We read with interest the articles by Grant et al. (1 ) and critique by Mortazavi and Doss (2 ) with reply by the original authors (3 ). In their original article, on p. 523, under the subheading “Examination of Threshold,” Grant et al. assert (emphases added):
Further, near the end of the “Summary and Discussion” section (1 ), the authors say (emphasis added):
Grant et al. arbitrarily and implicitly enlist the absence of a nonzero threshold (or equivalently a threshold of zero dose) as their effective null hypothesis, which they find the data do not permit them to “reject”, i.e., their estimates for thresholds derived from the LSS data cannot be distinguished statistically from zero. Then they go on to commit a common logical fallacy. They, in effect, “accept” this null, absence of a (nonzero) threshold, as the valid scientific...