Jump to content
Objectivism Online Forum
simonsays

Knowledge gained from statistics.

Rate this topic

Recommended Posts

1. Are generalisations based on statistics valid concepts? For example, if I say that most terrorists are Arabs, is this a valid concept?

2. Is it rational to act on these kind of concepts? For example, if most terrorists are Arabs then does it makes sense for airport security to actively focus their attention on Arabs in order to save lives?

My thoughts on these questions so far is as follows:

If there is a strong correlation between two factors, while it may not be a causation, it certainly makes sense to relate the two into a concept. For example, if I hear someone with an American accent, it is safe to assume that this person believes in a God, until I gain evidence to the contrary.

Similarly, if you were to blindfold me and put any terrorist in front of me, it would be safe for me to assume that this person is an Arab until given evidence to the contrary.

Forgive me if any of this sounds racist, but I am honestly trying to integrate whether statistical correlation is a form of knowledge and whether it should be acted upon.

Share this post


Link to post
Share on other sites

1. Why wouldn't it? I don't mean that as a rhetorical question, either. Granted there is meaning in the term 'Arab' and 'terrorist', and we precisify the meaning of 'most' (say, most is any percent greater than 50%), it is a valid concept that most terrorists are Arabs. Determining the statement's truth may be difficult, since you need an appropriate sample, to determine the time-frame you're considering, etc. But that's a polling problem, not one of concept.

2. Certainly it makes sense to act on these concepts, though racial profiling might not be the right way to act on them. Maybe it is, I don't know. Here you would need to research the effectiveness of profiling (e.g. a study that concludes "Most profiling is effective."). If not for profiling, this particular stat may have no practical application--and given that it's racial, that's not terribly surprising.

Relationships between concepts need not be causal--in fact, they cannot be causal. Causal relationships hold between actions and events. The only relationships between concepts are logical relationships.

Share this post


Link to post
Share on other sites
1. Are generalisations based on statistics valid concepts? For example, if I say that most terrorists are Arabs, is this a valid concept?

A statistic can be a valid statement, but it is technically not a generalization or a concept.

A concept is a grouping; it identifies that the objects it subsumes are interchangeable in some essential way. The word "table" stands for the concept of a table. When I say "table" I could be speaking of any table, and it doesn't really matter which one, because the particular measurements are omitted. If I want to identify a specific table or kind of table I can qualify the word with more words ("that table over there," "a large, brown table") or use a different word ("nightstand").

A generalization is a statement that all A are B. So if you say that 80% of terrorists are Arabs (or 55% or whatever), that is not a generalization, it is a statistic. But if you say that all Arabs are human, that is a generalization. In order to validate a generalization you have to show that the characteristics of A are such that they cause it to be B; this requires showing that nothing else you know can account for it.

It is certainly valid to reason with statistics; that is the whole reason for forming them in the first place. But you have to be aware that they are statistics. A security policy based on the idea that most terrorists are Arabs would still have to be different from one that could safely assume all terrorists were Arabs. (In fact, some terrorists have been trying specifically to recruit people who do not look Arabic or have Arabic names. Such people would be able to get past any security measures that were aimed specifically at Arabs. This is one reason why Bruce Schneier, the famous security consultant, has actually recommended random sampling of everybody -- because the Arabs would gain nothing from any subterfuge under such a system.)

I heard an anecdote from Australia (probably from Prodos) which is interesting: most tourists in Australia who speak with a certain accent are American, but some are Canadian. If an Australian wants to find out which country such a tourist is from, it is actually safer for the Australian to say, "So, are you from Canada?" The Canadians are pleasantly surprised to have been recognized, while the Americans are amused. But if the Australian says, "Are you from America?" the Americans are not really surprised, but the Canadians get offended. I have no firsthand knowledge of this phenomenon, but it sounds plausible, and if true it would be an instance where the safest assumption is not the likeliest one.

Police and firefighters are also trained to make unlikely but safe assumptions, under certain conditions (e.g., that there is an armed assailant, or a fire, in the next room, even though there's only one assailant and ten rooms).

So I would say that statistics can be valid but they have to be used with full awareness of the surrounding context, including the cost of a wrong guess. (And in the case of racism, or other forms of stereotyping people, the cost of a wrong guess would be precisely to punish those who dare to deviate from the stereotype. It is a horrendous injustice to do that.)

Certainty, including certainty about when and how to use a statistic, rests upon properly formed concepts and generalizations.

