Human Rights Promotion (11): Intentionality Bias Causing the Surge in Human Rights Talk

First, there has indeed been a surge in human rights talk over the past decades and even centuries. This is particularly obvious for the period since the end of WWII. Human rights have become the lingua franca among the oppressed, the persecuted and the bleeding hearts worldwide, effectively replacing language based on benevolence, honor etc. (No insult intended, I’m a bleeding heart myself). There’s something about the notion of a human right that captures the strength of demands for freedom and equality like nothing else. It makes a claim sound very strong and difficult to ignore.

Other reasons for the popularity of human rights – or better the fascination with human rights – are their clarity and simplicity, their obvious universality and the fact that they cover most if not all areas of human suffering, depravity and failing, including persecution, violence, lack of freedom, discrimination, poverty, work and the family.

A further, and as yet unexplored reason is the so-called intentionality bias. The intentionality bias is a psychological bias where actions are viewed as intentional even when they’re not.

Three studies tested the idea that our analyses of human behavior are guided by an ‘‘intentionality bias,” an implicit bias where all actions are judged to be intentional by default. In Study 1 participants read a series of sentences describing actions that can be done either on purpose or by accident (e.g., ‘‘He set the house on fire”) and had to decide which interpretation best characterized the action. To tap people’s initial interpretation, half the participants made their judgments under speeded conditions; this group judged significantly more sentences to be intentional. Study 2 found that when asked for spontaneous descriptions of the ambiguous actions used in Study 1 (and thus not explicitly reminded of the accidental interpretation), participants provided significantly more intentional interpretations, even with prototypically accidental actions (e.g., ‘‘She broke the vase”). Study 3 examined whether more processing is involved in deciding that something is unintentional (and thus overriding an initial intentional interpretation) than in deciding that something is unpleasant (where there is presumably no initial ‘‘pleasant” interpretation). Participants were asked to judge a series of 12 sentences on one of two dimensions: intentional/unintentional (experimental group) or pleasant/unpleasant (control group). People in the experimental group remembered more unintentional sentences than people in the control group. Findings across the three studies suggest that adults have an implicit bias to infer intention in all behavior. This research has important implications both in terms of theory (e.g., dual-process model for intentional reasoning), and practice (e.g., treating aggression, legal judgments). (source)

If there is indeed a tendency to view actions as intentional, then there will also be a tendency to frame problems in terms of human rights. For example, if the intentional actions of an oppressive majority assisted by prejudiced legislators and law enforcers are believed to be the main cause of discrimination of a racial minority, then holding those intentional actors legally and judicially responsible for rights violations makes sense and may be effective. When, on the other hand, a lot of this discrimination is in fact the result of unconscious bias, or when it is statistical discrimination rather than taste-based discrimination, then judicial action based on human rights is much less effective.

And it’s my opinion that a lot of human rights violations are unintentional, unconscious and statistical. That doesn’t mean we should stop framing the underlying problems in human rights terms, but it does mean that our efforts to do something about them should be non-legal and non-judicial. Story telling, making people aware of their unconscious biases against certain groups of people, incentivizing people and other strategies can then be more successful in stopping rights violations.

The intentionality bias can be understood as an example of the fundamental attribution error: the tendency to over-value dispositional or personality-based explanations for the observed behaviors of others while under-valuing situational explanations for those behaviors. A simple example, if Alice saw Bob trip over a rock and fall, Alice might consider Bob to be clumsy or careless (dispositional). If Alice later tripped over the same rock herself, she would be more likely to blame the placement of the rock (situational).

More on human rights and intentionality is here, here and here. More on biases is here.

What is Equality? (4): Equality of Opportunity

I wasn’t very pleased with my previous attempts, so here’s one more. Equality of opportunity is a type of equality that’s usually seen as a very moderate one, one that’s not too demanding – especially compared to other types of equality that focus on equal outcomes – and hence it’s supposed to be acceptable to those of us who are a bit squeamish about equality. However, I’ll try to show that this is a mistake. Real equality of opportunity is a very ambitious and difficult project. In order to show that, I’ll talk first about some of the causes of inequality of opportunity, and then about the things we can do to reduce this inequality.

