The Devil's Excrement

Report from the TDE’s roving reporter in Palo Alto for the Stanford
Seminar by Prof. Taylor.

The Seminar by Prof. Taylor began with a brief story of how Prof.
Taylor was handed down the problem by a senior member of the
Statistics Department at Stanford. Taylor began by giving a

<!–
D(["mb","summary of the political situation in Venezuela. Here are some points
mentioned by Taylor in his summary:

* Chavez has been a popular figure among Venezuelans, but has not been so
 popular for the US government probably because of Chavez\’ strong ties
 with Castro. The relationship between Chavez and Castro has been beneficial
 for Venezuelans as many Cuban doctors and teachers are now working in
 Venezuela.
* The opposition is mainly composed by businessmen and members of the
 political class that used to rule the country before Chavez. The private
 media (TV & newspapers) are controlled by the opposition and so, it is very
 hard to get objective information about what\’s happening in Venezuela.

Prof. Taylor then described studies he has finished recently using the
models he has on his web page (he mainly referred to the last three). He
revisited the hypothesis testing problem using FDR (False Discovery Rate)
did not find any weird behaviour in the referendum data under the models
he assumed. However, he pointed out that it would be possible to come up with
relatively simple mechanisms of fraud that would be basically impossible to
detect with these models. He then presented some new results about Benford\’s
Law. He said that there is no reason to believe that Benford\’s Law would
“,1]
);

//–>

summary of the political situation in Venezuela. Here are some points
mentioned by Taylor in his summary:

* Chavez has been a popular figure among Venezuelans, but has not been so
 popular for the US government probably because of Chavez’ strong ties
 with Castro. The relationship between Chavez and Castro has been beneficial
 for Venezuelans as many Cuban doctors and teachers are now working in
 Venezuela.
* The opposition is mainly composed by businessmen and members of the
 political class that used to rule the country before Chavez. The private
 media (TV & newspapers) are controlled by the opposition and so, it is very
 hard to get objective information about what’s happening in Venezuela.

Prof. Taylor then described studies he has finished recently using the
models he has on his web page (he mainly referred to the last three). He
revisited the hypothesis testing problem using FDR (False Discovery Rate)
did not find any weird behavior in the referendum data under the models
he assumed. However, he pointed out that it would be possible to come up with
relatively simple mechanisms of fraud that would be basically impossible to
detect with these models. He then presented some new results about Benford’s
Law. He said that there is no reason to believe that Benford’s Law would

<!–
D(["mb","apply to the results of the recall referendum. (My comment: How come the
“,1]
);
D([“mb”,”2000 Presidential election does follow it?). He showed that data simulated
under the assumption that Elio Valladares\’ model was the "fair model"
did not followed Benford\’s Law and looked pretty similar to the recall
referendum data.  He mentioned that he had talked to experts on electoral
“,1]
);
D([“mb”,”data who also said that there is no reason to believe that Benford
“,1]
);
D([“mb”,”would be followed by data like these.

He then mentioned some recent work with exit poll data but didn\’t
have time to go over it. He is now working with the Carter Center on
providing recommendations for future elections.

“,0]
);
D([“ce”]);

//–>

apply to the results of the recall referendum. (My comment: How come the
2000 Presidential election does follow it?). He showed that data simulated
under the assumption that Elio Valladares’ model was the “fair model”
did not followed Benford’s Law and looked pretty similar to the recall
referendum data.  He mentioned that he had talked to experts on electoral
data who also said that there is no reason to believe that Benford
would be followed by data like these.

He then mentioned some recent work with exit poll data but didn’t
have time to go over it. He is now working with the Carter Center on
providing recommendations for future elections.

This entry was posted on February 3, 2005 at 10:00 pm and is filed under Venezuela.