Chapter 13 
DR. COHEN: Ranking information has only an ordinal property. The 
absolute numbers have no significance, nor do the units of measurements 
between — like 16, 17, 18, 19. In a ranking system, you never assume those 
units are equal. 
In this system, presented to consumers, consumers have a right to 
assume equal appearing units — 16, 17, 18, 19. It goes beyond a ranking 
system. If you just want a ranking system, then there are ways to do that, 
to build on this kind of variance. 
So, if all you want is a ranking system, the one that is in place now 
attempts to do more than that, and I thought your evidence indicated that 
it did not do it with great validity. 
DR. TOWNSEND: What variance are you speaking of? 
DR. COHEN: The interaction of human smoking topography and cigarettes 
having different design features. 
DR. TOWNSEND: Have you quantified that? 
DR. COHEN: 1 have not done it. 1 think we have discussed it today. I am 
not a technical expert on it. 1 am asking from the standpoint of consumer 
usefulness. 
If a ranking system could be preserved to meet your objectives that 
you set out and, at the same time, it would have more validity because it 
wouldn't represent units that do not exist because there is too much variance 
around them; it is only a ranking system. 
DR. DEBETHIZY: It sounds like you have put a proposal on the table, and 
my impression is that is what we will do tomorrow. 
DR. COHEN: It sounded like you were saying, "Well, if it achieves its 
purpose as a ranking system and leaves consumers to know which brands 
are lower and which are higher, and that is fine." 
DR. DEBETHIZY: And we think that the current method does that. So, I will 
be looking forward to the discussion tomorrow about alternative methods. 
DR. HENNINGFIELD: This is actually a nice introduction to my point. 
Compensation is one of the reasons that you get a good correlation with 
machines and a lousy correlation with humans. It is not validated that the 
FTC method predicts what humans get. 
But you have seemed concerned that compensation was not perfect, 
as though the fact that it was not imperfect rejected the notion of 
compensation. 
DR. DEBETHIZY: 1 think what 1 was doing was challenging the notion that 
people get the same amount from every cigarette on the market. 
DR. HENNINGFIELD: Do you know of any drug study with humans with 
addictive drugs where you do get perfect compensation? 
187 