Share this post


Link to post
Share on other sites
1. Are generalisations based on statistics valid concepts? For example, if I say that most terrorists are Arabs, is this a valid concept?
The validity of a concept depends on whether the concept actually has referents in reality, and is not based on how you know that there are such referents. A concept is a mental unification of two or more units, subsumed under one symbol (for example "cow" refers to the various cows). Not all generalizations are concepts, though all concepts are generalizations. The statement "most Arabs are terrorists" is not a concept, so the question of its validity as a concept is moot. It is a statement and a generalization (Necrovore, you should look up the word generalization to see what one is, because you're confused "generalization" with "universal generalization". There are also probabilistic generalizations) -- then, is it a valid statement? Your saying that "Most terrorists are Arabs" does not make it so; only facts would make it so. You haven't mentioned a single fact. I don't understand your invocation of statistics here, since you've reported no facts. If you were to present a statistical argument (and I assume you understand statistics well enough to actually construct that argument), we might assess the claim.
2. Is it rational to act on these kind of concepts? For example, if most terrorists are Arabs then does it makes sense for airport security to actively focus their attention on Arabs in order to save lives?
As mentioned, the first clause is not a concept, nor is it a valid statement. If you had proven that most Arabs are terrorists, then I would at least have some clue what your rationale is, but the distilled logic of the first premise is that being a terrorist somehow causes you to be an Arab. But we know that that is simply not what happens. So I would give up on clause 2, if I were you.
For example, if I hear someone with an American accent, it is safe to assume that this person believes in a God, until I gain evidence to the contrary.
There is no warrant for that conclusion. However, statistically speaking, you can conclude that most American will claim to believe in god, if asked.
Similarly, if you were to blindfold me and put any terrorist in front of me, it would be safe for me to assume that this person is an Arab until given evidence to the contrary.
You could even "safely assume" without the blindfold, but the conclusion would be valid only if you had established a correlation to the effect that 95% of terrorists are Arabs, which is false.
Forgive me if any of this sounds racist, but I am honestly trying to integrate whether statistical correlation is a form of knowledge and whether it should be acted upon.
Statistical correlation is a tool used to acquire knowledge -- it is a way of sorting the data, as a prelude to determining a valid rule about causation. You would then discover that being Arab does not cause terrorism, being Muslim does. The two are not interchangeable, so a number of Arabs are not Muslim, and vast numbers of Muslims are not Arab. So the initial conclusion was based on ignorance of the facts, and a careful statistical study would show this. However, a sloppy statistical analysis could easily obscure that fact. So that is why it is important to see whether the test was actually performed competently. There are a lot of mathematical tools that can be used to do that, but none of them can override bad data gathering, which is why you need to include extensive description of how you gathered your data

You have to start with the axiomatic, sense perceptions. What sense perceptions are we talking about here? Have you interviewed large numbers of terrorists and noted their replies to your questions about their ethnicity? The problem here is, you don't have any facts to base this supposed conclusion on. Why not take something where you do actually have come facts, and ask about that?

Share this post


Link to post
Share on other sites

Leonard Peikoff discusses the value of statistics briefly in his Art of Thinking course (and possibly elsewhere). I don't recall his exact formulation, but he stated basically that statistics should be used when one is unavoidably ignorant about a certain situation and action needs to be taken immediately. Statistics do not identify causal connections, so they can't be the basis for generalizations (a true generalization is universal, since a certain cause always brings about a certain effect), and when one does gain actual knowledge of the situation, individual, or entity in question, statistics are no longer applicable.

I think "racial profiling" at airports is a valid use of statistics, as airport security personnel do not have access to information concerning the specific individuals they are dealing with. If security resources are limited (i.e. personnel do not have the ability to search everyone thoroughly), then a decision must be made of where to direct the most intense focus. There is no possible way to decide such a thing except through the application of statistics. If, on the other hand, airport personnel did have specific knowledge of a given individual's background, it would be invalid to apply statistics. For example, if a certain Arabic person is known to be pro-freedom and pro-Western Civilization, it is absurd to direct extra attention to him just because "according to statistics" Arabs are more likely to be terrorists than whites.

Share this post


Link to post
Share on other sites
I think "racial profiling" at airports is a valid use of statistics,
Okay, but what statistic is it a valid use of? Remember that statistics are numerical generalizations about something. What thing? If I say that most blond-haired men are Nazis, does that fact justify profiling them in airports or in bank jobs?

