenquist team had spritzed cinnamon bun aroma and it boosted the amount of volunteerism and returned money. Would they conclude that sweetness is next to godliness? (Awkward!) They’d be reduced to “nice smells promote nice behavior.”

The authors invoke the “broken windows” theory of criminal behavior which holds that “damage and disrepair in the environment promote lawless behavior.” So why not run an unpleasant scent condition? Volunteerism and returned money should decrease in a stale-smelling room. Working the cleanliness/reciprocity link in both directions would be a far more compelling result. But then it wouldn’t be nearly as nifty a sound bite.

The authors’ disconnection from matters olfactory is matched by their odd view of virtue, morality, and moral behavior. In setting the stage for moral implications they name check various deep thinkers: Aristotle, Ralph Waldo Emerson, Marcel Proust, and . . . George Lakoff. (Which of these names doesn’t belong with the others?)

Then they float their big idea: due to the symbolic association between physical and moral purity,
clean smells might not only regulate physical cleanliness, but may also motivate virtuous behavior.
So how does a clean smell “motivate” virtuous behavior? What’s the thought process? I think it might be something like this:
“Hmmm . . . smells clean in here. A decent sort of place. And I’m a decent sort of person. How much money to return? Here, take $10.”
A less moral person might think:
“Hmmm . . . smells like someone just cleaned this room. This place is a bit fussy—best not to annoy them. Habitat for Humanity? Sure, I’d volunteer.”
So how do Liljenquist et al. explain the motivation? They don’t!
The link from cleanliness to virtuous behavior appears to be a nonconscious one: in neither experiment did participants recognize an influence of scent on their behavior.
[Another yadda-yadda moment: we’re never told when or how the participants were debriefed about smell. As phrased, they could have told the researchers they noticed a scent but denied that it influenced them. Who on earth reviewed this paper?]

Team Liljenquist thinks Windex smell produced a more “virtuous” result without the conscious awareness of the participants. But can a person be said to behave virtuously when he is unaware of doing so? Am I virtuous when the wind pulls a $5 bill from my pocket while I’m distracted and floats it into the hands of a hungry orphan? Hardly. Yet by Liljenquist logic I’m a saint.

Moral reasoning holds no interest for the authors. Here’s how they see things:
These findings carry important implications for environmental regulation of behavior. . . . the current research identifies an unobtrusive way—a clean scent—to curb exploitation and promote altruism.
Despite the hype, this study isn’t about virtue or godliness, nor does it engage morality at a conceptual level. This paper is about a smug idea for controlling the behavior of others for their own good without their awareness. It’s about encouraging charity at the point of a trigger spray.


olfacta said...

B-school academics? Windex? Hmmmm. I smell product differentiation.

Avery Gilbert said...


Something like "Give your home that holy glow with new Ultra-Sanctified Windex"?

Anonymous said...

avery, glad to see some skepticism about this study. i posed my own questions about it on my blog. here's the link to it:

you can remove the link if you want to - just wanted you to see you're not alone in your skepticism, when the whole world seems to be jumping on board their bandwagon.


Anonymous said...

for me, the biggest signal that this is b.s. is the fact that this product doesn't even exist on the windex website. (well, there wasn't a citrus-scented windex when i looked for it on october 28th.)

this is product-launch hype disguised as science. it wouldn't surprise me if citrus-scented windex hits store shelves and the website in the next few days.

and doncha love the publicity shots with the squeegee? the p.r. person who came up with that idea gets a gold star.

- minette

Avery Gilbert said...


Glad to see I’m not the only one wrinkling his nose over this study. Too bad the mainstream media doesn’t exercise a little more skepticism.

You touch on a couple of topics I almost included in my post. I left them out because it was nearly as long as the original article and because I didn’t want to go all insane clown posse on it.

I’ll have more to say soon in full post.

Anonymous said...

oh, so you let ME go all insane clown posse on them ;).

yeah, i'm a bit surprised by the lack of "journo" skepticism. has critical thinking been outlawed in j school? i mean, these are basic questions journalists should be automatically asking, whether they work online, in print or on air. instead, they're reprinting the press release nearly verbatim, even hailing it as some great psychological discovery in places. very stepford.

if no one in the mainstream media calls this out and gets some answers, it will have been a genius marketing ploy on the part of j&j. i hope there is an investigative reporter - or jealous ad exec - working on it somewhere right now.

in the meantime, i look forward to reading your full post.

cheers, minette