Four source of inequality of opportunity

1. Unequal endowments and circumstances

Equality of opportunity means that different people should have an equal chance of success in a certain life project, conditional on the willingness to invest an equal amount of effort. Of course, in reality, people will never have such an equal chance. The lottery of birth means that we are unable to choose to be born in a wealthy family with caring parents who can finance our education and motivate us to achieve our goals. It also means that we can’t choose which talents and genes we are born with. Talents and genetic differences are no more a matter of choice than the character and means of our parents. And genetic differences affect our talents, skills and maybe even our capacity to invest effort. (It’s not impossible that they even determine our choices of projects in which we want to be successful). So two people with the same life projects will only rarely have the same chance of success.

What can we do to equalize their chances? We can’t (yet) redistribute beneficial genes or disable harmful ones, and we don’t want to intervene in people’s families (and force parents to behave in a certain way or possibly even redistribute children). So we can’t remove the impact of genes and parents. But we can correct it, partially. For example, we can compensate people born with a genetic defect that reduces their chances of success in their life projects. We can offer people suffering from a genetic disorder that has left them paralyzed certain instruments to enhance their mobility. We can offer children born in dysfunctional or poor families free education, child benefits and encouragement. Etc.

2. Discrimination

Equality of opportunity also means correcting for lack of opportunity not resulting from the lottery of birth. If African Americans are systematically discriminated in employment, then they don’t enjoy equality of opportunity. They don’t have an equal chance of success in employment. If working for a certain company is part of an African American’s life project, and this company prefers white employees, then this African American doesn’t have an equal chance of success in his life project compared to whites with the same project.

The rule of equality of opportunity is only violated when the African American is rejected for no other reason than his race, and when this rejection diminishes that person’s opportunities (in other words, when this rejection is common and widespread rather than occasional; see here). If his skills, talents, merit and efforts are equal to those of other candidates, he should have an equal chance of employment or advancement. Equality of opportunity means that he should be allowed to compete for positions on equal terms, and that the difference between winners and losers in such competitions can only be a difference based on skills, talents, merit or efforts. However, even when he is rejected for the position because his skills, talents, merit and efforts are below the level of those of other candidates, he may not have been granted equality of opportunity. That is because the lottery of birth (point 1 above) has landed him in a discriminated group and because his lesser skills and willingness to invest effort and strive for merit may be caused by this discrimination.

Even if all are eligible to apply for a … position and applications are judged fairly on their merits, one might hold that genuine or substantive equality of opportunity requires that all have a genuine opportunity to become qualified. (source)

3. Misfortune in life

The natural lottery can reduce your equality of opportunity. Misfortune in the circumstances of your upbringing (bad parents, bad schools etc.) can also do it. And discrimination throughout your life as well. On top of that, other types of misfortune can limit your opportunities: you may get sick or have an accident. So we have to promote equality of opportunity at every step in people’s lives.

4. Neglect of abilities and talents

And there’s yet one additional cause of inequality of opportunity. Until now, I’ve assumed that equality of opportunity means that different people should have an equal chance of success in a certain life project. But maybe people have an equal chance of success in whatever life project they choose (as long as the project is morally acceptable of course). If society recognizes, rewards and encourages only certain talents and abilities, then some people will not be able to be successful in the life projects that they choose and that are compatible with their talents and abilities. For example, it’s fair to say that someone like Elton John would not have enjoyed equality of opportunity in Sparta or Saudi Arabia.

How to promote equality of opportunity?

If we accept all that, then the promotion of equality of opportunity involves different things:

1. Social structures or traditions

At the most basic level, it means getting rid of social structures or traditions that assign people to fixed places in a social hierarchy, to occupations or to life projects on a basis that has nothing to do with skills, abilities, talents, merit and efforts. Patriarchy, in which women are forced to focus on family life and raising children, is incompatible with equality of opportunity. As is a caste society, a society in which racial or other minorities (or majorities) are systematically discriminated against, or a class society in which the class of your parents, your blood line, your religion, your friends and relationships (nepotism) determine your chances of success in life. Getting rid of such social structures and traditions may simply require legislation outlawing them, or may also require affirmative action or positive discrimination and other forms of compensation for past wrongs (if some still benefit in the present from past wrongs, then equality of opportunity will not be respected simply because the wrongs have ended).

2. Equalizing skills, abilities and talents

But the promotion of equality of opportunity also means equalizing skills, abilities and talents, to the extent that this is possible (e.g. offering poor children free education of the same quality as the education and private tutoring offered to children born in wealthy families). And compensating people when this isn’t possible (e.g. give a blind man some help if we can’t cure his blindness).

