Open source would not be nearly as impressive as reproducible.
I guess you could conduct the same poll with multiple sets of people and see how the results compare, but in a way that's what poll aggregators do.
The result is qualified, comparisons of results shouldn't ignore that.