Based on my back-of-the-napkin math, you will likely hear about 127 bajillion polls between now and Election Day. It’s important to know how to properly interpret these polls if you want to have a better understanding about what’s going on in the election.
Three Kinds of Polls
Broadly speaking there are three kinds of polls:
- Public Polls
- Internal Polls
- Persuasive Polls
It's important to not confuse the three. That's easier said then done since campaigns will work hard to spin polling and even intentionally confuse voters about what kind of poll is which.
Public polls are generally paid for and released by three kinds of organizations:
- News organizations (e.g The Wall Street Journal, Fox News, CNN, etc.),
- Universities (e.g. Quinnipiac University, Monmouth University, etc.),
- Polling firms (Rasmussen or YouGov)
Public polls are done to help the media better gauge who may win the election and as a public service. Part of this public service often releasing the underlying data. The underlying data for any public polls can be found in their entirety somewhere online in what's know as the crosstabs.
Crosstabs break down the responses to the different polling questions by age, race, gender, political party and more. This allows the media and readers to gauge candidate or issue support by group. Are older GOP voters more supportive of Rubio or Walker? You can find out by checking the cross-tabs.
Cross tabscan also used to examine if a poll is accurate. If you notice an an anomaly, say for instance 95% of Democratic women were supportive of a constitutional amendment banning abortion, you'd be correct in thinking there’s might be something wrong with that poll. This also suggests that other poll results might be similarly skewed.
Public polls are relatively accurate as polling goes. They can predict trends such as the national Democratic trend in 2008 or the national Republican trends in 2010 and 2014. However, public polls usually fail to anticipate wave elections. For example, polling for both 2014 was generally accurate as it showed Republicans winning control of the Senate. Public polls mostly failed to correctly predict Republicans would win by such wide margins in swing states or that Republicans would even win North Carolina.
You will never see a true internal poll written about in the media. Internal polls usually include questions about several different races in a certain area, as well as questions testing the effectiveness of various messages. Campaigns hire polling firms, who then make an educated guess about who is going to vote in the upcoming election and then poll a representative sample of that electorate to figure where they stand and if they have a chance to win.
Internal polls are strategically important to a campaign. They are shared with legislative leadership, candidates, their strategy team, their consultants, and major donors or potential major donors (favorable poll results are used to raise money). Because of their strategic value, full internal polls are not released to the public or media because they may contain information campaigns don't want the public or opponents to know about their strategy, strengths, or weaknesses.
When you read about an "internal poll" being released by a special interest group, campaign, or political party you're actually reading about a persuasive poll. Persuasive polls are released to persuade people to support a cause or candidate with good poll numbers or to generate good press for a candidate. Persuasive polls may be separate from an internal poll (purely to make a candidate look good) or just the most favorable numbers taken an internal.
These numbers are fairly useless because they are unaccompanied by underlying data and thus difficult to verify. Their origin point also makes them highly suspect. Persuasive Polls may also be released by partisan polling firms themselves in order to make their clients or party look good. Public Policy Polling (PPP) is high profile, national Democrat polling firm, and Harper Polling is an example of a national, Republican polling firm.
Now you know a little bit about the three kinds of polls. If you don’t have it totally figured out yet, don’t worry. In Part Two we will look at several polls and analyzes them to see which category they fit into.