TCHQ Tip of the Week: Polling Part 1
(Please Note: After reviewing my first draft of this tip I realized it was way too long so here is Part 1 of 4 on Polling:)
By now you might have noticed a recurring theme in many of these tips. I keep refer to your polling or focus groups. Let’s take a look at what polling or focus groups are and how your campaign will benefit from them.
I’ll deal with focus groups in greater detail in another post, but for now if you are not familiar with the term focus group, it is a diverse group of people assembled to take part in a guided discussion about a topic. In our case; a political race. Depending on budget, time and other factors a focus group would have about a dozen people and you might do anywhere from a handful to several dozen different focus group session. It will give you a very nuanced picture of how people respond to different aspects of your campaign.
Polling asks a sample group of people questions. A sample group means a group of people who statically represents (age, D to R to I, sex, income and other demographic variables) the likely voter population. The questions ask for they think and depth of commitment on particular issues, such as do they support a political party, candidate or issues in the election. It can also asks questions about their personal habits. Favorite news source, radio and TV stations etc.. Depending on the type of poll, budget, size of the district and the number of likely voters pollsters usually contact between 300-1000 people, by phone and ask their questions.
In short, polling tells you what people think, focus groups tell you why they think it.
Margin of Error and Confidence Level
All polls have a margin of error and a confidence level. A margin of error is the maximum expected difference between the actual results and the poll results. Confidence level is the level of confidence that the margin of error is correct. These numbers come from a formula that takes into account the total number of likely voters that will vote and the total number of people in your sample. For example, a sample size of 400 in a race where 100,000 votes would be cast would have a margin of error of +/- 4.88% with a 95% confidence level.
You might think, “Great let’s call 800 and the margin of error would only be +/- 2.44%” Sorry, but it doesn’t work that way. 800 would only lower the margin of error to +/- 3.45 and would double, or more, the cost of your poll. You will need to balance your cost against margin of error.
One last note on margins of error. If you have a margin of error of +/- 4% (phrased as plus or minus 4 percent) and results say you have a 40% approval rating that means the range is 36%-44%. That may sound like a lot, but polls are a single snap shot, if you do 3 more polls and your approval rating is 40%, 39%, 42% and 41% you can be pretty sure your approval is right at 40%.
Next week I’ll discuss the difference between Benchmark Poll and Tracking Poll, as well as the roles they play in your campaign.