If that's true, it would destroy the notion people aren't buying radios anymore. I linked to only 9 models on just Amazon, and their total sales were 324,000+ per year (27,000+ per month).As to your data on "emergency radios" keep in mind that most of those are also used just to listen to something. Most are, in fact, not sold for emergency preparedness but for working in the garage, while gardening, when at the pool or camping or on a picnic.
Also, if people simply wanted radios for everyday household activities, why are they buying so many "readiness commando" looking emergency models when they could get normal, visually pleasing models instead? For fun, I just searched for the same number (9) of best-selling non-emergency radios on Amazon. The current total sales figure for those? 312,000+ units per year (1 (3K+), 2 (5K+), 3 (2K+), 4 (6K+), 5 (6K+), 6 (1K+), 7 (1K+), 8 (1K+), 9 (1K+)).
The only explanation I can see for this is that the large numbers buying emergency radios are doing so because they actually want them for their emergency features (NOAA and crank/solar charging), whether they plan on using them casually between emergencies or not.
The technical minutia of what happened in Minot involved comms systems little known to the public and virtually every media report presented overly generalized accounts. Had those esoteric details been published, people would have known that Minot was more a case of Murphy's Law scoring a trifecta (and then some) than of catastrophic training failures or apathetic emergency people. The Minot police did in fact attempt an EAS activation, first using equipment whose NVRAM, unbeknownst to them, had become faulty due to power spike damage, and then directly using its county's police band frequency allocated for EAS alert uplink transmissions. That uplink transmission worked, but unfortunately, the LP station designated for monitoring it had it incorrectly configured in its receiver, and never received the uplink transmission itself. Where the police department's training failed outright -- because they were unaware of this backup method existing -- was in not making a second EAS activation attempt using a hardline into NAWAS, which would have let them send their alert through the region's local NWS transmitter to the local LP(s). Instead, they went directly to the final backup method defined in their county EAS plan, making ordinary phone calls to the LP stations. But because the market was fully automated when the train derailed, nobody answered. The police department ultimately found some home phone numbers for station employees, but by then it was too late.The people in Minot had, apparently, not been well enough trained and were simply incapable of activating the EAS. They did not even react by finding someone who could step in and activate the system.
I know no way of finding the official EAS plan documents for Minot in 2002, but you can see typical EAS monitoring plans for a few California counties here, here, and here. Note the frequencies monitored by their LPs with names like SHASCOM, LA County Voice, CLERS, and Control One. Every county has designated frequencies like those upon which local agencies can uplink EAS alerts to monitoring LP station(s). It was one of these type of frequencies that the Minot police used and that failed to be received. Most of the long technical papers on how Minot went down seem to have fallen victim to internet data rot, but you can find a reference to this -- how the police actually did get an alert sent -- by Ctrl-F searhing this PDF for "crystal". (The station's receiver for that radio frequency dated to the EBS days and was crystal-controlled rather than using PLL synthesis. Only it's baseband output would have been routed into the station EAS unit, so its age wouldn't have prompted anyone to replace it. Only its frequency configuration mattered.)