Stereo types and statistics

[InfiniteNow]

This would be considered a sexist stereo type.

Everything sebby said, and also this would be considered an individual's interpretation of study data, not a sexist stereotype.

[EWright]

First of all, this is a very poor thread on the subject of stereotypes. There is a significant difference between generalities and stereotypes, as stereotypes tend to be more harmful despite this post:

 

Stereotypes always have a basis in truth. People dont have the time or inclination to treat everybody as an individual, so we use shorthands, clumping people together on the basis of a shared custom or habit. The problem, if it is one, comes when the steroetype becomes outdated. A group may find itself stuck with a steroetype that simply does not fit, not because the steroetype is wrong per se, but because it is outdated.

 

so NOT always have a basis in truth. Some stereotypes are CREATED as a means to be degrading or hurtful. As a person of color, I'll point to the example of black people eating watermelon and imagery in American history that surrounded this fashioned stereotype. This imagery was meant to portray black people as ignorant characatures of less-than-human watermelon sucking animals.

 

This sort of stereotyping is meant to keep a social group "in its place" and to keep a group from achieving better things for themselves, by persons who are hateful or fearful of it. There were many such examples inthe early part of last century, and unfortunately many of those still exist.

 

Now generalizations are usually based moreso in truth but can be perceived in greatly exagerated ways that then become stereotypical in themselves. One post stated as fact that black people can run faster than other people. I would say that sort of thinking is problematic. As far as speed it may be that the at the very top levels, the fastest black athletes usually outperform whites, but I don't think that among "average" blacks and whites that is always true.

 

It is also not necessarilly "bad" to generalize. As the post I quoted above points out, there are reasons for grouping people in order to simplify our perceptions and there's really no way around it. If you're alone on a dark street and you become uncomfortable when a certain type of person or group of persons is around, that may be a mechanism that causes you to take measures to keep yourself safe, such as leaving the area. The unfortunate side is that is may also be a perfectly respectable person or persons who recognize the fact that their appearance caused you to be afraid, and end up being the ones who are hurt by the situation. An important point here is that stereotypes can lead to this sort of fear in an individual which can cause a cycle of problems in a society.

 

Myself, I play cards for a living. I often make generlizations about the way people play based on their race and sex and age. "In general" I'll assume black men are more likely to bluff (but certainly not always); women are more likely to call bets, so you don't want to bluff at them, but you will often be more assured of getting their money if you have a hand; and retired white men do not bluff and they won't even raise you unless they have the absolute nuts. Now these generalizations can and do cost me money at times, because they're not anywhere near 100% and must be adjusted according to other factors at the table, but they help to a certain degree and can be given a certain amount of value in weighing decisions.

 

So to summarize, since I'm rambling on much longer than I intended to, I think stereotypes tend to be more harmful. Some do have some basis in truth, but it is often exagerated in a way to be additionally harmful. Some generalizations are also blown out of proportion and can lead to ignorance by those who choose to believe them to such an inflated degree... and they can also lead to becoming harmful stereotypes... certainly the line blurs at some points. But overall, be careful about latching on to ideas that are spouted by others, especially when they seem to be demeaning... there's probably exagerated meaning there, with some ill-will to boot.

[HydrogenBond]

The point I was making is that statistical studies are very similar to stereo types in that they create generalizations that do not apply to everyone but are often sold that way. The exceptions to the rule are built into the study but are usally left nebulous to the layman. The average becomes the stereo-type. This interpretation may not be true to the scientists and mathematcians because they know better, but it is true for most average people.

 

If one looks at commercials peddling the latest drugs, they present the data, with marketeers pumping in the sunshine, with the hope that the average statistical stereo-type will be swallowed. If they presented the data, like a scientiist to scientists, it would be wasted advertising.

 

The alcohol blood level and the law is a good example. The statistical study has an average with respect to cost/benefit. From the point of view of a scientist the data also shows exceptions on both sides of the average. Yet society will market this study based on the stereo type average, making no provisions for those outside the average. I would think science would speak up and tell leadership they have butchered the interpretation of the statitisical math into a stereo-type. A stereo-type that can cause harm to people who are not part of the average, just like other negative stereo types do.

