Hi Fade. I know you're intelligent, and we have debated this before, but many of your claims here are ignorant.
Here we go....in order
Fade said:
]Last time this question came up, I stomped it into the ground. It's hard to summarize WHY they are so flawed, because the flaws are multi-levelled.
I'll try to put forth as much information as I can, so I don't need to go back and refute anything point by point by people that simply don't understand what IQ is, and why it is.
First, nobody has ever established an objective metric measuring actual intellect. In fact, few can agree on exactly what constitutes intellect in the first place. Is it our ability to adapt to changing situations? Is it our memory? Is it our application? Is it a combination of all these things? What part of us can be empirically measured to give us our intelligence readout? I don't know. Neither do you.
There are many things we can measure-- reliabily and validly-- even if we do not completely understand the underlying construct being measured.
Use your same argument for "Grade Point Average." We can indeed measure someone's GPA. It's easy.
What does the underlying construct we're measuring get at? That's more difficult to answer. GPA is probably a function of many things: the schools you attend; motivation levels; the types of classes and teachers you take; whether or not you work full time and go to school, and -- heaven forbid-- perhaps your IQ.
No one agrees on just how important these individual compenents are to determining one's GPA. But, to go from that, to saying "therefore we're not trully measuring GPA"
is absurd.
Second example: how accurate were our tests for AIDS (or cancer) before we had a handle on indeed what AIDS (or cancer) was?
Second, the bias is overwhelming. Let me break this down:
1. Any test that calls itself a test will automatically give you a higher fail rate. I am a "good" test taker. I over-perform on every single standardized test I've ever taken. This is because I suffer from exactly 0 testing anxiety.
No test ever made is perfect. All tests contain error (i.e., the part of your total score that is due to factors other then that supposedly being measured by the test).
But. for over 100 years now, we've had objective, proven techniques for measuring just how much error is contained in any test. It's called reliability. 1-reliability = the variance in test scores not due to the construct being measured. In other words, 1-reliability is a measure of the test's error.
Standardized IQ tests have reliabilities in the mid to high .90's. That's about as good as you're going to get for any test in existence (most college exams, for example, have reliabilities barely pushing .6 or .7)
There is a whole branch of statistics showing nearly every racial minority underperforms on the overwhelmingly white created/educated tests out there.
True-- the data go back all the way to WW I. Any g-loaded IQ test (independent of the color of the person who made the test) will show large race differences at the group level.
This is probably the most replicated finding in all of psychology: Mean IQ differs across racial groups.
You're assuming that because the race difference exists, the IQ tests must be biased.
But, if you would look at the scientific literature-- 1000's of studies literally show that IQ tests have neither slope nor intercept bias as predictors for minorities. There is no underprediction.
The prediction accuracy of the test is race blind, though the total scores on the test are not. The US legal system now accepts the FACT that IQ tests are NOT biased against minorities.
Now, why do races differ on IQ tests? That's another big thread we can save for later.
I would challenge any "skeptic" here to produce data / list citations showing that modern IQ tests are either legally or psychometrically biased against ethnic groups.
I bet my IQ world view that the above paragraph will be ignored, as there just are no studies showing the tests to be biased.
2. Use bias. The sheer comfort of using tool, or where that person is, effects the test. For a quick study yourself, try taking an IQ test every few days. Do it when you're tired, when you're just waking up, in the middle of the day. Do it at work, do it in the bath (if you can), do it on the sofa, in various places. You'll see that your level of comfort strongly alters how well you perform.
You are again talking about reliability, in this case, test-retest reliability. Again, the r's are in the .90's
Standardized IQ tests (not the crap you take off the web) won't allow for you to test in the tub, or on your sofa.
The standardization process rules out many of your examples; the being tired or sick points you raise would indeed add to the error of the test, which is about 5% (given reliabilities of .95)
And, I'd submit to you that any professional test administrator would likely reschedule if you came in on a day where you were sick, hungover, whatever-- anything that would invalidate the test.
It's like testing Michael Jordan's basketball skills on a day when he had a sprained ankle. When Mike scores low, who's fault is it-- the test, or the person who gave him the test?
3. Every test fails to eliminate these biases, and is thus suspect. The use of WORDS and NUMBERS will drastically alter the results, because the use of WORDS and NUMBERS is drastically different in different people. I was raised to be highly literary, and I've used numbers mathematically from a young age. Even though I may be exactly as intelligent at a man that was never educated, had to work on a farm his entire life, and has never stared at a text book in his entire life.... he will perform abysmally while I will perform very well.
The IQ test showing the largest race differences is the Raven's Progressive Matrices.
It has no word nor no numbers.
There are also reaction time / speeded cognitive tasks that don't use words or numbers, yet show large group differences, while correlating strongly with traditional paper and pencil IQ tests.
It would not be said to be a test.
The person taking it would not know they were taking it.
There would be no manipulation of shapes.
There would be no manipulation of numbers,
or words,
or ideas,
or anything physical.
There would be no time limit, or interference by the people giving the test.
Any person of any nationality that speaks any language and that has lived at any time would be able to take the test without any preparation at all.
The effects would not change from morning to night,
or day to day,
or year to year (barring injury of course, whether traumatic or slow destruction as per drugs).
Look at the literature-- again. Many of these studies have been done. The results still come out the same**
the exception is taking speed out of IQ testing, as one best guess as to what IQ is, is speed of mental processing. So, controlling for that would be throwing the baby out with the bathwater.
