With the schools, it seems to come down to economics; that is, although the best recommended use of OLSAT and CogAT is probably ONLY as a screener to catch ADDITIONAL gifted children (who were not caught by other, additonal screening devices), it is sometimes (and I would argue probably erroneously) used as a CRITERIA for programming. Group tests are definitely cheaper to use than individual IQ tests, but are generally thought to be less accurate:

http://www.hoagiesgifted.org/tests_tell_us.htm