If you take everyone’s IQ score in the US and divide that number by the number of people in the US with an IQ, will it be 100?
That’s the general idea about it.
Yes.
What does “the number of people with an IQ mean”?
IQ is a norm-referenced number. That means that it will always yield an average score for the group upon which the norms have been developed. If the norming groups are non-representative of the whole population, then the average could be off.
The average has been somewhat arbitrarily set at 100.
Yes, because if it doesn’t the powers that be will re-norm the test to make it so.
Remember, IQ isn’t a real thing. It’s a score on a test where the average is artificially set to be 100. IQ scores cannot be properly compared over time because the test changes and the norm changes. If you asked people of varying ages who took the test at different times, you might not get an average of 100 but that doesn’t matter. It’s like saying that the value of a dollar doesn’t equal one hundred cents because the buying power of the dollar has varied over time.
IQ is not intelligence. It’s a number the collection of which at any given point in time, by definition, has an average of 100. Whether that answers your question I don’t know.
So we’re not counting people with no IQ? Like dead people?
Is all of the above correct however if taking the IQs only of people in the US?
Are the standard tests normalized country-by-country (i.e., IQ 100 in Italy is not the same as 100 IQ in Ireland?) Or is there as worldwide (or at least, I don’t know, OECD?) norm.
One thing my psychology teacher told us is that a person’s IQ is the score he gets on the particular test that measured it. There is no such thing as an absolute IQ.
I mean people who haven’t taken an IQ test. So my IQ is always changing because of other people’s IQ?
As stated above, IQ is completely dependent on the specific test you took and grading algorithm. It’s an artifact of our desire to assign a single number to something that is complex and multi-dimensional.
What I always heard was, “Think how stupid the average person you encounter is, and realize that half the people are even dumber than that!” 
I believe the IQ of 100 is tagged to the median, not the average. I’m curious whether the average is higher or lower than the median. It’d be fun to tell people that most people have lower than average IQ.
Most people in the US have lower than the average income.
A troublingly high number of people don’t understand the difference between average and median.
Nope, 100 is the mean, with a standard deviation of 15.
IQs are designed to be normally distributed, so the median and mean coincide.
No. The norm groups are static for a given test.
But the set of people who have a measured IQ is surely not representative of the entire population. Most IQ tests are administered because there’s some reason to believe that the subject has an extreme IQ (high or low), and the tester wants to know just how extreme. That’s at the very least going to mean that the tested population has a larger standard deviation and kurtosis than the general population (the distribution may even be bimodal), and it’s not hard to imagine that it might also skew the mean.
I also wonder if this might account for the Flynn effect. When IQ tests were first invented, they were mostly used for distinguishing imbeciles from morons, so the average of all those tested would be low. But increasingly, they’re also used to determine whether smart children are gifted, so the test population is now also sampling the high end of the population.
From the wiki link given above
You don’t norm a test by giving it to a biased sample.
I think intelligence is a conserved quantity. Thus, every time somebody posts something on the Internet, his brain loses the amount of intelligence that was placed into the post. (I am less sure if intelligence is also entangled – if so, then the poster would lose a certain amount of intelligence every time a viewer reads the post.)
The rate of intelligence transferred is proportional to the amount of intelligence present in the mind of the poster, and thus decreases exponentially as more intelligence is posted. This may be seen in posts on SDMB, where we have mostly very intelligent posters with a high “intelligence density”, who may write posts with high intelligence content while losing rather little intelligence themselves – versus, say, low-intelligence posters on Yahoo (for example), who transfer only a small amount of low-density intelligence while losing much more of their own in the process.
There’s no centralized data base with scores for everyone who has been administered an IQ measure. Research is done using representative samples, and as pointed out, norms for measures are also established based on representative samples. These are typically much larger than a typical research study.
does this mean that people who score very poorly might be just ss smart as everyone else?