What Is an Election Poll? Understanding the Various Methods Politicians Use to Poll and Survey Voters

A political poll is use of survey instruments to elicit and record an individual’s opinions, attitudes, and personal information. Instruments can be designed to capture qualitative data by asking open-ended questions of voters and recording answers in their own words. Quantitative instruments provide fixed options for their respondents’ answers, such as “What is your opinion of the candidate? Choose one: favorable, somewhat favorable, not sure, somewhat unfavorable, unfavorable”.

Modern polls are a form of survey research using a random sample of the population to produce as accurate a picture of public opinion as possible while still accounting for a margin of error.

Election polling has been around since at least the early nineteenth century. One of the earliest examples of a political public opinion poll measured presidential preference in the town of Harrisburg, Pennsylvania between Andrew Jackson and John Quincy Adams. During this era, most polling was done informally generally in the form of a straw poll. The name straw poll refers to the use of straw to tally individual votes. Small straw polls like this one were popular through the nineteenth century until more sophisticated public opinion polls were developed around the turn of the twentieth century. The 1916 election saw the first attempts made at nationwide polling using a representative sample. Nowadays, nationwide polls are continuously run by many different organizations and campaigns, usually increasing in quantity in the lead up to a presidential primary or general election.

Political campaigns generally run three types of polls: benchmark, brushfire, and tracking polls. The difference between these polls mostly has to do with the sequence in which they are run and how they are expected to inform campaign strategy. Generally speaking, each of these types of polls is conducted over the phone using a list of phone numbers rented from larger political organizations. The three stages of polling noted here are used for campaigns regardless of size: from city council candidates to presidential election campaigns. The three major categories of polls are:

1. 1. Benchmark polls: A benchmark poll, sometimes called a baseline poll, is conducted at the beginning of a campaign to establish baselines levels of voters’ perceptions, knowledge, and opinions of a candidate. Benchmark polls are usually made up of a short series of survey questions that provide the poller with a snapshot of the existing political landscape. This allows the organization or campaign to get a baseline of public support and plot their initial campaign strategy.
2. 2. Brushfire polls: These are conducted to gauge changes in voter sentiment during a race. A common brushfire poll seeks to measure a candidate’s popularity by checking “favorable” and “unfavorable” ratings. Brushfire polls are also used to measure support on important issues and to craft messaging, like stump speeches or political advertising. Campaigns will use these polls to gauge public opinion on hot button topics like foreign policy or economic issues.
3. 3. Tracking polls: Tracking polls are regular polls that a campaign runs up until election day. Tracking polls generally ask consistent questions and track upticks or slips in a candidate’s popular support.
4. 4. Exit polls: An exit poll is a type of poll conducted at a polling place on election day. Exit polls require campaign workers to poll voters as they leave their polling location asking them who they voted for. Exit polls are used to predict results and dictate election day strategy changes.

Whether you’re running for congress or conducting local issue-based political research, polling is an essential component of any political campaign. In order to design an effective poll you need to consider the following:

1. 1. Decide what you’re measuring. It’s incredibly important to be very specific about what your poll is trying to measure. Presidential campaigns and political parties run countless election polls in the run up to a general election that are specifically targeted to measure different things. When designing your own opinion polling, it’s important to know exactly the issue you’re trying to track, whether it be name identification within a specific demographic or popular support across the entire electorate.
2. 2. Choose your population. Once you’ve decided what your political survey is trying to measure, it’s time to choose a target population for your polling. There are many different populations that polls target with a wide variety of criteria, and they can often be very specific. For example, one poll might test presidential preference among Republicans, while another might gauge support of a candidate amongst California Latinos over the age of 65. What’s important is knowing who you are hoping to target and how you are hoping to use those results.
3. 3. Select a sample size. Once you have the target population for your poll, you need to choose the scale of your sampling. Polls can’t possibly reach every single individual in their target population because it would be astronomically expensive and logistically impossible. To get around this, polls use a smaller random selection with demographic percentages similar or identical to the target population they are trying to reach. For example, if you were running opinion polling for a Democratic California Senate primary election you might choose to run a poll of a thousand households with roughly the same percentage of demographic groups that make up Democratic voters in California. This way your poll can get an accurate cross section of the target population.
4. 4. Decide your Method of Polling. Polls are run in a variety of ways including: in-person, phone, or online. Political science holds that phone polls are generally the most accurate and cost effective. That being said, great strides have recently been made in improving the accuracy and reach of online polling. Choose a method that fits your budget and intended level of accuracy.

All polls have the potential for error. It’s important to view your poll results with a skeptical eye and be on the lookout for potential biases. Common errors include:

• Margin of error: A margin of error is an unavoidable part of any scientific poll results. Strong survey data based on a sample of at least 1,000 people is expected to have a margin of error of plus or minus three percent, for example. What this means is that if the same poll was conducted with a different sample of 1,000 people at a different time, the results are expected to be within three percentage points of the original poll. Margin of error isn’t like other potential polling errors in that it’s essentially baked in to poll results. The more fundamental mistake would be in not accounting for margin of error as you interpret and implement survey data.
• Nonresponse bias: Response rates are generally very low in most national polls conducted in the United States. It becomes much harder to get a representative sample when only a relatively small portion of your random sample population are willing to respond to your opinion polling.
• Response bias: Response bias is when poll respondents give answers that are not fully reflective of their actual opinions. Other times respondents might give more extreme answers than their actual opinions as a means of pushing poll results further towards their candidate or position. Occasionally poll respondents might feel some social pressure to not voice their support for a particular candidate or position that might be viewed unfavorably by the pollster or public at large.
• Wording issues: Pollsters have to be very careful to word questions in a precise and objective way. Asking questions with an assumption baked into them can lead respondents to answer in a way they feel is expected of them, producing an inaccurate poll.

Polling is an incredibly important part of running a political campaign. Getting accurate poll results can help a candidate do a variety of things including increase voter registration amongst supporters or target potential independents with ads that appeal to them. Understanding polling methodology is vitally important for anyone considering a career in politics.

Learn more about political campaign strategy in David Axelrod and Karl Rove’s MasterClass.