Some excerpts follow below.
Yes, this test is old, and PLB models are probably better now, especially since this old test had better results than an even older test . . . but nobody really knows that for sure.
SPOT and inReach often get a bad rep because their capabilities are always being tested, and the relatively few anomalies (on a % basis) generate lots of internet posts.
But for PLBs, there's no readily available way to test them other than inadvertently calling out the cavalry. (Well, except for the ACR 406Link feature.)
http://www.equipped.org/406_GPS_beacon_test_2005_summary.htm
"The few anomalies experienced in this evaluation bear out the reality that even electrically operated emergency signaling devices manufactured and tested to very high quality standards may still be less than 100% reliable in the field and that it remains good practice that users should always self-test beacons prior to embarking upon any excursion or being involved in any situation where they may have to be relied upon in an emergency."
"NOAA and other government agencies involved in the operation of the COSPAS-SARSAT system should investigate the apparent anomalies experienced during these tests that could possibly be attributed to faults in the system."
"There is an obvious and urgent need for government agencies involved in operation and regulation of these beacons and the COSPAS-SARSAT system to develop a more expedient means by which real world testing of these beacons can be conducted with a minimum of bureaucratic hurdles. It should be possible for any legitimate organization representing consumer interests to schedule a test of beacons on relatively short notice. For relatively small numbers of beacons, the use of operationally coded beacons should be facilitated, as the need to use test protocol-coded beacons is a very substantial impediment to the independent testing of these beacons."
"Delay in receiving system performance data (satellite data) is detrimental to the expedient and effective testing of 406 MHz emergency beacons with the potential for devastating data loss and potential for invalidation of testing that, at best, is difficult and expensive to organize. This evaluation experienced just such a loss of irreplaceable data on one test. It should be a priority for the government agencies involved to enable testing organizations to receive immediate automated feedback, perhaps via the Internet, of the system performance during a test."