I agree with your assessment about the IQ test. One particular question which irked me:

#17:
Which one of these five is least like the other four?
a)Mule
b)Kangaroo
c)Cow
d)Deer
e)Donkey

That's a fucked up question. A mule is sterile, and the offspring of a horse and a donkey. A kangaroo is a marsupial. Of all those, I believe deer are the only ones shown to be polygamous in behavior (I could be wrong on that one, though).
What is their benchmark for "least like"? Beats me.

And there is cultural bias in the test because some of the "fill-in-blank with the best choice" statements are based on hackneyed expressions one hears in the US and perhaps not other places. I'm not sure I had heard all of them myself.

I give the IQ test developers 100. It's a bit below average, but probably generous.

Yes, the "least like" ones are very annoying. Question i forget what: which is least like blah blah: Coconut, apple, banana, grape, pear.

My first thought was coconut, because it is the only one that's hollow, and that to the best of my knowledge is a single seed. However grapes come in bunches, although bananas come in groups. Bananas are the only one that isn't generally roundish in shape. Grapes are the only ones i know of that come in a seedless variety, although no one ever notices the seeds in bananas anyway. Apples are the only ones i know of that contain cyanide in their seeds, although that might be true of pears and grapes as well.

Or question number seven: John likes 400 but not 300; he likes 100 but not 99; he likes 2500 but not 2400. Which does he like: 900, 1000, 1100, 1200.

At first glance i thought it was that he didn't like things divisible by threes, but there were more than one answer that wasn't divisible by three. Then i discovered that all the numbers he liked when divded by three had a remainder of a third, and only one of the four choices when divided by three had a remainder of a third. Without any other information to go on, this is just as valid an answer as any other. However knowing that it probably wasn't what they had in mind, i kept going. They don't want you to figure out what the pattern is, because there could be many valid patterns. It's not even stated outright that there is a pattern, it's just assumed. They want you to figure out which patter is obvious to the test makers. Isn't it obvious from the way the question is worded?

Each of the numbers he likes is a number squared times 100, which is also only matched by one of the choices. I went with that once since it seemed a little more "obvious" that the previous valid pattern, and i was tired of trying to intuit what the questioners meant.

I agree about the extreme cultural bias as well, however i personally find the lack of clarity issue to be more annoying, just because i've been nailed by it many times before. Test makers don't seem to get the idea that intelligent people will view problems in more than one way, and if they aren't specific, there may be more than one answer possible. There were a couple of other questions that i forget now, that i could have answered completly differently if i'd applied strict logic to the statement, but in which the questioners unstated intent was obvious.

I'm reminded of Zelazny's second Amber series, when Merlin is given a riddle by a Sphinx, and Merlin gives an answer that fits the Sphinx's description, and the Sphinx claims that the answer is wrong. Merlin counters that his answer fits, and if the Sphinx is unspecific enough in his riddles to narrow it down to only one answer possible, then he is obliged to accept any answer that matches his description.