One of the larger problems with tests that last as long as some of these IQ type tests is what is termed "scale drift". This has happened to the SAT (see their website for numerous studies on the topic) and requires a re-norming every so often. Thus, if the test hadn't been equated properly (a psychometric/statistical link), there's a good chance that comparing these numbers from that long ago would be more or less meaningless. Likely, it is more valid, and less prone to measurement error, to compare people of a more similar and recent population. All this of course depends on which assessment of IQ we're talking about, what methods of equating and norming were used, and a lot of other factors that can lead to additional measurement error.
Please cite your sources on the cultural bias of the IQ. Also, which measure of IQ has bias? There are some more "culture neutral" versions of the quantitative reasoning and other forms of intelligence (non-language dependant and reviewed by multiple language groups for potential bias). See the PISA and TIMSS assessments for a large-scale multi-cultural/multi-lingual assessment.
Please do cite your sources on the Google tests. As a professional in psychometrics, I'd be very curious to know what evidence they provide for the efficacy of their tests. I'm also very curious as to what outcome they are looking for--e.g., most creative, but also compliant employees? Most loyal?
As a psychometrician, I can only say that any mental measurement professional would certainly argue that a single measurement is rarely enough evidence to make a high stakes decision. That is the reason that most college admissions are based on multiple criteria (e.g., weighted grades based on your school, etc.). For hiring decisions, it's probably safe to assume that most employers use certain minimum criteria for a large pool of applicants, but other factors such as experience (past performance), references and personal interview(s) are used to cull that pool further.
See--that's what I was hoping for and what should be modded insightful/informative. That's what I used to expect from/. mods. Thanks for going a bit further.
Why are we still talking about p-values? Effect sizes are truly a much better indication. That would of course, require some understanding of relative effect sizes for this type of analysis, so perhaps I've answered my own question.
When you consistently establish policy that eliminates non-college educated jobs, what do you expect the population to do? We would not have this "problem" if Obama/the corporate kleptocracy had refocused our government-owned to car companies to actual build and employ our skilled US auto-workers to produce: trains, train rails, solar panels, wind-turbines, and nuclear power plant components as opposed to out-sourcing jobs to increase corporate profits and maximize campaign contributions.
At the risk of being modded into oblivion: The whole crime prediction thing often uses predictor variables that correlate highly with educational and economic opportunity. How about we just call it needs based intervention? If we know what risk factors are for kids, e.g., they didn't eat breakfast or have a learning disability, perhaps we could invest more resources into at risk kids (e.g., fully fund after school programs).
How is this considered troll or off-topic? There is some serious pro-MS modding on this article, which is weird. Maybe I'm not on Slashdot? Better check the date... um, not April yet.
Slashdot Mirror
One of the larger problems with tests that last as long as some of these IQ type tests is what is termed "scale drift". This has happened to the SAT (see their website for numerous studies on the topic) and requires a re-norming every so often. Thus, if the test hadn't been equated properly (a psychometric/statistical link), there's a good chance that comparing these numbers from that long ago would be more or less meaningless. Likely, it is more valid, and less prone to measurement error, to compare people of a more similar and recent population. All this of course depends on which assessment of IQ we're talking about, what methods of equating and norming were used, and a lot of other factors that can lead to additional measurement error.
Please cite your sources on the cultural bias of the IQ. Also, which measure of IQ has bias? There are some more "culture neutral" versions of the quantitative reasoning and other forms of intelligence (non-language dependant and reviewed by multiple language groups for potential bias). See the PISA and TIMSS assessments for a large-scale multi-cultural/multi-lingual assessment.
Please do cite your sources on the Google tests. As a professional in psychometrics, I'd be very curious to know what evidence they provide for the efficacy of their tests. I'm also very curious as to what outcome they are looking for--e.g., most creative, but also compliant employees? Most loyal?
If their test looks like this: http://cruftbox.com/blog/archives/001031.html#001031, the psychometric reliability of a 20ish item test is probably not all that high.
As a psychometrician, I can only say that any mental measurement professional would certainly argue that a single measurement is rarely enough evidence to make a high stakes decision. That is the reason that most college admissions are based on multiple criteria (e.g., weighted grades based on your school, etc.). For hiring decisions, it's probably safe to assume that most employers use certain minimum criteria for a large pool of applicants, but other factors such as experience (past performance), references and personal interview(s) are used to cull that pool further.
Seriously horrible troll of an article.
You are so right. The equating is more like A^2+B^2=C^2.
Sure they can do that--for a small "servicing" fee, much like professional scalpers. I see how this is so much better.
See--that's what I was hoping for and what should be modded insightful/informative. That's what I used to expect from /. mods. Thanks for going a bit further.
How about something actually insightful. I may agree, but the mods... oh wait. This is /.
This guy absolutely wins romantic of the year. Imagine the proposal:
Him:Hey woman, do you want to sign a legal document to get a tax break? We might even procreate to garner further tax exemptions.
Her: Um...
It isn't strange at all. It's a long-standing practice within many religions. It's called syncretism. http://en.wikipedia.org/wiki/Syncretism
wait, are you implying that only actresses may not be paid? Whew...
"If you tilt the unit away from your face so it's almost at a 180-degree angle, you can still see the 3-D effect."
How the fsck do you manage to see the screen when you've turned it completely away from your face?
Why are we still talking about p-values? Effect sizes are truly a much better indication. That would of course, require some understanding of relative effect sizes for this type of analysis, so perhaps I've answered my own question.
Not to start a UID fight, but you must be new here. ;-)
Next on tap... energy produced from your schwetty balls.
When you consistently establish policy that eliminates non-college educated jobs, what do you expect the population to do? We would not have this "problem" if Obama/the corporate kleptocracy had refocused our government-owned to car companies to actual build and employ our skilled US auto-workers to produce: trains, train rails, solar panels, wind-turbines, and nuclear power plant components as opposed to out-sourcing jobs to increase corporate profits and maximize campaign contributions.
but it's Apple. hence, shinier and brighter and better.
At the risk of being modded into oblivion: The whole crime prediction thing often uses predictor variables that correlate highly with educational and economic opportunity. How about we just call it needs based intervention? If we know what risk factors are for kids, e.g., they didn't eat breakfast or have a learning disability, perhaps we could invest more resources into at risk kids (e.g., fully fund after school programs).
How is this considered troll or off-topic? There is some serious pro-MS modding on this article, which is weird. Maybe I'm not on Slashdot? Better check the date... um, not April yet.
mod over-rated without a link. Otherwise, this may as well be astro-turf.
Mod this up. I wish more people took the time to understand the depth of the problem.
I can't believe I'm saying this, but is this a play by the government to have the PO replace IPs so that we can have our rights reinstated?
Is it bullying when the "authorities" aren't doing their job?
The face hugger was to stop the inevitable comments about the prime directive from the Trekkies. Aw crap... I started it.