[sebbysteiny]

The alcohol blood level and the law is a good example. The statistical study has an average with respect to cost/benefit. From the point of view of a scientist the data also shows exceptions on both sides of the average. Yet society will market this study based on the stereo type average, making no provisions for those outside the average. I would think science would speak up and tell leadership they have butchered the interpretation of the statitisical math into a stereo-type. A stereo-type that can cause harm to people who are not part of the average, just like other negative stereo types do.

 

It seems to me that you have based most of your punch behind this example. You've mentioned it twice and stated that the system is wrong. So I think I will tackle this problem.

 

The issue about how spastic people who have drunk can drive is not about how much they drink but about how much alcahol is in their blood. The %age necessary to be a drunken killer is very similar in everybody.

 

A breathilisor measures the %age of alcahol in the breath of the driver. This depends entirely on how much alcahol is in the lining of the lungs, ie the alcahol blood level. Because the measurement is taken straight from the source of the problem, it doesn't measure how much someone drinks but how they are coping with what they drink. So all individual variation over amount they can tollerate wouldn't make less diference than hitting Mike Tyson with a feather.

 

I could drink 2 pints and not be noticed. A girl could drink half a pint and still fail the breathaliser test. It's true that the government gives a guidline of about 2 units or something, but it clearly says this may vary and one should know one's own tollerance.

 

The current policy is scientifically justified and needs no changing.

 

I think that has dealt with that example and therefore the entire argument.

[InfiniteNow]

The point I was making is that statistical studies are very similar to stereo types in that they create generalizations that do not apply to everyone but are often sold that way.

Neither statistics nor statistical studies create generalizations nor similarities. That act is reserved to the individual interpreting the statistics. The statistics themselves are just that, numbers resulting from observation. While lay people are often deceived by the creative use of statistics, it is still their responsibility to ensure they have an accurate understanding.

 

Statistically, your posts tend to lack adequate support and evidence to support your claims, but I still like the fact that you speak with your own voice and try to offer new ideas. However, the generalization I make that you are often FOS is my own interpretation. :doh:

[HydrogenBond]

I agree that most statistical studies are done in a scientific way. But after the results hits the mainstream it is left in the hands of subjectivity, where it is turned into scientific stereotypes. Such studies should be controlled by the scientific commnuity for proper interpretation instead of handed over to be subjectively massaged.

 

Here is an interesting example of scientific statistics gone crazy. On an AM talk radio station they mentioned a new study that correlated the increase in current childhood hyper active disorders to cigarette smoking. This may be valid scientific statistics, but it defies common sense. Think of it. About 50 years ago, smoking was fashionable. This ment that not only people who liked smoking smoked, but also those wanting to be part of the fad. It was everywhere and children were plagued by smoke. Yet the hyper activity was lower. Since that time, smoking has steadily decreased while hyper kids have increased, yet scientific statitics make a reverse correlation. Statistics, can be used to push forward subjective agendas.

 

The second half of the study was lead and hyper activity. Lead paint was the mainstream 50 years ago. Again low hyperactivity at that time where even lead gas was common. It was outlawed back in the 1970's and its exposure decline since, yet more kids are hyperactive. Again science statistics leading to a realtionships that defied common sense.

 

What that tells me, statistics can be used to justify illusion of science even before subjective manipulation by polititians and marketeers. In my opinion statistical science should be called pre-science or elementary science to distinguish it from rational science.

[sebbysteiny]

Neither statistics nor statistical studies create generalizations nor similarities. That act is reserved to the individual interpreting the statistics. The statistics themselves are just that, numbers resulting from observation. While lay people are often deceived by the creative use of statistics, it is still their responsibility to ensure they have an accurate understanding.

