What's wrong with stereotyping?

Okay, so according to Google, a stereotype is,

a widely held but fixed and oversimplified image or idea of a particular type of person or thing.

Everybody knows what a stereotype is, and most people think that it's wrong to hold stereotypes, particularly with respect to groups of humans, but why?

Consider that we use stereotypes in science all the time, and nobody thinks that there's anything wrong with that.  We apply stereotypes when we examine plant and animal behavior.  We use oversimplifications and analogies when we consider the behavior of fundamental particles.

So what's so bad about using stereotypes of human beings?

Any stereotype that is false is bad.  So, we might think that stereotypes about groups of humans are bad because they are false.  Most probably are, so that's a fair point.

Furthermore, stereotypes can be false in different ways.  They could be completely false, meaning that no one in the group possesses the feature claimed by the stereotype.  For instance, a stereotype like "All Asians are capable of backwards time travel" would be completely false, since no Asian is capable of backwards time travel.  But quantified claims admit of degree.  So, stereotypes could be false in a manner that is not completely false.  For instance, the stereotype, "All Asians are hard working" is not completely false.  It is still false, because there is almost certainly at least one person of Asian descent that isn't hard working.  When you make a complete generalized claim like "All x's are y," all it takes is one exception to show that claim false.  Even though the claim is false, it is not completely false, since it is almost certainly true that there are at least some persons of Asian descent that are hard working.

So, we might think that stereotypes are bad because they are false.  Either they are completely false, or they are false to some degree.  Any stereotype given as a complete generalization (All x's are y) is very probably false.

But suppose that we have some stereotypes that are not complete generalizations, but rather partial generalizations.  These could come in the forms like "Most x's are y," or "A lot of x's are y."  These claims seem more likely to be true.  Suppose that there are such claims that are true of certain groups.  For example, let's imagine the claim "A lot of Asians can't drive well" is actually true.  Social scientists went out and actually did the right kinds of studies and found that the majority of Asians are bad drivers.  Would it still be bad to believe in this stereotype?  If so, why?

One reason why it might be considered bad is that it assumed that human beings have free will, whereas the rest of nature doesn't have free will.  This means that even if 99.99% of a certain group of humans behaves in some way, we are not justified in inferring that the next member of that group will behave in the same way.  Since we have free will, we are not bound by the sorts of laws and principles that the rest of nature is constrained by, and thus predictions will always have a chance at failing.

Maybe that's why some people think stereotyping is bad.  Stereotyping denies our identify as free individuals.

Okay, well suppose that you don't believe that humans have free will, but you still think that stereotyping is bad.  Why would you think that?

Well, it's important to consider how stereotypes are used.  They play a role in predictions and decision making.  We often do these sorts of things with imperfect information, so we often take risks in making decisions.  Usually this isn't a problem, since the rewards typically outweigh the costs.  The reward for eating a medium rare steak at a high end restaurant is probably far greater than the risk of the steak being contaminated.  Perhaps when making decision based on stereotypes, the costs begin to outweigh the benefits.  Making a decision to violate an individual's human rights based on information pertaining to the individual's group membership may be a risky one, and may be the reason why people think stereotyping is bad.

Or, perhaps there is some confusion over what the term "stereotyping" means.  If stereotyping included all inductive generalizations about groups of human beings, and if stereotyping was bad, then it would follow that all of the social sciences is morally suspect, since it is essential for social sciences to make inductive generalizations about groups of people.

So, maybe we want to back and say that stereotyping does not include all inductive generalizations about people.  So what then do we include?  We probably want to include false generalizations that are popular, but what about true generalizations that are unflattering to certain groups?  What if it turned out to be true that most Asians were bad drivers?  Would this be included as a stereotype?