Commentary of Eugene Laska, Statistician

The proof offered is that different investigators readily replicate estimates of relative potencies. This is proof that relative potency can be measured with high replicability. Missing from the statement are absolutes, i.e. does a response in level of pain of moderate in one pain type mean the same thing in another pain type? The question of whether it can be measured "scientifically" seems to be unanswered. the argument on consistent estimates of relative potency speaks to reliability but says nothing about validity.