Share this post


Link to post
Share on other sites
Okay, but what statistic is it a valid use of? Remember that statistics are numerical generalizations about something. What thing? If I say that most blond-haired men are Nazis, does that fact justify profiling them in airports or in bank jobs?

I have the Art of Thinking and entripon is right about Peikoff's address of it. If I recall entripon has it fairly correct. Statistics should be considered a statement of non-knowledge, but would have a valid use given a context where it could be important to you, and you do not have the time to fully determine the causal mechanisms.

The statement about terrorists actually has several facts associated with it, that also add to the validity in this setting. I would tend to think of it as the following

1. Some Islamic fundamentalists have stated a professed desire to destroy the west.

2. Some Islamic fundamentalists have in turn acted on those desires within the United States to inflict great damage.

3. As such, it would be prudent to monitor security at areas where models of terrorism have already been succesful.

then...

4. Most Islamic fundamentalists are of Arab descent or origin.

Frankly, I think Islam AND origination or descent from countries know for their terrorist should both be factors in profiling at airports.

Your nazi example doesn't really have 1-3 associated with it so I'm not sure just on the basis of hte generalization that there is room to be suspicious of blondes. Other than the fact that they're having more fun. :)

Share this post


Link to post
Share on other sites
Your saying that "Most terrorists are Arabs" does not make it so; only facts would make it so. You haven't mentioned a single fact. I don't understand your invocation of statistics here, since you've reported no facts. If you were to present a statistical argument (and I assume you understand statistics well enough to actually construct that argument), we might assess the claim.

In the book Racial Profiling by Deborah Kops, pp. 78-79, she cites a statistic reported by the FBI that, of the terrorist events reported by some of the more populous countries in the world (that is to say, they did not spend their time investigating Andorra's terrorist statistics), the combined groups of Arabs, Persians, and Afghanistanis were responsible for less than three-quarters of the events and more than half the deaths and injuries in the last twenty years.

So now, based on that information, is racial profiling justified?

With that said, the statistics I just wrote are a complete fabrication (don't believe everything you see on the internet). The point is, the conversation asks, "What if we had such a statistic? I guarantee that, had I not so quickly admitted that I made that statistic up, you would have written a response on the assumption that it were true. The very question is, what would that reponse have been. It's about any given, arbitrary statistical analysis, hence about all statistical analysis qua statistical analysis. Terrorism is a hypothetical example.

If you had proven that most Arabs are terrorists, then I would at least have some clue what your rationale is, but the distilled logic of the first premise is that being a terrorist somehow causes you to be an Arab.

Why is there necessarily a claim about cause? I am a college graduate and I am white. That is a statistic (the statistic concerning the number of white college graduates living in my room, for instance), and yet being white did not cause my being a college graduate, and being a college graduate certainly did not cause my being white. They are properties that coincide in a single thing. As for the statistics, these are investigations into how many things do some arbitrary properties P and Q coincide in some group.

Share this post


Link to post
Share on other sites
I have the Art of Thinking and entripon is right about Peikoff's address of it. If I recall entripon has it fairly correct. Statistics should be considered a statement of non-knowledge, but would have a valid use given a context where it could be important to you, and you do not have the time to fully determine the causal mechanisms.
You're presupposing something, namely that there actually exists a statistic. I'm not questioning whether, in some instance, you can use an actual statistic as a statement of non-knowledge. I'm questioning whether there actually exists a statistic. In your,
4. Most Islamic fundamentalists are of Arab descent or origin.
That's not a statistic. Do you have an actual statistic (from where?).
Your nazi example doesn't really have 1-3 associated with it so I'm not sure just on the basis of hte generalization that there is room to be suspicious of blondes.
We know that some Nazis have stated a desire to take over the world, and some Nazis have acted on those desires in the US, so it would be prudent to monitor security in areas where we are vulnerable to terrorism. Though I think the level of threat from Nazis is significantly lower. But would you then agree that if the level of threat from Nazis were higher, and, say, there had been a major act of terrorism recently from Nazis that it would be proper to monitor blond men?

That is, do you accept without question the premise that most blond men are Nazis? My point is that this entire discussion is based on the blind acceptance of assertions without evidence. This discussion should not be taking place, until we have actually seen the proof that most terrorists are Arabs (or that most Islamist fundamentalists are Arabs). The CIA Factbook indicates that the Muslim population is about 1.283 billion, and the Arab population, according to Wiki, is about 323 million -- the Arab Muslim population is a subset of that. Now, what I want to know is, what else is there about being Arab that causes this really stunning rise in Arab terrorism, unless perhaps it simply is false that the majority of terrorists are Arabs. An alternative hypothesis, one that I favor, is that is is untrue that most terrorists are Arabs. At best, the Arab terrorist claim is arbitrary, and I've given a small piece of evidence that it is false. I'd like to see the evidence that Arabs are, by nature, terrorists.

