At the same time every individual school will have dozens of people with graduate degrees. Much of those graduate degrees will be based on classes and research about how to acquire exactly this type of data. And yet, they are unable to make these things themselves and go out and purchase them in the most expensive way possible.
The amount of waste involved in the education process is mind boggling.
I was trying very carefully not to say that there's no value in these surveys, because they are designed by experts: PhDs, running statistically significant experiments. Sample sizes on the ones I've looked at are in the hundreds, which is pretty huge for a psych experiment. Just slapping one together based on a few classes you took would be irresponsible and probably not very useful as a diagnostic. My complaint is mostly that since we're publicly funding these experiments anyways (one government body buys them from a company, who gives some money to another government body), we should cut out the middle-man.
It's an interesting problem, because the companies involved fund studies to improve the diagnostic power of the test (tweaking the questions and scoring), but then lock up what they find out.
(Here's an example of a commercial psychometric survey: http://www.talentlens.co.uk/select/giotto)