Pulse Poll: Counting votes and Fake votes

Written by: Tim Watson on Nov 6, 2007 4:00 PM EST

Over the past few weeks, we here at DFA have fielded many questions related to how we are using the 2nd and 3rd place votes and to the security we had in place to deter any fraudulent voting. Charlie asked me to explain these two topics, so here it goes.

We decided to ask for 2nd and 3rd place votes as a way to get a better feel as to who you supported after your first choice. It was obvious from before the poll even began that we would get a good number of votes for Al Gore. In fact over 38,000 people did, and since it now appears Gore will not be in the running, we can see who those people may vote for in their states primary. Those Gore votes end up going to: Edwards: 30%, Obama:18% and Kucinich: 15%.

Many of the questions about the other choices were wondering if we were going to use IRV (instant runoff voting) to determine a winner. Although while our poll winner was crowned solely on first place votes, our poll results page allows one to simulate an IRV scenario. (Interestingly enough, you have to remove all but Kucinich and Edwards to get one candidate above 50%. Kucinich wins 51% to 49%)

On the security side of things, several attempts where made to steer the poll in a certain direction with fraudulent votes. Twice one candidate had huge spikes that when investigated were obviously not unique people, but computer programs written to submit votes. All the votes were coming from the same ip address and the email used had a counter on it to make the email unique. (something like abc01@gmail.com abc02@gmail.com and so on) Obviously we removed those votes as soon as we determined them fake.

In the end, we found just under 1% of all votes had an invalid email address, so we are confident we caught any fake voting that would have caused significant changes to the poll outcome.

Tags: