Bad Stereotyping

I was in Texas and Oklahoma last week. In the course of the trip, I was in a number of situations where I had to make conversation with people I didn’t know. Looking back on those conversations, I realize that when I was talking to white, male businessmen and needed to come up something to say, I generally chose the subject of college football.

    For lack of a better word, let’s call this “conversational discrimination.” I don’t assume that every stranger I meet wants to talk about college football. But I drew an inference about my conversational partner, based on his membership in the “white-male-businessmen of Texas and Oklahoma group” and used that inference to direct my behavior. As Judge Posner reminded us, in his review of Blink, in situations where one doesn’t know a lot about an individual, it may “sensible to ascribe the group's average characteristics to each member of the group, even though one knows that many members deviate from the average.” As it turns out, my assumption was largely correct. I had lot of really great conversations about college football. (Let's be clear this was not a hardship: I'm happy to talk about college football until the cows come home).

     The reason this stereotype was so useful was that I used as much of the available information about my conversational partner as I could. The fact that I was in Texas and Oklahoma mattered a lot. I wouldn’t have assumed that I could talk about college football with a similar group of white male business types from, say, Silicon Valley. The fact that they were businessmen mattered, and not, say, graphic designers or actors. The fact that they were men and not women mattered, and I know from experience that if I’m choosing a sports topic for conversation with an black male businessman, I’ll probably guess basketball—particularly if the person I’m talking to is from the East Coast. The point is the accuracy of stereotypes is a reflection—in large part—of their specificity: the more information you can use to build a generalization, the better off you are.

     This is my third (and last) comment on the Ayres study. My first point, as those of you who have been following my thoughts on this know, is that price discrimination against black males by car salesmen is morally wrong. My second point is that it is a bad business strategy. My third—and in some ways most important point—is that its lousy stereotyping.

    Let’s go back to the study. The male and female, black and white testers who Ayres sent out to car dealerships all gave the salesmen the same set of facts. They were all roughly the same age (late twenties). They all drove the same kind of car into the lot. They all dressed neatly and conservatively. They identified themselves as college-educated professionals (sample job: systems analyst at a bank). And they said they lived in the upper-income Chicago neighborhood of Streeterville. The car salesman, then, has several pieces of data from which to create his stereotype. He has the gender, race, age, occupation, educational level, and class (or at least a class proxy) of his potential customer. And what did he do? With the black men, he zeroed in on age and race, and ignored everything else.

      In his critique of my analysis of Ayres, Judge Posner did the same thing. When he says that it may be  “sensible to ascribe the group's average characteristics to each member of the group,” the “group” he’s talking about is race. But why is Posner—like the car salesmen—so hung up about race? Wouldn’t it be just as sensible, in the case of black men, to define their “group” as the group of college-educated, upper income professionals? So too with Steve Sailer. He says that car salesman are acting rationally, based on the fact that black men—as a group—like to be seen overpaying for cars. I have made my feelings known about what I see as the motivation behind that particular comment. But let’s just focus here on its appropriateness. Why is Sailer—like Posner and Ayres’car dealers—so intent on zeroing in on what is only one of many available and relevant facts about the customer?

     The short answer to that question, I think, is that this is what racial prejudice is: it is the irrational elevation of race-based considerations over other, equally or more relevant factors.

    But let me make two other points. First, thinking of the Ayres study this way gives us, I think, some insight into the anger that continues to be felt in the African-American community over discrimination. Put yourself in the shoes of one of those black males in Ayres study. You go to college. You get a good job. You make a lot of money. You move to a posh neighborhood. And when you walk into a car dealership all of those achievemens—and what they signal about you—vanish, and the salesmen only sees  the color of your skin. Can you understand now why I’ve been hammering away on this subject?

    Second, some of the commenters to my previous posts seem to have been of the opinion that price discrimination represented a kind of shrewd, profit-maximization strategy by salesmen. Shrewd? Tell me what’s so shrewd about being given four critical facts about a potential customer, and deciding to discard three of them?

Comments

[Byrne Hobart]

This post is missing any evidence that the car salesmen discarded the data. The only variable in the experiment was race, and the experiment demonstrates that race has an effect -- but you'd need to do another experiment (keep race constant and vary age, for example) before you can argue that the car salesmen were only focusing on race.

[Malcolm Gladwell]

I'm not sure I agree. The reason salesmen use race in constructing customer stereotypes is, presumably, because they think it is a proxy for sophistication. The black man is assumed to be an easier target. But here the black men present with additional information that invalidates that stereotype. They are educated, wealthy professionals--and the car salesmen ignore that and cling to the race-centric proxy. As do, more importantly, Posner and Sailer.

[Sweetness]

Gladwell writes: "My third—and in some ways most important point—is that its lousy stereotyping....
The male and female, black and white testers who Ayres sent out to car dealerships all gave the salesmen the same set of facts [age, drove same car, dressed neatly, college-educated, and lived in upper-income neighborhood]...The car salesman, then, has several pieces of data from which to create his stereotype....And what did he do? With the black men, he zeroed in on age and race, and ignored everything else."

Or he held everything else constant. Wasn't the study designed to see if race mattered? All other things being equal, it did. But perhaps the salesmen would treat upper-income blacks differently from lower-income whites. We don't know.

The stereotying isn't lousy, from an amoral profit-maximizing perspective, if it works. Perhaps upper-income blacks are even more likely to negotiate ineffectively because they either aren't good at it or don't care (i.e., they have enough money to buy the car and don't want to bother with annoying haggling).

From the profit-maximization standpoint, the stereotying is only lousy if it doesn't work. And Ayres study does not prove that.

[Byrne Hobart]

The Posner/Sailer argument is more nuanced than that. You're claiming that dress, neighborhood, etc., totally invalidate race as a datum. They're arguing that even given that information (which the salesmen don't necessarily believe is true), race could still be meaningful.

Once again, you can't claim that the salesmen 'ignore' wealth and education unless you present some kind of data. Are you saying that they treat all white customers -- whether they wear suits, jeans, or nothing at all -- the same? Without studies that let one variable at a time float, the only conclusion you can draw is that race has an impact on price -- whether that's in addition to or regardless of other variables can't be determined from the data.

In addition, I'd have to wonder if the experience was the same at each dealership. Don't you think there would be some variation in prejudice? Was there any relation between the race of the car salesman and the treatment of customers? If black salesmen are fairer, it would validate the assumption that the other salesmen were being discriminatory. If sales behavior is constant across all races, it's a resounding affirmation of the Posner/Sailer theory that these car dealers are just making the best business decisions they can.

[DDrexler]

They are educated, wealthy professionals--and the car salesmen ignore that and cling to the race-centric proxy.

maybe their experience has taught them that black clients, regardless of education, class, etc., tend to be an easier target/big spender, etc.

even if this is a racial or cultural thing, i don't think it would be unchangeable. if people are made aware of this (blacks and non-blacks alike), they'll be better educated consumers and will know when and how to haggle. i'm not black, but in the past, i probably wouldn't have haggled very much - i don't really know how to do it and its not really something i ever saw my parents do. as i've gotten older, i've seen it done and am better at it now. this would also incentivize car dealer to be more "moral" by providing a fair price to everyone. easier said than done, i'm sure.

[Nate Strauch]

Great post, Malcolm. I think, however, you fail to provide any evidence whatsoever to the contrary. It's impossible to prove your assertions without numbers that it's a bad business strategy to quote blacks a higher price. Unless you can show -- all other variables held equal -- that black men are, on average, just as shrewd car buyers as white men, than you argument can't hold water.

[Santiago]

"Why is Sailer—like Posner and Ayres’car dealers—so intent on zeroing in on what is only one of many available and relevant facts about the customer?"
I don't think you're being fair to Posner, here. As far as I can see, he is the only one not taking a view about whether or not it's rational. He's accusing you of assuming it's irrational, the same way you are rightly accusing Sailer of assuming it's rational. Posner, as usual, is the only rational person in the debate. (I wish he were actually in the debate. Anyone know if the Becker-Posner blog does requests?)

