« Satire: "College Profs Denounce Western Culture, Move to Caves" | Gene Expression Front Page | Borg Sperm »
May 17, 2005

IQ Irrelevancies

Since Jensen & Rushton's article a few weeks ago (but really since Day 1 of graduate school), I've heard a lot of arguments against using IQ instruments, some are great (the best were from Jensen himself), some are mediocre (from Robert L. Williams, creator of Ebonics and the BITCH) and some I can only shake my head at, because the arguments they make are irrelevant. Some of my favorites are listed below.....

But I read an article saying X_1 group improved their performance by Z points on Y measure after a week of coaching from [usually someone with access to the assessment instrument].
So what? If they didn't gain any points, that would be the interesting point. The real question is not if X_1 group can increase their score after coaching, but, giving the same coaching to groups X_2 . . X_p is the rate of increase the same across groups1. Moreover, the score(s) at the end of the coaching are not measuring the same thing as the scores before the test, so to compare them is akin to comparing apples and doorknob handles.
Smitty took an IQ test and said s/he got X score (usually X < 80), but I think Smitty is brilliant. S/he always engages in conversation, offers lots of opinions about this, that, the other, and sounds so eloquent when s/he speaks.
So what? When did a given amount of interactions with [the person making the argument] indicate an objective, reliable way to indicate intellectual capacity?
I don't think a 2-hour exam can really give a good measure of someone's [put moral construct here, but it usually is worth].
So what? 1) No scientist would ever advocate that IQ=worth. 2) You're entitled to have whatever opinion you desire to have (whether supported or not supported by data).
IQ tests just measure performance on discrete skills. They don't really (directly) measure intelligence [usually followed by some asinine argument about why either intelligence doesn't exist or that there are so many varieties that it isn't worthwhile just to measure "1 kind"] .
So what? You can't directly measure evolution or gravity either, but few scientifically literate folks would say that they don't exist. You can quantify the constructs' effects and make (quite accurate) predictions about their potency on a given variable.
IQ researchers [in the past] have used it to discriminate against X group and say derogatory things about [the group] [The (often unstated) conclusion I am supposed to draw is that IQ/intelligence/g research should stop].
So what? People have misused fire to set women's garments ablaze, does that mean fire is bad? E-bay has been used to sell teenage boys, does that mean the Internet (or computers, in general) shouldn't be used? The ethics of a given research program and the object under investigation by the research team are two separate issues, not to mention use by folks who have no idea what they are doing with the object.

1. Note. I am not advocating the analysis of change scores.

Posted by A. Beaujean at 10:57 PM