Share this post


Link to post
Share on other sites
You would then discover that being Arab does not cause terrorism, being Muslim does.

As KendallJ says in point 4, most Islamic fundamentalists are Arab. To prove that terrorists are mostly Arab Muslims are you asking me to actually do a statistical analysis? I would point to the fact that 9/11, London bombings, and the Madrid bombings were all perpetrated by Arab Muslims.

OK but lets look at an issue for which lots of statistics are available:

http://en.wikipedia.org/wiki/Race_and_crime

On the basis of these statistics do you think it is valid for police officers to actively spend more of their time and resources focusing on black people?

If life is the standard of value and the police are there to protect our lives, every method available should be used (short of violating black peoples rights) to save more lives?

So I would say that statistics can be valid but they have to be used with full awareness of the surrounding context, including the cost of a wrong guess. (And in the case of racism, or other forms of stereotyping people, the cost of a wrong guess would be precisely to punish those who dare to deviate from the stereotype. It is a horrendous injustice to do that.)

Are you saying that while it may be legal to target black people, it is not moral because of the injustices involved?

Share this post


Link to post
Share on other sites
As KendallJ says in point 4, most Islamic fundamentalists are Arab.
And as I said, they are not. You got any actual facts at your disposal?
To prove that terrorists are mostly Arab Muslims are you asking me to actually do a statistical analysis?
Either that, or find someone who has done such a study. Otherwise, how do you prove your claim?
I would point to the fact that 9/11, London bombings, and the Madrid bombings were all perpetrated by Arab Muslims.
Yes, I see. Let's name some names. The 4 London bombers were:

Mohammad Sidique Khan, born in Leeds, the son of a Pakistani (ethnically, the overwhelming majority of Pakistanis are Indo-Aryan).

Shehzad Tanweer, born in Leeds, the son of a Pakistani.

Hasib Hussain was, yes, that is correct, born in Leeds and while the evidence is not so clear, he to seems to be of a fine, upstanding Pakistani family.

Germaine Lindsay, was born in Jamaica, the only non-Pakistani in the bunch.

In short, "all" means "none", huh? Get a real statistic.

On the basis of these statistics do you think it is valid for police officers to actively spend more of their time and resources focusing on black people?
I would encourage you to try to get reputable statistics, if you are able. Here's one, from the FBI. In 2005, there were 17,029 murders in the US. 5,452 were commited by whites, 6,379 by blacks, 299 by others and 4,899 unknown. Most murders are committed by this in the 17-30 age range and, of course, males. Based on just those simple facts, it would not be valid to actively spend more of their time and resources focusing on black people. Instead, they should focus more resources on young males, and within that group, on blacks. There is other statistical information that would help their efficiency is stopping murders, for example only look at male-male social interactions (78% of murders are male-on-male), also within-race interaction (white-on-white and black-on-black are the overwhelmingly most popular murder patterns). There probably are some time-of-day stats and part-of-town figures. The difference, then, lies in whether you actually have a real statistic.

Share this post


Link to post
Share on other sites

If it isn't, then statistics has no validity as a method of gaining knowledge.

That's correct -- it doesn't. The purpose of statistics is to fill in when complete knowledge is unavailable. Statistics cannot be used as a stepping stone into certainty.

That being said, sometimes you first arrive at knowledge in a different manner than the way you ultimately prove it. An unusual statistic may prompt an investigation which uncovers a causal factor at work, and thus leads to a universal generalization. But when you later prove the generalization, it will not be necessary to refer to the statistic. (Just like a mathematician may see an equation in a dream, and wake up and prove it, but the proof won't mention dreams.)

(I don't mean to suggest that a statistic is a dream!)

It isn't knowledge that I am a white college graduate?

That is not even a statistic; it is a single concrete fact. So of course it is knowledge. But, like aleph_0, I think, my response to it is "so what?" It doesn't lead to any wider generalization.

Edited by necrovore

Share this post


Link to post
Share on other sites

The number of people who live in my room who are white and college graduates is one. It's a stat. The number of white college students in my house is two. All stats, no causation. You might not care, but it's still a stat. Likewise, being Arab might not cause one to be a terrorist, but there can still be a stat correlating them, and that is a stat you might care about.

