Editor’s note: This commentary is by Rich Clark, who is the director of the Castleton Polling Institute and an associate professor of political science at Castleton State College.
[N]ot long after arriving in Vermont in 2011, I became aware of the survey that state senator, and fellow political scientist, Bill Doyle issues every Town Meeting Day. I sought out the results and pored over them. I considered the methodology and understood that the data are not collected in such a way as to allow for scientific population projections. Only surveys that employ probability sampling techniques have the claim to representativeness of larger populations by scientific standards.
Nevertheless, I love the Doyle Poll. It shares the same goal as the Castleton Polling Institute to include the people’s voice in political conversation and consider public preferences on important and timely policy debates. It reaches Vermont voters every year precisely where the voters are — at the voting precincts and town halls. And it asks voters about topics their legislators are discussing in the halls, meeting rooms, and chambers of the state capitol building.
Vermonters, with the help of the media, can understand that some public input is collected such that the results have some generalizability and other input lack such representativeness without dismissing the latter out of hand.
That the Doyle Poll does not employ probability sampling procedures and lacks the rigors of data collection in scientific polling is problematic if we are to stipulate that the completed responses are representative of Vermont voters in a given election. They may be, but it cannot be proved. In fact, they are more likely not representative, and even if the final results are close to those found with polls using probability sampling.
Despite the methodology, the data from the Doyle Poll do represent the views of a large number of voters who took the time to complete the survey and offer their opinions to the researchers. Those opinions should become part of the conversation.
Keep in mind that those who write letters and testify before legislative committees are not likely representative of the entire population of Vermonters or of Vermont voters, but we listen and consider those views as part of the broader conversation. That is how we might consider the Doyle Poll report. Because self-governance is so important to us, and it requires an involved citizenry, we should be open to many sources of public input, including scientific polling, to broaden the public discussion. Vermonters, with the help of the media, can understand that some public input is collected such that the results have some generalizability and other input lack such representativeness without dismissing the latter out of hand.
I look forward to the Doyle Poll and the reported results, and while I don’t conflate them with polls done by other methodologies, I do think that they are a valuable contribution to our political dialogue in Vermont.