[ANM]

FWIW, bad strategy and lousy stereotyping are synonomous in this case.

The fact that salesman continue to hone in on race suggests that it is more influential than a college degree, clothing, etc.

"Can you understand now why I’ve been hammering away on this subject? ... Tell me what’s so shrewd about being given four critical facts about a potential customer, and deciding to discard three of them? " Discarding is appropriate if those other three facts don't matter much (i.e. affect the buying price). That is, ignoring those other facts is rational if they are of little consequence. And besides, those facts may well influence the behavior of the salesmen, but that [i]all other things equal[/i] race matters - a black man doesn't become white after he dresses well and goes to college in the eyes of the salesmen.

If some form of clothing were associated with hard bargainers, and whites wore the clothing of the weak willed and the blacks of the strong negotiators, the black-white difference would probably be smaller.

[Howard Lew]

When I am being described, I am always taken aback when the first (and sometimes only) thing said about me is my race (I am chinese). On a day to day basis, my race affects my behavior much much less than my income, my profession, politics, etc., yet I am often reduced down to race.
Reading the comments on this blog sporadically over the last week or so, I am often amazed by how much rationality the writers ascirbe to economics and people's behavior. What has struck me over a lifetime is people's illogic and their ability to follow an idea down a road no matter how contrary the surrounding facts are and how dire the consequences. (one example: our current President)

[Malcolm Gladwell]

Let's go back to the Oklahoma example. Do I chose a different topic of conversation if I'm talking to a white Oklahoman businessman or a black Oklahoman businessman? No. In that case (unlike in New York, a more basketball centered town) the significance of race in constructing my stereotype is trumped by the sigificance of geography, profession and gender. What the Ayres study was picking up--in isolating race--was whether car salesmen were willing to make that adjustment, to discount the role of race in the presence of other potentially confounding variables. And what does the study find? That the salemen still make race a heavy consideration, to the tune of $1000 a car. That's a BIG discount.

[Sweetness]

"Reading the comments on this blog sporadically over the last week or so, I am often amazed by how much rationality the writers ascirbe to economics and people's behavior."

That's one of the underlying assumptions of economic analysis: people are rational actors; businesses seek to maximize profits, etc. Market-wide, this is generally true, but it is an assumption -- because it's how you isolate things to examine. Without the assumptions, analysis would be impossible.

[Sweetness]

Gladwell writes: "What the Ayres study was picking up--in isolating race--was whether car salesmen were willing to make that adjustment, to discount the role of race in the presence of other potentially confounding variables. And what does the study find? That the salemen still make race a heavy consideration."

But wouldn't you only know that if the variables were differentiated? E.g., we don't know from Ayres if the salesmen would treat a lower-income black buyer differently than a higher-income black buyer. If they would treat them differently, then they _would_ be taking into account variables other than race. But Ayres' work is silent on this.

[Byrne Hobart]

"What the Ayres study was picking up--in isolating race--was whether car salesmen were willing to make that adjustment, to discount the role of race in the presence of other potentially confounding variables. And what does the study find? That the salemen still make race a heavy consideration, to the tune of $1000 a car. That's a BIG discount."

Okay. What would be a more accurate number? If it should be lower, how much lower? If it should be equal, doesn't that discount cultural factors which correlate with race regardless of neighborhood or economic status?

[Sweetness]

Byrne writes: "If it should be equal, doesn't that discount cultural factors which correlate with race regardless of neighborhood or economic status?"

Is Gladwell's point that there are no cultural factors that correlate with race regardless of economic status? I'm having trouble seeing it, but maybe this is it.

If so, is that true? Are black and white businessmen from the same region earning the same income and living in the same neighborhood completely fungible? If not, could they be different (on average) regarding, say, spending habits?

