Philip Stark a UC Berkeley statistics professor has developed a new method to better determine the accuracy of voting results, a formula which will be used in some counties today.
"There was previously no method of auditing that certified that you got the right answer," Stark said.
The method takes a random sample of precincts and tests to see whether each precinct's margin of error between the computer's count and a manual count is small enough to be considered insignificant.
"What we are basically trying to see is whether there are a large amount of samples with a small enough error," he said.
Before this method, it was difficult to ensure that counts provided a reliable level of confidence, Stark said.
Stark's method came after Bowen commissioned a statewide Post-Election Audit Standards Working Group in order to develop a way to improve the accuracy of election counts.
While no future plans have been developed for his method, Stark said it is currently in the experimental phase.
"We're just testing to see if the results are feasible," Stark said.
more here and here
No comments:
Post a Comment