3. Upgrading ambitions

And the promotion of equality of opportunity means reducing differences in merit and effort that are not the consequences of people’s voluntary choices. E.g. a child raised in a poor and dysfunctional family may have involuntarily adapted her ambitions downwards. Helping her at a young age may allow us to prevent this down-scaling of ambitions. Perhaps this down-scaling of ambitions is the result of the structures and traditions described in 1 above.

4. Other types of misfortune

The promotion of equality of opportunity also means helping people whose skills, abilities, talents, merit and efforts have been limited by misfortune different from the misfortune caused by the lottery of birth. If two people have the same ambition, talents and skill to become a lawyer – perhaps after social corrections to their initial starting positions in life (e.g. free schooling for the poorest one of them) and after legislation providing equal employment access (e.g. for the black lawyer-to-be) – but an accident leaves one of them blind, maybe society should provide that person with law books in Braille and such.

5. Recognizing abilities

And, finally, it means that a broad range of talents and abilities should be recognized and rewarded in society, with the exception of those that involve limitations of other people’s talents and abilities.

Limits of equality of opportunity

So equality of opportunity is a very ambitious and far-reaching project, contrary to what people usually believe about this type of equality. Hence we have to limit it somehow. For example, it shouldn’t extend to people’s private lives. You can’t demand that the girl next door marries you even if that’s your project in life and even if you think you’re the best candidate who didn’t have his equal opportunity. And the girl can decide not to marry you simply because you’re black. A club of racists can decide not to accept your membership request. A racist restaurant owner can decide not to serve you food on his private property. None of this diminishes your equality of opportunity, at least not as long as enough of the same opportunities exist for you elsewhere.

There will be a problem of equality of opportunity if all or many restaurants, clubs etc. turn you away. But if that’s not the case, and enough of the same opportunities remain elsewhere, even businesses can discriminate on the basis of race in their employment decisions, as long as this practice is not widespread and not part of a wider system of discrimination not limited to employment. If, in a perfectly tolerant and egalitarian society, there’s one bakery insisting on being racist and refusing to hire or serve blacks, who cares? (More here).

Equality of opportunity and statistical discrimination

However, discrimination in employment doesn’t have to be taste based, as they say. It can be mere statistical discrimination. Is that a violation of equality of opportunity? I would say yes, because discrimination is discrimination and whatever the motives are – a taste for discrimination or just prudence based of statistical averages – it diminishes the opportunities of those affected by it. People who engage in statistical discrimination make no effort to assess the skills, merit and talents of individuals.

Discrimination (12): Is Price Discrimination Immoral?

Price discrimination – or price differentiation – is a commercial policy. A seller may want to sell identical goods at different prices to different classes of customers in order to increase market shares, reach otherwise unreachable groups of customers or profit from customers’ willingness to pay.

The question is whether we should treat this type of discrimination like we treat other types. In other words, should we label price discrimination as something that is morally reprehensible, and should we also make it illegal (which are two different things – lying is often reprehensible but only rarely illegal)?

The answer is: it depends. Some forms of price discrimination are morally neutral or even praiseworthy. Offering students or poor people discounts for museum tickets or public transport has some moral benefits. Other forms, however, are clearly reprehensible:

[O]ne field experiment examined discrimination against disabled people in the context of car repairs, finding that disabled customers received higher quotes than the non-disabled customers. To get at the nature of this discrimination, the authors first conducted a survey, which revealed that “mechanics believe the disabled approach 1.85 body shops for price quotes while the non-disabled approach 2.85.”  In a second field experiment, the authors instructed participants to say the phrase, “I’m getting a few price quotes.”  This significantly changed outcomes — disabled participants received much lower offers: “Importantly, the lower offers received by disabled testers after signaling a willingness to search are not statistically different from those received by the abled,” write the authors. (source)

This is an example of so-called third degree price discrimination, which is in fact a form of statistical discrimination because the price is differentiated on the basis of an attribute of a customer segment (in this case disability), and this attribute is taken as a proxy of the customers’ ability or willingness to pay (in this case, the disabled are willing – or believed to be willing – to pay more because they can’t or won’t be troubled with shopping around).

Clearly, not all third degree price discrimination or statistical discrimination is wrong (student or senior discounts are cases of third degree price discrimination), but in this example it is wrong. Offering the disabled higher prices simply because they are disabled and hence less likely to shop around, is clearly immoral, even if it’s not based on animus against the disabled. It aggravates the disadvantage that nature has imposed on the disabled, and it’s a typical case of exploitation.

