This statement seems wrong:
"The noise amplitude when converted to 1/3 octave bands, is 15 to 20 db below the accepted threshold for hearing in this frequency range. This suggests that the hearers are not extracting the Taos hum stochastically from acoustic noise." The tones would be much less than 1/3 octave wide, and thus should be expected to be heard more easy than the writer claims.

I also found this gem: "Another interesting fact is that some HUM sufferers are either partially or totally deaf." http://homepages.tesco.net/~John.Dawes2/agecurv.htm

And a curve showing a peak at age 55-60 years, with an onset in the late 20's, despite the fact that hearing only worsens with age, at the same link.

This is the conclusion of the most careful study on that same site:
http://homepages.tesco.net/~John.Dawes2/report.htm
"SRL had expected to measure specific low frequency noises at sufferer's property. The lack of measurable real noises and the findings of the audiological research suggests that there may be noise problems with physiologically related causes. One sufferer was very adamant about a noise which could not be found. However, since undergoing a detailed audiological examination they have been diagnosed as suffering from tinnitus. They have now accepted this and are trying to come to terms with it and are to some extent relieved at knowing what the problem was."