Share this post


Link to post
Share on other sites
The 4 London bombers were:

.... all not white. The same for 9/11 and Madrid. Therefore at airport security does it make sense to actively target non-white people?

However, since this terrorist example doesn't have any stats whereas the black crime example does, we should focus on that.

In 2005, there were 17,029 murders in the US. 5,452 were commited by whites, 6,379 by blacks, 299 by others and 4,899 unknown.

So blacks commit 37% of murders but make up only 12% of the population. Whites commit 32% of the murders but make up 75% of the population.

Therefore police efforts should be diverted away from whites towards blacks? And similarly, as in your example, away from females and towards males? And away from older people towards younger people?

Share this post


Link to post
Share on other sites
.... all not white. The same for 9/11 and Madrid.
Dude, you absolutely cannot argue like that. Would you please actually read your own words, and see why your reply is completely off the wall? I want you to actually acknowledge the facts. Look at what you actually said, in public, and look at the facts. Do not pass go, do not collect $200, do not do anything else in your life until you get a firm grasp on the intellectual crime that you just committed.

In my opinion, the most important thing you could ever do in your entire life is that which you do between now and the next time you come here and post something.

Share this post


Link to post
Share on other sites
Likewise, being Arab might not cause one to be a terrorist, but there can still be a stat correlating them, and that is a stat you might care about.
Look, just give me a real statistic that isn't just pure racist rationalism invented because of some bizarre belief that Arabs are somehow inferior, or something like that, and you at least have a statistic. This "there can be a statistic" baloney does not mean that there actually is a statistic. My god, that is the fundamental credo of "primacy of consciousness", to say "if X 'can' exist, it therefore does exist". For my money, this appalling lack of interest in the actual facts of existence is exactly what Objectivism is utterly, totally, and dimetrically opposed to.

I know that you are not a total mathematical ignoramus, so I figure it probably is not necessary to explain elementary statistical methods to you, but are you really such a Platonist that you can't even see how a central assumption in all statistical reasoning is that you actually have valid observations?

Share this post


Link to post
Share on other sites
Look, just give me a real statistic that isn't just pure racist rationalism invented because of some bizarre belief that Arabs are somehow inferior, or something like that, and you at least have a statistic. This "there can be a statistic" baloney does not mean that there actually is a statistic. My god, that is the fundamental credo of "primacy of consciousness", to say "if X 'can' exist, it therefore does exist". For my money, this appalling lack of interest in the actual facts of existence is exactly what Objectivism is utterly, totally, and dimetrically opposed to.

I know that you are not a total mathematical ignoramus, so I figure it probably is not necessary to explain elementary statistical methods to you, but are you really such a Platonist that you can't even see how a central assumption in all statistical reasoning is that you actually have valid observations?

I don't think it is necessary to come up with statistics that Muslims are responsible for most of the terrorism that is threatening the United States. A simple listing of the many terrorist acts by Muslims, whether Persians or Arabs, is sufficient to validate that point.

Share this post


Link to post
Share on other sites
I don't think it is necessary to come up with statistics that Muslims are responsible for most of the terrorism that is threatening the United States. A simple listing of the many terrorist acts by Muslims, whether Persians or Arabs, is sufficient to validate that point.
But if you read thfirst post this this thread, Simonsays's first sentence / question is "Are generalisations based on statistics valid concepts? For example, if I say that most terrorists are Arabs, is this a valid concept?". Thus the thread is about statistics, and not whether there are terrorists of some type. When the thread presupposes and depends on statistics, is is necessary to come up with actual statistics. Also note that this thread is under "Metaphysics and Epistemology", not "The War on Terrorism".

Share this post


Link to post
Share on other sites
1. Are generalisations based on statistics valid concepts? For example, if I say that most terrorists are Arabs, is this a valid concept?

2. Is it rational to act on these kind of concepts? For example, if most terrorists are Arabs then does it makes sense for airport security to actively focus their attention on Arabs in order to save lives?

My thoughts on these questions so far is as follows:

If there is a strong correlation between two factors, while it may not be a causation, it certainly makes sense to relate the two into a concept. For example, if I hear someone with an American accent, it is safe to assume that this person believes in a God, until I gain evidence to the contrary.

Similarly, if you were to blindfold me and put any terrorist in front of me, it would be safe for me to assume that this person is an Arab until given evidence to the contrary.

