They could call Reich alone.
A quick look through Google Scholar reveals that Reich has done some decent science in his past, but his analysis, thus far, is seriously flawed. In analyzing the word "stop" he reports that the first formant frequency is 10% higher than an adult. A quick read of several articles on formant frequencies versus age reveals that the mean of the first formant frequency of most vowels declines sharply until age 16. After that it levels off but declines slowly in older age. The variability within age groups swamps any difference between 17 and 28 year olds. Of course, this is why he needed to try use the exemplars, but didn't. This does not even address the issue that the extreme emotional state and loudness introduces big changes to the spectral signatures of speech.
I would like to see the graphs he included with his report.
H&H seem to have done the best they could, but they should have never tried to produce numerical confidence levels. It would have been very difficult (impossible?) for them to have determined a sampling distribution for their methodology.