Exploitation occurs when one party in a voluntary exchange between two (or more) partners gets an unfair price for the goods or services exchanged, and when this party enters the exchange from a disadvantaged position. A price is unfair when it is below what it would have been in a fair auction. It is beyond doubt that a fair auction would have allowed the disabled, who enter the exchange from a disadvantaged position, to pay less.

More posts in this series are here.

Discrimination (11): Types of Discrimination

After this post about the reasons why discrimination is wrong, I thought it might be useful to add something about the differences between some types of discrimination and in the process link to some older and forgotten posts.

The most common type or at least the most commonly referenced type of discrimination is explicit discrimination: “whites only” signs, apartheid, miscegenation laws etc. When this type starts to become more and more unacceptable, hidden discrimination often takes over. Someone who wants to discriminate no longer does so explicitly – because that’s illegal or socially frowned upon – but instead adopts a policy or a law that doesn’t reference the target group, that remains facially neutral and that, when applied, accomplishes the intended discrimination. The most famous examples are the Jim Crow era literacy tests for voters. These tests effectively excluded blacks from the franchise although blacks weren’t explicitly targeted. Decades of educational discrimination made sure that very few blacks passed the tests.

Both these forms of discrimination are intentional, but some cases of discrimination are unintentional. People may not intentionally aim to impose disadvantages on other groups, but the structures of society, as they have been influenced by decades of previous – intentional – discrimination, make it very hard to avoid the imposition of systematic harm on some groups. The enduring effects of slavery in the U.S. are an example. Some would argue that this isn’t discrimination at all, since discrimination is typically defined on the basis of intentions and not purely in terms of consequences or outcomes.

Then there’s unconscious discrimination (see also here). People often have unconscious motives for their actions. And finally there’s statistical discrimination (see also here), which is discrimination that doesn’t arise from prejudice or bias, but occurs when people use aggregate group characteristics, such as group averages, to evaluate individual personal characteristics (for example, employers avoiding to hire African Americans because it’s statistically more likely for an African Americans to be an ex-convict).

More posts in this series are here.

Discrimination (8): What’s Wrong With Discrimination?

Let me try out a few possible answers to the question in the title of this post:

  1. Is discrimination wrong because it’s differential treatment of persons? That can’t be the case, because we treat people differently all of the time: some people earn more than others, have better grades in school, have a lot of friends or none at all, etc. None of those types of differential treatment seem to pose a moral problem per se.
  2. Is discrimination wrong because it’s differential treatment of persons based on the mere fact of their membership of a group? That can’t be true either. We don’t allow the blind to drive a car, we don’t allow the mentally ill to run for political office, and we don’t allow inmates to move freely across the country. Again, usually this is done without violating our moral intuitions. Differential treatment based on group membership may be wrong in some cases but it’s not discrimination per se.
  3. Is discrimination wrong because it’s disadvantageous treatment of some relative to others? Again, the answer is no. We certainly impose a disadvantage on inmates when we confine them to a prison building.
  4. Is discrimination wrong because it’s disadvantageous treatment that is morally objectionable? If I, as the sole racist member of a tolerant society, refuse to eat at a restaurant owned by an African American or serve African Americans in my own restaurant, I act in a morally objectionable way, and impose a small disadvantage on some people. And yet these people are not discriminated against in any meaningful sense of the word.
  5. Is discrimination wrong because it’s differential treatment of persons based on some of their immutable characteristics, for example their gender or skin color? Again, the answer has to be no. Sometimes it’s justified to treat persons differently based on their immutable characteristics or on characteristics that are beyond their control. We don’t allow one legged basketball players to compete in the NBA. And, conversely, sometimes it’s not justified to treat persons differently based on their self-chosen lifestyle. Forcing vegetarians to eat meat, for example, is wrong.
  6. Is discrimination wrong because it’s differential treatment of persons based on characteristics that are superficial or irrelevant, for example when we refuse to eat somewhere simply because the restaurant owner is black? That can’t be the case either. If I prefer to marry a blond because I believe blond women are better wives, I’m not discriminating against non-blonds. And if I’m the only one refusing to eat in a black man’s restaurant, it’s hard for him to claim that he’s discriminated against.
  7. Is discrimination wrong because it fails to treat people as individuals and relies on inaccurate stereotypes? For example, because I believe African Americans can’t cook properly and I therefore don’t bother to check out the food of an individual African American restaurant owner? Again, discrimination must be something else, for two reasons. First, there are laws against discrimination based on correct stereotypes, so-called statistical discrimination. Employers can’t just refuse to hire young women simply because young women are more likely to have children and hence be absent from work. Second, I may rely on inaccurate stereotypes and not engage in discrimination. For example, if I organize a successful boycott of German classical music based on the erroneous belief that this music is inferior to French classical music, I don’t discriminate orchestras or record labels that produce German classical music.

