# Thread: MBTI Type and I.Q.

1. Originally Posted by Evan
I'm not sure I'm following your logic. Are you saying that people with 50 IQ are outliers and only negligibly affect the median?

All I was trying to point out is that it isn't necessarily true that 50% are above average and 50% are below average.
You're missing an underlying assumption: IQ tests are made such that the distribution of the scores is Gaussian (or Normal), thus symmetric, thus 50% above average and 50% below average.

2. Originally Posted by FDG
You're missing an underlying assumption: IQ tests are made such that the distribution of the scores is Gaussian (or Normal), thus symmetric, thus 50% above average and 50% below average.
Well, I don't think they are made to be normally distributed. But found to approximate normal distributions. Minor detail...

3. Originally Posted by Evan
Depends what you mean by 'think'. If you mean "conscious true/false judgments", this is true. But that's pretty narrow...

"Thinking" in the common sense of the word is probably more like Intuition in MBTI than Thinking.
hahaha

4. Originally Posted by Evan
Don't mean to be a dick, but this is not technically true. Just because 100 is average doesn't mean 50% are above and 50% are below.

The average of the sequence {110, 110, 110, 110, 110, 50} is 100 but 83% of people are above average.
Yes, not necessarily true. But IQ falls along a bell curve as well that is perfectly symmetrical along the 100 axis.
For example: 96% of the population is between 70 and 130. 48% is between 70 and 100. 48% is between 100 and 130. 2% is below 70. 2% is above 130.
(Again, if this appears too low, it is likely because this 2% above 130 invades these kinds of websites, and the 2% below 70 rarely communicate online.)

5. Originally Posted by FDG
You're missing an underlying assumption: IQ tests are made such that the distribution of the scores is Gaussian (or Normal), thus symmetric, thus 50% above average and 50% below average.
^ Umm, yeah. That's what I was trying to say

6. Originally Posted by Modern Nomad
Well, I don't think they are made to be normally distributed. But found to approximate normal distributions. Minor detail...
I actually think (I'm pretty sure about this) that the scoring system is set up such that the scores will be normally distributed. Of course, the system is based on past experiences, thus strictly speaking it could be completely wrong for tomorrow's reality; nonetheless, that's the way it's constructed.

7. Originally Posted by FDG
I actually think (I'm pretty sure about this) that the scoring system is set up such that the scores will be normally distributed. Of course, the system is based on past experiences, thus strictly speaking it could be completely wrong for tomorrow's reality; nonetheless, that's the way it's constructed.
The Stanford Binet test (the major IQ test that is professionally administered) gives you a score where they match your "mental age" vs. your "real age". It is usually considered to be most accurate when you take it at age 8ish.

So if you take it at age 8, and you score what the "average" 12 year old gets. Then your IQ is 150. (12/8 * 100). there is nothing inherent in this scoring system that will make the distribution normal. Scores are based on nominal performance. Not adjusted.

For instance, if half of 12 year olds think like the average 9 year old, and half thinks like the average 15 year old, then half IQ's would be 75, half would be 125, for an average of 100 within that population for a barbell type distribution. Also, it is entirely possible that 12 year olds are normally distributed, 11 year olds are right tailed, and 13 year olds are left tailed, 15 year olds have kurtosis. But overall together, they are normally distributed for the overall population which is what the Stanford Binet results indicates.

8. Originally Posted by FDG
You're missing an underlying assumption: IQ tests are made such that the distribution of the scores is Gaussian (or Normal), thus symmetric, thus 50% above average and 50% below average.
How can you make a test such that the results are exactly Gaussian? Maybe a Gaussian is a really good approximation...

I dunno, I guess I'm getting a little nitpicky; I was just pointing out that there's a difference between median and mean.

9. Originally Posted by Modern Nomad
The Stanford Binet test (the major IQ test that is professionally administered) gives you a score where they match your "mental age" vs. your "real age". It is usually considered to be most accurate when you take it at age 8ish.

So if you take it at age 8, and you score what the "average" 12 year old gets. Then your IQ is 150. (12/8 * 100). there is nothing inherent in this scoring system that will make the distribution normal. Scores are based on nominal performance. Not adjusted.

For instance, if half of 12 year olds think like the average 9 year old, and half thinks like the average 15 year old, then half IQ's would be 75, half would be 125, for an average of 100 within that population for a barbell type distribution. Also, it is entirely possible that 12 year olds are normally distributed, 11 year olds are right tailed, and 13 year olds are left tailed, 15 year olds have kurtosis. But overall together, they are normally distributed for the overall population which is what the Stanford Binet results indicates.
That method has been gone for quite a while now.

10. Originally Posted by laughingebony

Thus, Binet developed the concept of mental age (MA), which is an individual's level of mental development relative to others[1].
After taking a standardized test, an individual's mental age is divided by his chronological age and multiplied by 100, yielding an intelligence quotient (IQ). Thus, a subject whose mental and chronological ages are identical has an IQ of 100, or average intelligence.
Mental age - Wikipedia, the free encyclopedia

You really didn't know what it was referring to right? Are you trying to say because the Army uses letter grading, that means civilian population use of IQ scores on a numerical basis (mental age) no longer has any statistical significance? Please clarify.

Or did you just not know? I need to know which one of the two it is.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
Single Sign On provided by vBSSO