Not too much point arguing about whether they're any good or not ...
Here's a Geoffrey Robertson hypothetical - say you're the boss of a big flying organisation picking people off the street to pilot your aircraft, and you're oversupplied with prospective employees. How do you cull them, or more charitably speaking, whittle away the excess so you're left with the best?
Hmmm ... use education as an initial sorting device, then have a first stage of testing and interviews (using what you judge to be the best available psych and aptitude tests available, which (crazily enough, pardon the pun) are likely designed with a lot of input from trained psychs. Medicals too.
Going on the results from the first round, pick the cream of the crop and get them back for further interviews and/or testing. Based on who is judged to do best and how many you need, send out offers.
Sounds eerily familiar, doesn't it?
On the subject of whether the psychomotor tests are any good as predictors, dunno - anyone got any good long term data?