So then, why is discrimination wrong? Discrimination is wrong because it is a denial of rights. Not all denials of rights are discrimination – for example I imagine that Kim Jong Il denies the rights of all North Koreans consistently without discrimination. But discrimination is a denial of rights, and a denial of a particular type, namely the unequal denial of rights.

But not every case of unequal denial of rights is a case of discrimination. For instance, if some people in police custody are beaten up and others aren’t that’s not necessarily discrimination. Discrimination is the unequal denial of the rights of members of a socially salient group, and for no other reason than their membership of that group. It must be a socially salient group, i.e. a group that is an important and stable distinction in society, not a fleeting and inconsequential identity such as the group of people with blue eyes or people of the age of 40. Those groups aren’t salient and hence won’t be discriminated.

So, discrimination occurs when rights are granted to some and taken away from others, for no other reason than their membership of a socially salient group.

What exactly is a denial of rights? It’s an action (or a law, a custom etc.) that makes it impossible or very difficult to exercise one’s rights. That means that there won’t be discrimination if for example one lonely racist shop owner refuses to service blacks. Those blacks can easily exercise their rights by avoiding the racist shop. Only if the number of such cases is high and those cases occur systematically will there be denial of rights and hence discrimination. The policy of an army to reject homosexuals is discrimination because homosexuals can’t simply go to another army to serve their country. Racial discrimination at the time of Jim Crow and segregation was real discrimination because it substantially reduced the options of blacks who could not exercise their rights elsewhere.

If we revisit some of the examples given above we can see that they are not necessarily cases of discrimination as it is defined here. Giving people low grades or low salaries can be justified if it’s done for good reasons, i.e. not simply because of people’s membership of groups. But even if it’s done because of membership, it’s not discrimination if it’s isolated and if people can easily go elsewhere for their proper grades or salaries. Discrimination isn’t just about unequal treatment but also about options. Not all unequal treatment is discrimination. It’s discrimination when it also takes away the rights of those treated unequally. And their rights are taken away, not when there’s a single instance of unequal treatment based on group membership, but when there are many such instances and people can’t easily exercise their rights elsewhere. Their rights in these examples are the right to education and the right to fair wages. The same is true in the case of the sole racist member of a tolerant society refusing to eat at a restaurant owned by an African American. The restaurant owner isn’t denied his right to sell his food because one guy refuses to eat there.

Racism (14): Race and Consumer Behavior

There’s strong evidence of racist sorting by people looking for a job (white job seekers often avoid working for black managers, and white workers quit their jobs more rapidly when a white manager is replaced with a black manager). A similar phenomenon is race discrimination by buyers.

Do buyers discriminate based on race? This column describes an experiment in the US that advertised iPods online from black and white sellers. Black sellers received fewer offers at lower prices, doing better in markets with competition amongst buyers and worse in high-crime markets. The authors find evidence of both statistical and taste-based discrimination. … [I]t appears that discrimination may not “survive” in the presence of significant competition among buyers. Furthermore, black sellers do worst in the most racially isolated markets and markets with high property crime rates, suggesting a role for statistical discrimination in explaining the disparity. (source)

The important question is indeed to what extent this “sorting” on the part of buyers is motivated by statistical discrimination or by taste-based discrimination:

  • Statistical discrimination means that race is used as a proxy for unobservable negative characteristics, maybe in this case a judgment about the probability that black sellers will be happy with a marginally lower sales price, given their statistically higher rates of poverty. Or perhaps there’s distrust based on unclear statistical judgments about the risk of buying fake or stolen goods, meeting sellers in an inconvenient or dangerous neighborhoods, or dealing with unreliable sellers who might not complete the transaction.
  • Taste-based discrimination occurs when people just don’t like dealing with black people for no particular reason apart from the difference in race.