 

Statistically, your posts tend to lack adequate support and evidence to support your claims, but I still like the fact that you speak with your own voice and try to offer new ideas. However, the generalization I make that you are often FOS is my own interpretation. :lol:

 

I think I'm not understanding this correctly. Are you saying that statistics is just the raw data but any attempt to make a meaningful interpretation of them is a generalisation??

 

This is like saying that it is a generalisation to say that smoking significantly increases risk of lung cancer in the same way that saying being black significantly increases the risk of being a mugger.

 

One is a generalisation founded on ignorence. The other is a simple interpretation of fact founded by correct statistical analysis. Of course Hydrogen Bond will say "what's the difference?" right?

 

It is true that a disproportionate number of black people commit crimes (albeit not significantly) and it is true that a disproportionate number of Muslims end up becoming fanatical Islamic terrorists hell bent on killing as many innocent people as possible. But saying 'being race x makes behavior y more likely' is morally and intellectually wrong.

 

Cmon everybody. I dislike the point Hydrogen Bond is making as much as anybody but we can't ignore it. We must find the distinction between interpreting statistics based on medical data to find conclusions (right) and interpreting statistics based on racial data to find conclusions (wrong).

 

Hydrogen bond is saying that since the latter is clearly wrong, then as there is no distinction between the two concepts, the former must also be wrong.

 

I'm beginning to think that there is no statistical difference between the two. However, there is a huge moral difference between the two. If you like, Racial steriotypes are a special case of stastics where it is morally wrong to apply proper statistical analysis.

 

The reason for this is not because of the consequences of applying the correct statistics. Lets say there is a link between muggers and colour of skin. The correct statistical conclusion would be that we would stand a slightly higher chance of being mugged by a black stranger than a white stranger. Lets say this increased probability of being mugged is as high as a 30%. Then how should one respond walking past a black man?

 

Firstly, our level of fear when we walk past a white man is almost 0/10 with 10 being having a loaded gun pointing at your head. Lets say it is 0.0005. According to statistics, our fear should become 0.00065 when walking past a black man. In other words, our rational reaction in light of correct statistical analysis is to still have almost no fear whatsoever.

 

However, how will people react? I think the rating could get as high as 1 or even 2 / 10. This may be true even in people knowledgable about statistics. Further, that fear could cause some people to commit horrible acts of unjustified physical violence.

 

In other words, owing to fears of the outsider and fears of differences, the human mind is not physically capable of being rational when it comes to differentiating between races. However it can be rational when interpreting scientific data on lung cance risks.

 

So what should we do as a society? We can't have an entire race being demonised, threatened and, as has happened in many societies, exterminated owing to a total inability of the human mind in general to interpret statistics of race rationally. So our only choice is to demonise it and to stop all efforts to find and research such stastistics.

 

Being a decendant of the worst racial extermination campaign in history, I cannot agree more with societies decision to demonise that field of statistical analysis in its entirety to the point that even simple observations like 'blacks make better runners' is a taboo.

 

So that is, in my opinion, the distinction that you were looking for Hydrogen Bond. The human mind is not capable of interpretting racial steriotypes correctly and unless there is some major compelling reason, that field of statistical analysis must be kept in the intellectual maximum security prison for all time.

[HydrogenBond]

Sebbysteiny hit part of the nail on the head. Statistical stereo-types is more obvious when applying science statistics to racial differences. A statistical study is just a study and should be taken with a grain of salt. This would be possible if statistical science came with a how to interpret guide for the layman. Without this the layman see increased risk or low deviation, etc., like it is a rational relationship.

 

To say smoking increases the risk of lung cancer does not imply that anyone who smokes will get lung cancer. Nor does it imply that anyone who does not smoke will not get lung cancer. But it will be interpretted that way by the layman. If brand=X lowers body fat by 20%, there is no guarentee it will work on eveyone or that everyone will get those results. But this is how it will be pitched to the layman. Statistics allows too much room for a meaningful correlation to be subjectively fudged into a rational illusion of one size fits all, i.e, stereo-type.

