> ...it is just the situation that millions people have them lying around is really a danger.
24 incidents per year [0] (or -assuming only one million sets are out there- one incident per ~41,600 sets out there) doesn't seem like much of a danger. For comparison this [2] leads me to calculate that there are ~141 serious accidental poison exposures per 41,600 children out there. [3]
By all means, levy a serious penalty for the importation, production, or sale of inadequately or improperly labelled SREM magnet sets. But the actual danger of said sets really appears to be overstated by the CPSC.
[0] https://www.federalregister.gov/articles/2014/10/03/2014-233... [1]
[1] Remember that the ALJ didn't find the CPSC's much larger injury estimates to be credible.
[2] http://www.poison.org/poison-statistics-national
[3] Poison control centers report 42.6 exposures per 1000 kids 6 and younger. NPDS reports that 8% of all reported poison exposures resulted in a moderate, or major effect. So, 42.60.0841.6. This isn't the right way to answer the question "How many young kids are seriously injured by poison exposure?", but it's certainly not a bad way to get an estimate.