A peer-reviewed study about Wikipedia's accuracy

[Ramius]

:lol: :lol:

 

Surely, you cannot be serious with this statement??

 

So let me get this straight, if I am playing a board game where there is one die, if someone asks me "what is the average of ONE roll of the die?" I should say: "Isn't it obvious, the result will be a 3.5???"

 

The other person will ask "But I do not see a 3.5 value on the die" and I say "look closer"

 

The next statement from the other person would probably include the words "crazy" or "delusional"

 

if you think thats good, wait til he gets into the part where he claims a die roll of 6 is in "error", and has more "error" than a die roll of 4, even though both outcomes have an equal probability of occurring.

[DC Tom]

if you think thats good, wait til he gets into the part where he claims a die roll of 6 is in "error", and has more "error" than a die roll of 4, even though both outcomes have an equal probability of occurring.

 

...thus causing the roll of 6 to regress toward the mean...a "die roll" being, of course, your attempt to measure the "true average roll" of the die.

 

 

You know, he could have stuck with my dice example, and thus avoided about half the problems with his analogy. Instead he changes it to a "die" and makes it completely !@#$ed.

[Cornerville]

if you think thats good, wait til he gets into the part where he claims a die roll of 6 is in "error", and has more "error" than a die roll of 4, even though both outcomes have an equal probability of occurring.

 

 

So if I am at a craps table in Vegas and roll 2 6's, should i yell "two errors right there!!!"

[X. Benedict]

So if I am at a craps table in Vegas and roll 2 6's, should i yell "two errors right there!!!"

 

Are you on the pass line?

[Ramius]

So if I am at a craps table in Vegas and roll 2 6's, should i yell "two errors right there!!!"

 

screw snake eyes. Bet the house on rolling twin 3.5's.

[Cornerville]

Are you on the pass line?

 

 

Im usually not when it hits!!

[Cornerville]

screw snake eyes. Bet the house on rolling twin 3.5's.

 

 

 

We should make t-shirts with your avatar on it and with the saying underneath...we could make millions!!!

[Orton's Arm]

:lol: :lol:

 

Surely, you cannot be serious with this statement??

 

So let me get this straight, if I am playing a board game where there is one die, if someone asks me "what is the average of ONE roll of the die?" I should say: "Isn't it obvious, the result will be a 3.5???"

 

The other person will ask "But I do not see a 3.5 value on the die" and I say "look closer"

 

The next statement from the other person would probably include the words "crazy" or "delusional"

Welcome to this discussion, Silent Bob. Why you chose this thread is a good question--there are more interesting places you could have chosen.

 

But since you're here anyway, I may as well give you a little background into our debate. Tom, Ramius, and I were arguing about I.Q. test scores. For the purposes of the discussion, I'd defined the term "true score" on an I.Q. test as the average score you'd get if you took the test 1000 times, assuming no learning effect or fatigue effect. You take the test one time to try to measure this true score. You may get lucky and score above your true score, or unlucky and score below it.

 

Tom brought up the concept of dice. I pointed out that if you wanted to make a die roll analogous to the above situation, you'd have to roll a die one time to try to measure the average roll you'd get if you rolled it 1000 times. Unfortunately, neither Tom nor Ramius appear to have understood the analogy. Or more accurately, Tom probably does have at least a rudimentary understanding of the analogy, but makes fun of it anyway. If it's a choice between intellectual honesty on the one hand, and the chance to make fun of me on the other, Tom will go with choice B every time. Ramius doesn't have to deal with that dilemma, because his understanding of these concepts is a lot weaker.

[Ramius]

Welcome to this discussion, Silent Bob. Why you chose this thread is a good question--there are more interesting places you could have chosen.

 

But since you're here anyway, I may as well give you a little background into our debate. Tom, Ramius, and I were arguing about I.Q. test scores. For the purposes of the discussion, I'd defined the term "true score" on an I.Q. test as the average score you'd get if you took the test 1000 times, assuming no learning effect or fatigue effect. You take the test one time to try to measure this true score. You may get lucky and score above your true score, or unlucky and score below it.

 

Tom brought up the concept of dice. I pointed out that if you wanted to make a die roll analogous to the above situation, you'd have to roll a die one time to try to measure the average roll you'd get if you rolled it 1000 times. Unfortunately, neither Tom nor Ramius appear to have understood the analogy. Or more accurately, Tom probably does have at least a rudimentary understanding of the analogy, but makes fun of it anyway. If it's a choice between intellectual honesty on the one hand, and the chance to make fun of me on the other, Tom will go with choice B every time. Ramius doesn't have to deal with that dilemma, because his understanding of these concepts is a lot weaker.

 

and by intellectual honesty, you mean rolling 3.5 on a single die roll? Hows that coming by the way? Molson PM'ed me and said that during your last attempt to roll a 3.5, the die rolled into the fryer, so it'll be a few days until your next trial.