[JewishAtheist]

Mr. Gladwell:

I still see only assumptions from your side. You don't provide any evidence that wealthy blacks and wealthy whites are equally good at -- and likely to -- drive a hard bargain. They might be, but you don't even pretend to prove it.

How can you claim Sailer is wrong when you haven't even investigated his claim? I believe the reason he is so passionate on this subject is not because he is racist but because the quality of discourse about race-related matters is so low and so infuriating.

In the complete absence of evidence for the rationality or irrationality of the car salesmen's decisions, you label one side correct and the other racist. That's hardly constructive.

[blueman]

Someone asked something like this in a previous thread - "If a white American man goes to Nigeria, and the locals charge him higher prices due to his background, is this racism, 'bad stereotyping' ... or does the salesman just want to maximize his profit?"

[different jeff]

Thanks for the follow-up post, Mr. Gladwell.

I think you made your point very cogently here. I missed the fact (and concommitant significance)that all the potential customers were presented as well-heeled professionals, that everything was controlled for but race.

I am curious about how you account for the fact (taking Sailer's word) that black salesmen overcharged as well.

Anyway, as a mid-thirties, white Okie who lives in the shadow of Owen Field (ain't a businessman though, just an aspiring librarian), thanks for taking another pass at this.

I look forward to reaching this passage in Blink.

[David]

A story and a couple thoughts. Sarte has an amusing anecdote in which he me meets a woman who is vocally anti-semitic. He asks her why she feels this way and she says that she was once had a bad experience with a Jewish furrier. Why, Sarte wonders to himself, doesn't she go around hating furriers? There actually is a good deal of psychological research that says that race is a more salient characteristic than class, weight, and just about everything else but gender. Lawrence Hirschfeld at Michigan is a good person to read on this.

Two thoughts - no matter how good a research paper is, it can't be examined in isolation. The Ayers papers is one of many, many papers on discrimination and many of the questions raised in this discussion could be easily answered by referring to the larger body of work. The issue here seems to revolve around whether the actions of the car salesmen are 'rational' and at least in so far as defining rational as maximizing sale price we can't know with the info we have. There are, however, numerous other studies in which similar discrimination occurs and it is quite clearly not rational in any meaningful sense. Unfortunately its hard to link to scholarly papers that most people would be able to access, here is a useful bibliography:
http://www.ilo.org/public/english/protection/migrant/papers/usempir/bibl.htm
If you look through some of the papers, one thing you'll notice is that the studies often go beyone Ayres in not just making the participants the same except for race, but often stack the deck againist the white confederate (e.g., white job applicants with criminal records vs black applicants without) and the results still tend to be the same. In some respects Gladwell chose an unfortunate example to make his case with b/c it does leave the door open to the possibility that this is rational behavior. Look at the larger body of work on the subject, however, and that door shuts awful quickly. It is only in the accumulation of limited studies that the larger picture starts to emerge.

Second, many of the arguements for the rationality of the salesman rests on the (tautological) assumption that it must be rational otherwise the market would punish it. As any non-ideologue economist will tell you, however, the market is full of inefficiencies and failures. As I infer above, the arguement that something must be rational simply becuase it persists in a market is pure tautology.

[ANM]

"Let's go back to the Oklahoma example. Do I chose a different topic of conversation if I'm talking to a white Oklahoman businessman or a black Oklahoman businessman? No. In that case (unlike in New York, a more basketball centered town) the significance of race in constructing my stereotype is trumped by the sigificance of geography, profession and gender. " It all depends on how much of the variance in preferences for basketball vs. college football may be attributed to race. If it's purely a function of one's region, then race wouldn't matter. But it isn't, as even in the city, baseball (and probably football) fans are more often white than basketball fans. In both this example and that of car salesmen, you refuse to admit that race matters after controlling for other factors.

I suspect that a greater proportion of black as compared to white, Oklahoman businessmen prefer basketball, even if most of both groups prefer college football.

