http://professor-lyman.livejournal.com/ (![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png)
professor-lyman.livejournal.com) wrote in ![[community profile]](https://www.dreamwidth.org/img/silk/identity/community.png)
fandomhigh2006-02-21 01:33 pm
professor-lyman.livejournal.com) wrote in fandomhigh2006-02-21 01:33 pm
Entry tags:
Political Campaigning (Tuesday, February 21, 4th period)
Josh smiled as the campaigning class came in. "Hope you all enjoyed your three-day weekend," he said. "Alanna, how many days until the election?" He waited for her answer before continuing. "Today we talk about polling."
He turned to the board. "Polling provides two major things that campaigns can't live without," he said, beginning to write. "It provides a political profile of the state, and an opportunity to see if your message is resonating with the people you want it to.
"When I say political profile, I mean it gives you more of a feeling about your voters than the numbers I gave you from the the voter files. You need to look at three distinct dimensions: a demographic profiling of the voters; an issues profiling of the state; and a profile of what the voters think about the current incumbent." He grinned. "You have to know who you're talking to before you can figure out if anyone's listening. Things like party identification, their education level, income, religion, occupation, age--who votes, who votes for you, and if possible, why. You also need to get hard numbers backing up your feeling on what issues would be important, and who thinks so. And make sure those numbers are the same people you want to be talking to. If you discover that the falling prices of beef is a huge issue, but no one who cares about it would be voting for you anyway, well, that's not an issue you have to really worry about."
He looked down at his notes. "Then you need to ask yourself when to poll. These sorts of things don't come cheap. The political parties will be doing polling, and various interest groups, and newspapers and CNN, but your internal polls will always be asking slightly different questions. If you ask yourself the question, 'would I be willing or able to do anything differently depending on the results of this survey?' and don't come up with a decent answer, don't spend the money."
He smiled. "Then comes the writing of the questions. I know that people are constantly talking about how you can twist numbers to make them say whatever you want, but in the case of internal polling, you don't want the numbers to tell you what you want to hear. You need the numbers to tell you the truth. So when you're writing questions, make sure you allow for people to say that they don't know, and give them a range of responses to choose from. When you're pulling a sample, especially in places like Montana where people living on reservations might not have phones, make sure to take that into account. Pulling names out of a phone book isn't that effective--voter lists tend to be better, as is random digit dialing, but both of those can get expensive.
"Polling doesn't have to be all phonebanking and thousands of people wearing stupid headphones. Your door-to-door work should also be asking the same kinds of questions: who are these people, what do they care about, what would it take to have them vote for your guy? Ask as many people as possible the same questions until you either get a definite no and can mark them off your list for potential voters or get a definite yes and you don't have to bother them until the weekend before election day."
He sat down. "Typically, you'd do a huge, intensive, incredibly expensive poll during the early stages of the campaign, then begin with tracking polls once a month if you can afford it starting in about June, and then down and dirty week-by-week polls beginning in late September or early October."
He looked around. "Questions?"
He turned to the board. "Polling provides two major things that campaigns can't live without," he said, beginning to write. "It provides a political profile of the state, and an opportunity to see if your message is resonating with the people you want it to.
"When I say political profile, I mean it gives you more of a feeling about your voters than the numbers I gave you from the the voter files. You need to look at three distinct dimensions: a demographic profiling of the voters; an issues profiling of the state; and a profile of what the voters think about the current incumbent." He grinned. "You have to know who you're talking to before you can figure out if anyone's listening. Things like party identification, their education level, income, religion, occupation, age--who votes, who votes for you, and if possible, why. You also need to get hard numbers backing up your feeling on what issues would be important, and who thinks so. And make sure those numbers are the same people you want to be talking to. If you discover that the falling prices of beef is a huge issue, but no one who cares about it would be voting for you anyway, well, that's not an issue you have to really worry about."
He looked down at his notes. "Then you need to ask yourself when to poll. These sorts of things don't come cheap. The political parties will be doing polling, and various interest groups, and newspapers and CNN, but your internal polls will always be asking slightly different questions. If you ask yourself the question, 'would I be willing or able to do anything differently depending on the results of this survey?' and don't come up with a decent answer, don't spend the money."
He smiled. "Then comes the writing of the questions. I know that people are constantly talking about how you can twist numbers to make them say whatever you want, but in the case of internal polling, you don't want the numbers to tell you what you want to hear. You need the numbers to tell you the truth. So when you're writing questions, make sure you allow for people to say that they don't know, and give them a range of responses to choose from. When you're pulling a sample, especially in places like Montana where people living on reservations might not have phones, make sure to take that into account. Pulling names out of a phone book isn't that effective--voter lists tend to be better, as is random digit dialing, but both of those can get expensive.
"Polling doesn't have to be all phonebanking and thousands of people wearing stupid headphones. Your door-to-door work should also be asking the same kinds of questions: who are these people, what do they care about, what would it take to have them vote for your guy? Ask as many people as possible the same questions until you either get a definite no and can mark them off your list for potential voters or get a definite yes and you don't have to bother them until the weekend before election day."
He sat down. "Typically, you'd do a huge, intensive, incredibly expensive poll during the early stages of the campaign, then begin with tracking polls once a month if you can afford it starting in about June, and then down and dirty week-by-week polls beginning in late September or early October."
He looked around. "Questions?"