 

Let me contrast rational science with statistical science with an example. If we look at rate of acceleration of an object in a vacuum due to gravity, it is governed by simple Newtonian gravity equations, which allow one to predict the result for any object in any gravity environment.

 

If there was no such thing as rational math or logic, one could get the same results with statistics, by running a lot of experiments. The problem with this is that each object would have to be treated separately as a potentially different situaiton. If a study used rocks of size A, and I was to suggest that rocks of size B should act the same way, the reaction would be "where is the data". I would have to repeat the experiments for size B. If I suggest golf balls will act the same. The reaction would be, "where is your data!: Because it is impossible to run studies on all objects at all gravity conditions, since the earth's gradient is limted, there will always be a margin of uncertainty left over until complete hard data is generated.

 

One may notice that one rational relationship can take the place of legions and legions of statistical studies. But if culture and science is biased toward statistics and chaos, rational relationships can be dismissed, with the statistical, i.e, where is your mountain of data. There is a tendacy in the human mind to see things rationally. The compromise is statistical studies with people attempting to draw some rational stereo-type to make random more in line with the rational nature of reality.

 

Statistics are useful but the blackbox approach is different than opening the blackbox and see what rationally occurs inside. But statistics does not want the blackbox to be opened due to the way it conducts business. Yet if the blackbox was opened there would be little need for statistics since a simple rational relationships would supersede this more primative form of science, i.e., pre-reason.

 

Hypothetically, if we were to outlaw statistics and require all science be done with logic and reason, we would all be better off. It would require people to think rather than depend on a statistical oracle to tell you how and what to think, i.e, blindman's prophesy. The statistical oracle leaves room for subjective interpretation unless one is a high priest. This is good for social politics but not science. Don't get me wrong, statistics is a useful oracle, but having something think for you atrophies your ability to think rational for yourself. The oracle becomes a crutch that should be used only when your leg hurts, but not all the time.

[InfiniteNow]

I'm not going to quote any one reference, cause there's a lot of stuff in this thread which is no where near the realm of "beyond personal opinion."

 

Statistics are mathematical tools developed in an attempt to help us try to better understand systems under study using our limited resources. Any generalizations or stereotypes resulting from the results of these calculations are generated by the interpretation of the observer. This further leads to problems with indidividuals focussing on spurious relationships, which they share, and a domino effect ensues.

 

Science is a methodology, not a group of people, not an administration, not an observer, not a religion, it's an approach. It's a dynamic way to approach the universe seeking better understanding, yet all those understandings remain relative to the interpretation of each observer.

 

Last, conclusions drawn from false premises are themselves false. Please be sure you know what you're talking about before speaking, or at the very least ask questions in an attempt to increase your own understanding.

 

Any fool can stand on a soapbox and yell, but the bigger fool is the one who listens to them.

 

Please feel free now to continue with whatever it was you were doing.

[sebbysteiny]

Hydrogen Bond. Do not get your hopes up. I believe the reason we seem to agree on anything at all that others are not agreeing with is because I have understood the point you are making and I have given it genuine consideration.

 

I think Infinite Now does not fully understand the debate.

 

But we have come to totally different conclusions. In my view statistics is a 100% valid and scientific way of finding links. I think that anybody who ignores statistics is willfully blind. They are, in my view, unquestionable. And I'm not just talking about the data, I'm talking about the conclusions, or generalisations as Infinite Now seems to be calling them.

 

Your dilema was that reaching conclusions based on statistical data about, say, medicine is considered intellectually correct. Reaching conclusions based on statistical data about race is considered intellectually incorrect. But since both are doing the same thing, they are either both correct or both incorrect.

 

I accept that.

 

But you have concluded that since generalisations on racial stereotypes are so obviously wrong (because a stereotype does not give the whole story), reaching similar conclusions about medicine must also be a generalisation and equally wrong.

 

However I have concluded the other way. I have said that reaching conclusions, or as you say, generalisations, about medicine is right AND SO IS MAKING GENERALISATIONS ABOUT RACE. They are both intellectually correct.