Howard Lew, people refer to you as 'that Chinese guy' so as to distinguish you from others (usually when they don't know who you are). In a posh law firm, if you say, 'that white guy,' you're only talking about at least half the lawyers. People use the details that set you apart from others so as to identify you. If I lived in a small Chinese town, I'd probably be called the American or the white man or even 'round eye.' Race (and gender) is memorable, quite unlike referring to the guy who say, wears Puma sneakers.

David, you're right, this study doesn't have enough information. But the discussion is occuring because
A) Gladwell assumes that race shouldn't matter, all else equal.
B) Gladwell deems far more plausible (possibly because of point A) the possibility that salesmen are unconsciously racist rather than rational.

[JewishAtheist]

There are, however, numerous other studies in which similar discrimination occurs and it is quite clearly not rational in any meaningful sense. Unfortunately its hard to link to scholarly papers that most people would be able to access, here is a useful bibliography:
http://www.ilo.org/public/english/protection/migrant/papers/usempir/bibl.htm
If you look through some of the papers, one thing you'll notice is that the studies often go beyone Ayres in not just making the participants the same except for race, but often stack the deck againist the white confederate (e.g., white job applicants with criminal records vs black applicants without) and the results still tend to be the same.

Thanks, David. I don't understand why people can't simply respond with facts and studies rather than name-calling and attributions of evil motives.

[juniper]

has anyone seen the "west wing" episode where josh decides to respond to a criticism on a message board, and donna says, no, stop, don't, and he says, no, no, it'll be fine, i'll clear everything up. and later that day he realizes that the readers don't seem to be taking his response in the spirit in which it was intended, and donna says, really, josh, no, stop, don't, but he can't stop himself and tries again. and by the end of the day josh is fully cognizant of the fact that this message board is the bloody thunderdome, but he still doesn't want to walk away...

i guess art really does imitate life. at any rate, i'll be happy to move on to a new topic--any new topic--after this post. i mean, three cheers for topical debates, but the snake-eating-its-tail progression of this one is kind of wearing me out.

[Jesse]

I like David's last comment -- just because a market does something doesn't mean it's rational (just ask Greenspan about the dot com era). Sometimes the behavior is rational but the underlying assumptions are irrational. If my nephew things there are monsters under his bed, it may be an irrational belief, but his unwillingness to get out of bed is completely rational given the “facts” at his disposal.

I am black, and the best way for me to summarize my experiences in buying things or professional settings is this (bear with me, I'm a lawyer): When it comes to many people I deal with for the first time, they often cling to a rebuttable presumption that I am less intelligent, refined, articulate, etcetera because of my race. When I walk in to court or to a job interview or the Lexus dealer, I have the burden of producing evidence that I can string a couple of sentences together, that I actually am somewhat intelligent, or that no, really I do not want the teal coupe with gold 20" rims. Once I meet my burden, I usually get treated the same as any one else; occasionally I get treated better, kind of like the Chris Rock bit about Colin Powell -- "He speaks so well! He's sooooo well spoken!"

The Nigeria tourist thing is an interesting case. I'm sure there are a few white Nigerians given the colonial history, but I would guess that the market vendor is using white as a proxy for rich and if I walked up and kept my mouth shut, I would get the "local" price. Then again, the Nigerian salesman is probably more "rational" than the car dealer because if approach him wearing a pair of Old Navy cargo shorts, a pair of my Doc Marten sandals, and a Polo shirt, he's going to ping me as a Yank right away and charge more. Travis Whiteside the car salesman is still going to make me prove to him that I am no rube even if I am wearing my Jos A. Bank blue pinstripe suit with the Harvard tie (reminds me of Trading Places).

[Aaron]

juniper,

That scene was based on an experience Aaron Sorkin (the main writer on the show) had on a message board devoted to the West Wing. So, yeah, art does imitate life.