[DC Tom]

Welcome to this discussion, Silent Bob. Why you chose this thread is a good question--there are more interesting places you could have chosen.

 

But since you're here anyway, I may as well give you a little background into our debate. Tom, Ramius, and I were arguing about I.Q. test scores. For the purposes of the discussion, I'd defined the term "true score" on an I.Q. test as the average score you'd get if you took the test 1000 times, assuming no learning effect or fatigue effect. You take the test one time to try to measure this true score. You may get lucky and score above your true score, or unlucky and score below it.

 

Tom brought up the concept of dice. I pointed out that if you wanted to make a die roll analogous to the above situation, you'd have to roll a die one time to try to measure the average roll you'd get if you rolled it 1000 times. Unfortunately, neither Tom nor Ramius appear to have understood the analogy. Or more accurately, Tom probably does have at least a rudimentary understanding of the analogy, but makes fun of it anyway. If it's a choice between intellectual honesty on the one hand, and the chance to make fun of me on the other, Tom will go with choice B every time. Ramius doesn't have to deal with that dilemma, because his understanding of these concepts is a lot weaker.

 

Intellectual honesty like pretending error in a single measure and variance in a population of measures, or subsequent tests of the same thing versus independent tests of different things, or purposely confusing "heritability" and "inheritability", you mean?

 

Admit you don't know what you're talking about, already. Accepting your ignorance is the first step in learning.

[Orton's Arm]

and by intellectual honesty, you mean rolling 3.5 on a single die roll? Hows that coming by the way? Molson PM'ed me and said that during your last attempt to roll a 3.5, the die rolled into the fryer, so it'll be a few days until your next trial.

Don't worry, I'm not accusing you of intellectual dishonesty. You're too stupid for that to be an issue.

[Orton's Arm]

Intellectual honesty like pretending error in a single measure and variance in a population of measures, or subsequent tests of the same thing versus independent tests of different things, or purposely confusing "heritability" and "inheritability", you mean?

 

Admit you don't know what you're talking about, already. Accepting your ignorance is the first step in learning.

I'm not saying that all your criticism of my posts is based on intellectual dishonesty. A lot of it's based on simply misunderstanding what I've been saying. Or the topics we've been debating. Or both.

[Wacka]

Holcomb's Arm,

What is this top 50 university,what is your degree, and what is it in?

 

My guess, the 3.5th top grossing McD and Doctor of Fryology.

[Orton's Arm]

Holcomb's Arm,

What is this top 50 university,what is your degree, and what is it in?

 

My guess, the 3.5th top grossing McD and Doctor of Fryology.

Hey Wacka,

 

This discussion has taken place before. I've categorically refused to announce the name of my alma mater, because Tom, his wife Ramius, and their puppies would immediately begin deriding it. While I'm sure that worse things have happened to my alma mater than that, I see nothing positive (from the school's perspective) that would offset that very small bit of negativity. Nor do I feel that the Toms or Ramiuses of this board have done anything in particular to have earned that particular bit of information.

 

I'd rather be the butt of McDonald's jokes myself, than have my alma mater subjected to the same kind of unthinking, stupid derision to which Tom and Ramius subject everything else.

[DC Tom]

I'm not saying that all your criticism of my posts is based on intellectual dishonesty. A lot of it's based on simply misunderstanding what I've been saying. Or the topics we've been debating. Or both.

 

Again...I understand perfectly what you're saying. And I understand why it's wrong.

[Orton's Arm]

Again...I understand perfectly what you're saying. And I understand why it's wrong.

You think you understand what I'm saying, but you don't. Let me put it to you this way: you've got two separate phenomena at work, both of which often get labeled "regression toward the mean." For clarity, I'll call phenomenon A the test/retest effect, and phenomenon B regression toward mediocrity.

 

I honestly think you've come to understand the test/retest effect. If you select people on the basis of test scores, and if there's an element of chance involved in determining those scores, the people you selected will have scores that overstate their true distance from the mean. The test/retest effect will always be relevant if you're studying the intelligence of smart people's children. Because the parents were selected on the basis of their test scores, those scores overstate their true distance from the population's mean. Even if the children are just as smart as their parents, their test scores will be closer to the population mean.

 

Then there's regression toward mediocrity--children actually being closer to the population's mean than their parents. Insofar as the narrow-sense heritability of a trait is less than 1; you can expect regression toward mediocrity. Unlike the test/retest effect, regression toward mediocrity involves real movement toward the population's mean, and not merely the appearance of such movement.

 

Suppose you notice that parents with high I.Q. scores have children with somewhat less impressive scores. At least some of that movement will be due to the test/retest effect. And you have to account for that effect before attributing the rest of the movement to regression toward mediocrity.