 

This seems to me to be an accurate summary of our positions.

 

If I am right, then why is making generalisations about a race considered 'racist' and intellectually wrong? The key stone in my argument is that the human mind, no matter how intelligent, is incapable of treating statistics of races rationally because of our genetically installed and unavoidable irrational fears.

 

From this, I wish to make a new meaning of 'stereotype'. A 'steriotype' is when you draw a conclusion far greater than the statistical evidence rationally permits. Thus stereotypes are only major intellectual problem for racial statistics.

 

I believe your attacks on statistics are again attacks on ignorant use of statistics. Lets look at your examples.

 

Here is an interesting example of scientific statistics gone crazy. On an AM talk radio station they mentioned a new study that correlated the increase in current childhood hyper active disorders to cigarette smoking. This may be valid scientific statistics, but it defies common sense. Think of it. About 50 years ago, smoking was fashionable. This ment that not only people who liked smoking smoked, but also those wanting to be part of the fad. It was everywhere and children were plagued by smoke. Yet the hyper activity was lower. Since that time, smoking has steadily decreased while hyper kids have increased, yet scientific statitics make a reverse correlation. Statistics, can be used to push forward subjective agendas.

 

If what you say is true, then it is physically impossible for the amount of hyper activity and cigerette smoking to correllate. After all, look at the data 50 years ago.

 

So either the statistics must be falsified or your arguments are flawed somewhere, ie based on a false premis. I would put my money on the latter.

 

So where's the flaw? Where did it say that hyperactivity has increased from 50 years ago (ie the premis of your statistical counter evidence was false)? What about the population increase? Does your argument consider that? Is smoking really much less popular today than it was then? Do you have evidence? And finally, in a grand catch all, there could be hundreds of alterior variables that could account for the difference including, for example, current living and working habbits.

 

So one needs to separate the variables and have what's called 'a fair test' where one has two groups whose only difference is that one smokes and the other does not. Then you continue to test the correlation until you are as certain as you want to be that it exists.

 

So in conclusion, I think your arguments disputing the correctness of the link are weak weak weak.

 

The second half of the study was lead and hyper activity. Lead paint was the mainstream 50 years ago. Again low hyperactivity at that time where even lead gas was common. It was outlawed back in the 1970's and its exposure decline since, yet more kids are hyperactive. Again science statistics leading to a realtionships that defied common sense.

 

Again, there could be a similar host of factors any one of which could fundamentally destroy one of your simplistic premises and therefore the whole argument. One of which might be increases in number of cars, a new use of lead that is commonplace, or even perhaps the effects of lead petrol from the development in the third world.

 

And on another point. So what if it defies common sense? Common sense said the world was flat. Common sense was wrong. Common sense is often wrong. That's why one needs training for certain jobs because common sense is not enough. If I had to chose between common sense and statistical evidence, I would chose the statistical evidence every time. That is yet another of your premises that is fundamentally flawed: that if something is contrary to common sense, it must be at the very least questionable.

 

If there was no such thing as rational math or logic, one could get the same results with statistics, by running a lot of experiments. The problem with this is that each object would have to be treated separately as a potentially different situaiton. If a study used rocks of size A, and I was to suggest that rocks of size B should act the same way, the reaction would be "where is the data". I would have to repeat the experiments for size B. If I suggest golf balls will act the same. The reaction would be, "where is your data!: Because it is impossible to run studies on all objects at all gravity conditions, since the earth's gradient is limted, there will always be a margin of uncertainty left over until complete hard data is generated.

Okay, this is the last argument I will expose as being heavily flawed because I think every one of your examples can be similarly treated.

