I completely agree with Bruce. First, does anyone really believe that the TSA needs all "72 airlines[' ...] June 2004 domestic passenger flight records" in order to test their system? Assuming the system were verifiable, a sample of real data would be sufficient. Second, how will they know the difference between a successful test and a failed test? Did they catch a set of actual terrorists through other means in June 2004, so that they have a group of expected true positive matches? Seems unlikely, given the publicity that each minor incident has raised.
No, the purpose of the TSA test is not to verify their system -- such verification isn't possible, and if it were, all 72 airlines would not need to participate for a valid test sample. The real test the TSA is running is whether or not the airlines will comply, and how badly the public will freak out.
The airlines have complied -- test 1 passed. So, onto test 2. This would be a great time to write your congresspeople and freak out.