The study cited above uses a number of clever ways to disentangle these two effects. For instance, the inclusion of white tattooed sellers, who also received fewer and lower purchase offers, suggesting that part of the differences are due to statistical discrimination. Another part, however, is just plain racism. Black sellers are at a significant disadvantage on average, and that’s due to both statistical and taste-based discrimination.

More on statistical discrimination here.

Discrimination (7): Statistical Discrimination v. Background Checks?

Employers often use background checks before deciding to hire someone. For example, they may check the criminal record of job candidates, their credit scores, health history etc. It’s somewhat understandable although not always acceptable that they are reluctant to hire someone who has been in jail, has been sick for a long time, or has proven to be undisciplined by not paying her bills.

Let’s focus on ex-convicts for the moment. These people have a hard time as it is, sometimes even for no good reason because they shouldn’t have been incarcerated in the first place (I argued here that many countries, and especially the U.S., put too many people in jail). So, allowing employers to use criminal background checks can force ex-convicts into a vicious circle: unable to find a job, they may be forced to go back to crime.

Furthermore, there’s a racial aspect to all of this: in the U.S., African Americans are more likely to be ex-convicts. According to some, this racial discrepancy is precisely the reason to allow criminal background checks. If employers aren’t allowed to check individual candidates, they will resort to statistical discrimination: they know that blacks are more likely to have a criminal record and so they won’t hire any blacks at all, just to be safe.

However, if you espouse this argument in favor of background checks, you essentially want to make things better for one disadvantaged group – blacks, who are generally disadvantaged in employment – by making things worse for an even more disadvantaged group, namely ex-convicts. And that’s assuming that employers will hire more black people if they can use criminal background checks; but assuming that means assuming there’s no racism. Helping a disadvantaged group by harming an even more disadvantaged group is plainly absurd, and you can only fail to see that it’s absurd if you have an overriding fear of government regulation. Regulation should be kept in check but not at any price. I think in this case regulating businesses and outlawing background checks is the appropriate thing to do.

Let’s turn briefly to another type of background check: credit scores.

[M]illions of Americans, as a direct consequence of looking for work, have lower credit scores. … The use of credit checks in employment decisions should be banned. It is a form of discrimination against the poor — the codification and enforcement of class barriers. It is therefore a form of discrimination against those groups more likely to be poor. (source)

It seems there’s a

growing tendency of HR departments to check the credit scores of potential employees apparently deeming this data to be an important predictor of employee behavior. This creates a Catch-22 scenario for the unemployed where you can’t improve your credit score unless you get a job and you can’t get a job until you improve your credit score. (source)

Apart from the obvious fact that credit scores seem to be a type of knowledge that is much less useful for an employer compared to a criminal record – if your house burned down and your credit score is low as a result, does that make you a bad employee? – there’s a real issue for the poor here. They shouldn’t be discriminated against just for being poor. It’s not just a lack of conscientiousness or discipline that can lower your credit score. Back luck and poverty won’t help either. Some say the free market and competition will take care of this: employers stupid enough not to hire good poor people simply because they have a credit problem will lose out. Their competitors who don’t engage in credit checks will hire them, and those businesses will acquire a commercial advantage. I don’t know. Seems awfully optimistic to me.

Discrimination (5): Statistical Discrimination

Definition

Here’s the New Palgrave Dictionary of Economics definition of statistical discrimination:

Statistical discrimination is a theory of inequality between demographic groups based on stereotypes that do not arise from prejudice or racial and gender bias. When rational, information-seeking decision makers use aggregate group characteristics, such as group averages, to evaluate individual personal characteristics, individuals belonging to different groups may be treated differently even if they share identical observable characteristics in every other aspect. … group-level statistics, such as group averages, are used as a proxy for the individual variables. (source)

Statistical discrimination, also called rational prejudice (or rational racism), occurs when some stereotypes are considered to be statistically “accurate”. Accurate here in the sense of statistical generalization or average behavior (or characteristics).

Some examples

Young drivers are on average more likely to be involved in traffic accidents. Hence, auto insurers impose higher auto insurance premiums on young people. This insurance discrimination is called statistical discrimination because it’s based on some level of accuracy regarding the stereotype, and not on plain prejudice of or disgust for young people.

