This is a chart displaying the "raw data" for each candidate that has been polled by at least one of the poll in my model. The poll results here are grouped under each candidate and color coded according to poll. The order of the candidates from left to right is determined by the unweighted average of the poll results for each candidate. Polls in which a candidate was not included are excluded from the average for that candidate.
Example: Mitt Romney got 19% in the only two polls that tested him. Therefore his average is 19% and not 7.6% which is what it would be if I used 0's in places where the candidate was excluded. For these preliminary purposes, I am concerned with how these potential candidates are testing in the polls of which they are included. Ergo, candidates are not penalized for not being a part of a particular survey.