IQ Tests. Slate, via Sailer has this interesting report on how the military has found that an aptitude test predicts performance of soldiers:

A study by the RAND Corporation, commissioned by the Pentagon and published in 2005, evaluated several factors that affect military performance—experience, training, aptitude, and so forth—and found that aptitude is key. This was true even of basic combat skills, such as shooting straight. Replacing a tank gunner who had scored Category IV with one who'd scored Category IIIA (in the 50th to 64th percentile) improved the chances of hitting a target by 34 percent.

Today's Army, of course, is much more high-tech, from top to bottom. The problem is that when tasks get more technical, aptitude makes an even bigger difference. In one Army study cited by the RAND report, three-man teams from the Army's active-duty signal battalions were told to make a communications system operational. Teams consisting of Category IIIA personnel had a 67 percent chance of succeeding. Teams with Category IIIB soldiers (who had ranked in the 31st to 49th percentile) had a 47 percent chance. Those with Category IVs had only a 29 percent chance. The study also showed that adding a high-scoring soldier to a three-man team increased its chance of success by 8 percent. (This also means that adding a low-scoring soldier to a team reduces its chance by a similar margin.)