A Quick Guide to Polling: Part 4
Do it yourself polling = Campaign Malpractice
Week one we went over some basic polling terms: Read Part 1 here. Week two was Benchmark and Tracking polls: Read Part 2 here. Last week we covered how to hire a pollster: Read Part 3 here. This week we cover why do it yourself polling is campaign malpractice.
So, now that you understand the basics of polling, why can’t you do it yourself? Here are several reasons.
First, there is an art to developing a reliable polling model that takes a lot of work and years of experience to be good at. If you aren’t polling people who are going to vote, the information you get is useless.
Second, although you will have input into what topics the questions will be there is a science to writing questions that will give you reliable results. Polling language needs to be carefully crafted.
Third, you want an experienced call center team to do your calls. Improperly trained (or untrained) callers can have a huge impact on your results.
Finally, pollsters use specialized software to crunch your numbers and come up with your results. You can’t simply plug a benchmark poll into an Excel spread sheet and expect to get useful data.
Two quick stories:
Several years ago I had a potential client tell me during a meeting in his office that the campaign had a poll in the field at that moment. Who was doing the poll? He took me to the back room of his office and showed me three teenagers making calls. How had they selected who they were calling? He said, “I gave the interns phone books and to them to randomly call people from the books.”
I decided right then and there I wanted no part of that campaign. He sent me an e-mail that night saying the poll show them up by 31 points and that they would win easily in the fall. They lost by 17 points and only spent about 1/3 of the money they had raised.
In a more recent campaign cycle I was polling in a community I was familiar with, using a pollster who had polled there many times before. The pollster called the first night we were in the field telling me it was going to take longer than he thought, because our response rate was so low. He went on to tell me that they were getting a lot of hang ups, and several of the people they had called had told them they were polled by someone the week before and that the callers were so rude they would never agree to be polled again.
Through the grapevine I found out that a candidate in another race had poll the week before us. He had some friends make his calls. He was also telling people that his poll showed he would win easily. He lost by less than 100 votes.
In the first instance, the campaign was run so badly a reliable poll probably wouldn’t have helped, but in the second instance a good poll would have almost certainly have effected the outcome.
Next week will wrap up with ROBO polls and live callers.