In 2010, many of the polling firms that conducted pre election polls in previous election seasons either did not poll or did not publicly release polls in the 2010 election cycle.
In light of this fact, it is becoming increasingly difficult to show how well SoonerPoll did using our old grading methodology. Nevertheless, we have updated this page with an analysis of how well we did in comparrison to the only other firm that publicly released polling data on statewide races, Rasmussen Reports.
In 2006 and 2008, SoonerPoll conducted more public opinion polls for the Oklahoma news media than all other pollsters COMBINED. We are proud of that fact and proud of the results we provided to our media partners. So much so, that we are willing to show how well we did.
Many pollsters are quick to promote their results during the weeks leading up to an election, but what about after the numbers are in? SoonerPoll conducts a post-election analysis of all pollsters who publicly released their results.
Our goal is to increase the level of accuracy of every poll we conduct, and to do that we must continually strive to improve what we do. One of the ways we seek to do better is through conducting an analysis following every statewide election.
Because we believe that pollsters should be accountable to the public, we make the results of our post-election analyses available here for anyone who wishes to read them.
The last published poll results from every known pollster will be included in each comparison. SoonerPoll makes every effort to ensure the validity of every poll listed here:
SoonerPoll uses several types of polling measurements, including the Mosteller measurements and other, more recently developed methods.
What are the Mosteller measurements?
Following the 1948 presidential election in which pollsters proclaimed Dewey the winner over Truman, a commission was formed to explore the sweeping failure of pollsters to accurately predict the election’s outcome. The resulting report, by Frederick Mosteller, established eight different methods of measuring the accuracy of a pre-election poll. These methods have endured and remain the most frequently used by researchers when evaluating pre-election polls. For our purposes, they have been adapted for use in all elections, not just those comprising one Democrat and one Republican.
SoonerPoll is unique among its competitors.
SoonerPoll’s outstanding performance is a testament to our strong methodology – developed by us, available only to our clients. We construct our own sample frames based upon a statistical model of likely voters, also developed by us, and we conduct our own data collection only.
Unlike our competitors, we do not buy data from vendors, nor do we outsource telephone interviewing. Because SoonerPoll directs and conducts every phase of a poll, the public can be certain that every measure has been taken to ensure the accuracy of the results we publish.