JewishAtheist,

I'd like to point out that there were some people interested in a rational discussion with facts and arguments. Those people got out-posted by more hostile people, but it wasn't just people from one side of the argument. It's hard to keep track of a discussion with 40 posts an hour or whatever.

[Charles H. Green]

Being white and having been married to a black woman, I have some second-hand clue of what Malcolm is talking about.
Every person whose skin is dark has a plethora of examples they can cite. In my ex-wife's case, it included being given an F on a freshman english paper because "you must have plagiarized it." (After much explanation, it was re-graded an A).

Or, showing up for a housing appointment to find the unit had inexplicably been rented in the last twenty minutes.

Or, being asked while leaving the hotel for the day's appointment to please clean the roon as soon as possible.

Here's the point, and I think, at least partly, Malcolm's point. Would any of the offending parties think that it advantaged them to be unconsciously racist? Did any of them profit from the experience of having said something horribly racist?
Is it good business in this day and age to hold stereotypical viewpoints about classes of people in these kinds of circumstances?

Word gets around. Reputation is faster and faster. Six degrees of separation is an overstatement. To believe that a reputation as a racist is somehow consistent with good business is increasingly obtuse.

A salesman might squeeze a little more margin out of a customer when his stereotype "fits," but he'll lose a helluva lot of repeat business when word gets around--which increasingly it does.

As the world gets more connected, offensive behavior becomes less and less profitable. In this regard anyway, what's profitable and what's "right" may be converging.

[Michael Blowhard]

Another example, if it all hasn't grown too stale already?

I once spent some time in Morocco. I'm white, and I was a college-age hippie backpacker at the time.

Where doin' business is concerned, Morocco is a little like a car dealership -- everything is up for bargaining.

Early on I went out shopping a few times, bought some stuff, and came back proud of myself. I'd bargained the Moroccan merchants down to 50% of what they'd been asking. Woo hoo. Could I play the game or what?

But when the Moroccan friend I was staying with heard the prices I'd paid, she looked at me in indignation, threw some insults my way, and accompanied me out next time I went shopping. She's dark-complected, and she speaks the language and she knows the game, and she bargained the price the merchant was asking down to *a tenth* of what was initially demanded.

What to make of this?

Pretty obviously, the Moroccan merchants who had seen me (white, blue-eyed, hippie backpacker) when I was on my own had jacked their prices way way up and had played me very effectively. They'd made me think I'd done a great job of bargaining when in fact they were squeezing top dollar out of me.

Now, the interpretive question: Did they behave this way because they were anti-white? Because they were anti blue eyes? Because they were anti hippie backpacker? (They might well have been anti all these things, but how could I be sure without getting to know them personally?)

Or did they jack their prices 'way up when they saw me coming because they thought they could get more money out of the transaction that way? Because they'd learned they can routinely squeeze top dollar out of white American hippie backpackers?

It seems to me that the first conclusion is lame and naive. Why should they care about "white," "hippie" or "backpacker" at all? They've got better things to do than cook up seething prejudices and act on them. They were busy people trying to feed their families.

And it strikes me that conclusion #1 is also pretty insulting toward the merchants I was dealing with. It's *assuming* they were acting out of racism (or anti-Americanism, or anti hippie backpackerism). Why on earth would anyone *assume* racism? In the first place, where's the proof? In the second place, shouldn't we be just a wee bit wary of assuming the worst about someone else?

It seems to me far safer, far more sensible, and far more gracious to conclude that the Moroccan merchants behaved the way they did because their motivation was making as much money as possible.

Were they making assumptions about me that might be construed as insulting? I suppose so -- white American hippie backpackers are gullible, vain, easily-taken-advantage-of, etc. I suppose so. I certainly wouldn't have felt flattered to learn it. But who cares, really? The smart response to learning such a thing (Moroccan merchants take advantage of typical characteristics of white hippie backpackers) is to wise up.

In any case, I'd never take this packet of assumptions (let alone their behavior) as racist. Why would I? I'd think instead, "Hey, they really know how to play the game, don't they? Time for me to raise the level of my game."