Regarding life insurance premiums, the customer’s likelihood of dying is one of the most relevant variables affecting the insurers profitability. Since gender is highly correlated with life expectancy, it is therefore optimal for the company to adopt a policy setting higher premiums for men, even for those men who may be healthier and less risk prone than the average woman. (source)

Employers often use the average performance of an applicant’s group to determine whether to hire him or her.

You don’t hire a guy with a Mohawk as a receptionist at a law firm – even if he promises to get a hair cut. Why not? Because on average, … guys with Mohawks have trouble with authority. (source)

Similarly, people with facial tattoos are unlikely to be hired as CEO’s, perhaps because they signal an anarchic, counter-cultural, contrarian, nonconformist, bohemian and “wandering” attitude.

Why?

The reason people engage in statistical discrimination is the high cost of making case-by-case judgments and the relative accuracy of group averages. Given that it’s not a product of bigotry, irrational prejudice, hatred and dislike, is it correct to call it discrimination? It is in the sense that people judge other people simply on the basis of group membership, not individual merit, as in “regular”, “irrational” discrimination or prejudice. Yet it’s not discrimination because people engaging in statistical discrimination often don’t do so because of their dislike or hatred of the discriminated group but because of economic self-interest and risk avoidance, as is shown by the examples above. The stereotypes are not baseless as most stereotypes are (“Blacks are lazy”, “Jews are money hungry”, “women are only productive at home” etc.).

This discrimination is “statistical” because contrary to normal prejudice it acknowledges that a relatively large number of members of a group do not conform to the stereotype. Just as it isn’t irrational to claim that it’s statistically accurate that republicans in general are pro-life while at the same time knowing that many aren’t. “Irrational” prejudice doesn’t have so much tolerance for exceptions. “People see others as average members of their groups until proven otherwise”. (source)

Is it acceptable?

So we know what it is and why people do it, but is it morally and legally acceptable? I think it depends on the type.

While such discrimination is legal in some cases (e.g., insurance markets), it is illegal and/or controversial in others (e.g., racial profiling). (source)

The difference between acceptable and unacceptable statistical discrimination hinges on several facts, I think:

1. How accurate are the group generalizations? You can forgive profit-seeking insurers for imposing higher auto insurance premiums on young people. However, wage gaps between genders resulting from theories about lower productivity of women (said to result from higher female investment in child rearing and a focus on “caring jobs” as opposed to jobs that tend to pay well) are less acceptable because the generalization is more dubious and the category of “women” too broad. Generalizations may be more accurate when they are not about behavior but about more “technical” and less controversial things such as life expectancy.

2. What is the “weight of history”? Statistical discrimination of blacks or women is by definition less acceptable than statistical discrimination of young people, given the history (maybe that’s a stereotype about discrimination…). Differentiated treatment, even if it isn’t really discrimination, can still cause harm given a certain history. The harm can be feelings of insecurity regarding a possible repetition of the past, feelings of continued attacks on self-esteem etc.

3. What are the consequences of statistical discrimination? Paying a slightly higher insurance premium is a lesser harm than being pulled over by the police a few times a week because of your skin color, even if your skin color is a statistically “accurate” prediction of the probability that you engage in crime in your neighborhood.

4. To what extent is statistical discrimination self-fulfilling? To some extent it clearly is.

People think teen-age males are criminally inclined (and they are), this angers the teen-age males, leading them to commit more crimes. (source).

If everyone assumes that boys are not good at domestic work, the benefit of learning domestic work is zero because no one will hire you as a domestic worker. In some areas, this self-fulfilling process is obviously harmful, and hence statistical discrimination is harmful as well.

It is possible that, because employers are less likely to hire women in jobs that require labour market attachment, then women are more likely to be involved in child-rearing than men, and are less prone to acquire the skills that are necessary to seek and perform well in those jobs, confirming the asymmetric belief that employers hold regarding labour market attachment. (source)

This phenomenon is called the stereotype threat. On the other hand, there can be self-reversing prophesies at work as well. A-typical members of a group, who of course suffer from the prejudice in the same way as typical members, may incite the latter to change their behavior. E.g. responsible young drivers who pay too much insurance because of the typical behavior of their fellow group members, may pressure the latter to change their ways. In this way, statistical discrimination may work to undermine itself. The same result will occur because a non-typical member of a group may get relatively more exposure. A woman who’s relatively good at soccer – compared to other women – but just as good as an average male soccer player, may get more attention than the latter. This excess of attention may undermine the stereotype on female soccer players.