Kicking the program up again last night, it did discover a 4th answer to the self-referential test (2 already found + 1 correct one + 1 new incorrect one). They are listed below in the table.
Working through the new solution, it indeed seems to be a correct answer assuming you don’t know the answer to 20 is.
Ultimately, I will say that I feel a little bit let down because of the arbitrariness of question 20.
20. Standardized test is to intelligence as barometer is to
(A) temperature (only)
(B) wind-velocity (only)
(C) latitude (only)
(D) longitude (only)
(E) temperature, wind-velocity, latitude, and longitude
A barometer is a device to measure atmospheric pressure. So, obviously the author’s intent was to make some claim that standardized tests aren’t a good way to measure intelligence. But, the analogy is malformed for at least two reasons. The first reason it is malformed is because you have one going test going to one attribute intelligence, the barometer measuring a bunch of different attributes would be one reason to prefer the other answers to the author’s chosen answer. Secondly, a relationship between an unrelated item to an unrelated item is meaningless. I mean, why not say standardized tests are to intelligence as a barometer is to black?
It might be sour grapes at this point, but I do feel let down that an arbitrary inside joke on the part of the author ruined an otherwise entertaining exercise. Precisely because it turns out to be essential in determining the “correct” solution, out of what turns out to be quite a few acceptable answers.
For those interested in the C++ source code, it can be found here.