Re: Testing SD2? - Absolutely! / scientific method

Mark parashakti108 at
Tue, 12 Apr 2005 22:21:58 -0000

--"denis bider" wrote:
> > -M: Are you going to use the Condorcet poll sheets as a data input 
to the PageRank ranking? That is how I would do it. Condorcet would see 
the assymetry of the data points as being ranked in order, but is only 
a first-order algorithm.

>D: That's a good idea, we can do that kind of analysis too. However, I 
was thinking of doing the SD2 poll separately from the Condorcet vote, 
and the names people enumerate on the SD2 question can be those of any 
member of Mensa, not just the candidates.

-M: *IF* you want to be *scientific*, the two processes should only 
differ by ONE variable. In this case, that variable would only be the 
algorithm employed to analyze the data. The data should be identical if 
you want to use only one experimental group. Your idea would require a 
second experimental group, which would complicate things. You could 
have a separate section on the data sheet where non-candidates can be 

1. PageRank of candidates    -  Control group
2. Condorcet of candidates   -  Experimental group #1
3. PageRank of candidates and non-candidates - Experimental group #2

-M: This way, both experimental groups differ from the control group by 
one variable.

>D: So that we get a real graph of all of the voters and the relations 
between that, and get the highest-ranking from that. It would be 
interesting to see how many of the resulting highest-ranking are 
actually candidates.

-M: Yee Haww! Fun stuff.

-Mark, Seattle WA USA