 

What you are saying is wrong if my physics degree knowledge is to be believed. F=ma was Newtonian's law. There was no real reason for this. It was found exactly because of statistical data. Newton noticed the statistical correlation (the line on the graph between Force and accelartion etc) of these properties and proposed the equations. For hundreds of years nobody found any object that disobeyed the equations. But then some very fast objects stopped obeying this and it was clear that at velocities so high that they could not have possibly been achieved in Newton's day, Newtons laws break down. In steps Einstein with his own theory and finds his answer which now succeeds in produces the whole correlation from objects at light speed speed to zero. He then says that it is not possible to exceed light speed and therefore the corrleations of Einsteins equations correspond to the statistical data for all possible speeds and therefore all possible experiments. However Einstein's theories are only accepted because they correspond to the data.

 

So that's how science developes. You get statistical data. You note the correlation. You then explain the correlation. You then get a theory. That theory then works. But then you find a new situation that the theory stops working so you modify the theory or come up with a new one altogether. Either way, it is the science that follows the statistical data and not the other way round. Statistics is at the very heart of science and if you distrust statistics, you distrust the foundations of the entire scientific establishment.

 

Ergo, I cannot see any intellectually acceptable way of discounting statistical evidence that is correctly interpreted. Every one of your supposed illustrative examples is flawed in my opinion.

 

 

Finally, it is my turn to give you an example to illustrate the absolute correctness of statistics even though no theory has been found.

 

Ever since the beginnings of humanity, people noticed that if you had a sword stuck through the heart, you had a high probability of dying. Thus they noticed a statistical link between the two. It would be unthinkable in my opinion to question that link based entirely on statistics simply because the theory explaining the link, ie the understanding of the respiritary system and the blood system, had not been thought up until tens of thousand of years later.

[HydrogenBond]

Ever since the beginnings of humanity, people noticed that if you had a sword stuck through the heart, you had a high probability of dying. Thus they noticed a statistical link between the two. It would be unthinkable in my opinion to question that link based entirely on statistics simply because the theory explaining the link, ie the understanding of the respiritary system and the blood system, had not been thought up until tens of thousand of years later.

 

Most soldiers would have drawn that conclusion with far less data than needed to run a statistically valid study. One such data point may have been enough for common sense and logic to know this trick. But statistics would not have accepted that bold conclusion, without a few hundred more data points so it could be recognized as a valid study.

[sebbysteiny]

Most soldiers would have drawn that conclusion with far less data than needed to run a statistically valid study. One such data point may have been enough for common sense and logic to know this trick. But statistics would not have accepted that bold conclusion, without a few hundred more data points so it could be recognized as a valid study.

 

Wrong wrong and wrong.

 

The number of data points does not really matter. 5 could be enough to establish a valid and rigorous statistical link. It's called the statistics of small samples and it is mathematically valid.

 

But the point is that people notice statistical links and draw conclusions that are still valid even if the theory and explanation behind it is not known. You however trumped the idea that statistics are only valid and reliable WHEN there is a theory behind it backing it up and not when taken in isolation. That you have even accepted my example and even added your own little peace about statistics of small samples proves that your idea is highly doubtful.

 

But I notice another thing. You said, because of the statistical data, it became common sense. If you follow that logic to its natural conclusion, then since all sampling of statistical data uses identical methods, you are saying that the conclusions drawn from statistical studies become common sense. I agree but would add that these statistical links are much more reliable than other forms of common sense.

[Kriminal99]

My first attempt is to say that steriotyping is fundamentally bad statistics born out of deliberately ignoring more important factors and also possibly drawing false conclusions as a consequence of cultural ignorance.

 

Exactly. If you touch a glowing red thing eminating heat because you don't want to stereotype it as dangerous, you are just stupid.

 

However if you touch something that is blue (using the excuse that blue things are usually low or moderate temperature) but looks like a flame because you like the color blue, then you are equally as stupid. You ignored that it looked like a flame which is a much better indicator than color, just because you liked the color.

 

If you fail to hire a person of a given race who was a rhodes scholar because people of his race are statistically less hard working (but really because you dislike that race) you are equally stupid because you have artifically reduced your labor pool making it so you will have to charge more to pay your increased labour costs which resulted from the reduced supply.