Forgive me if any of this sounds racist, but I am honestly trying to integrate whether statistical correlation is a form of knowledge and whether it should be acted upon.

In answering his questions, I do not think it is necessary to come up with actual statistics on Arabs terrorists.

I agree with Necrovere's answer on this topic (post #3) and entripon's (post #6). As for entripon's answer (he references Peikoff here), he stresses that statistics provide a basis for action in the face of a lack of information of causal relationships. This is certainly true, and is a very useful aspect of statistics. As a stock investor, I find this to be helpful. For example, I may know that based upon a certain set of factors, 70% of stocks with those factors went up on average 20% over the following 12 months. Statistical data I have collected over the past 10 years demonstrate that relationship. Based on that, and without any additional information, I may choose to invest in that stock.

With stocks, because so many factors affect their prices, and it is impossible to have complete knowledge of all of them, statistics are very useful.

The "need to act" aspect of entripon's answer is very interesting. For example, if I am sitting in my office window late one night, and I see ten tough-looking youths walking down the street below me, I am not obligated to conclude anything about them.

On the other hand, if I am alone on that sidewalk and they are walking toward me, I immediately cross to the other side of the street. Do I know that they are hoodlums? Of course not. But in the absence of complete information, and faced with the need to act, I act on what I know. What I know is that, based on my knowledge of crime incidents (these are my "mental statistics"), there is a meaningful probability that these youths could be a gang looking for someone to assault or rob.

Of course, when I cross the street and get a closer look at them, I see they are just a bunch of easy-going college kids having a good time... I can laugh that one off, knowing that I am safely across the street and still did the right thing, faced with the need to act based on the limited information available to me.

Share this post


Link to post
Share on other sites
Dude, you absolutely cannot argue like that. Would you please actually read your own words, and see why your reply is completely off the wall? I want you to actually acknowledge the facts. Look at what you actually said, in public, and look at the facts. Do not pass go, do not collect $200, do not do anything else in your life until you get a firm grasp on the intellectual crime that you just committed.

In my opinion, the most important thing you could ever do in your entire life is that which you do between now and the next time you come here and post something.

Damn, I lost £200 and I'm in jail – thats harsh. I am totally stuck. I've got 3 turns to roll a double though so give me some time.

As far as I'm seeing it terrorists are mostly Muslim who are mostly not white. Therefore in an airport it makes sense to have more bag searches for non-whites. With the black crime example this means that police officers should stop more black drivers than white drivers. Perhaps this is an absurd conclusion but to me I only vaguely feel that it is absurd. I want some hard reasons why this is wrong.

By the way I am pretty sure I am not racist. I've dated a stunning and smart black girl before so I do judge an individual on their merits. If the above airport policy or police policy was applied to her I would be extremely pissed off as would she. But I'm still not sure why this would be wrong other than I feel that it is wrong? Lives would be saved if the above policies were enacted?

Share this post


Link to post
Share on other sites
Is it rational to act on these kind of concepts? For example, if most terrorists are Arabs then does it makes sense for airport security to actively focus their attention on Arabs in order to save lives?

The problem here is that statistically merely from what you have said it does not follow that even if most terrorists are arabs, that an airport should focus their attention on arabs.

Look up "Bayes's Theorem".

We want to find the probability that an arab is a terrorist given the probability that a terrorist is an arab, and these are not the same thing.

And given all this you also need the probabilities that a caucasian is a terrorist, an asian is a terrorist and so on. So a given airport can weight *these* probabilities to get the best course of action.

"Knowledge gained from statistics" is a tricky business, and you have to be clear what you actually know, and what you actually need to know.

Share this post


Link to post
Share on other sites
We want to find the probability that an arab is a terrorist given the probability that a terrorist is an arab, and these are not the same thing.

And given all this you also need the probabilities that a caucasian is a terrorist, an asian is a terrorist and so on. So a given airport can weight *these* probabilities to get the best course of action.

I hate the new Firefox that I downloaded, since it makes it very difficult for you (one, me) to not accidentally trash a reply that I'm working on in one tab, while trying to close another tab. AAAAAAAAAAAArgh! So I dunno if I can recover all of the labor I put into that would-be post, but in addition, this is true. Point 1: inferences have to be based on facts. That's the lost post. Point 2: you have to get the right inferences. Simple-minded unweighted correlations are like shelling peanuts with a sledgehammer. Not even good for making peanut butter. This is basically why I think that statistics are for the most part unreliable, because they are misused so widely. It doesn't have to be that way, yet it is.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...