Professional pollster Matthew Jason contributed this article to The Campaign School to help candidates avoid the common mistakes he often sees in political polling.
Polling Mistake #1: Not Testing Campaign Issues or Messages
“People had an illogical, self-serving rationale when it came to interpreting the behavior of others.” — Best-selling author Marisha Pessl
Why is your campaign conducting a survey?
Do you have a clear set of goals, a valid rationale for expending limited resources on a voter survey?
Every campaign cycle I am approached by candidates or campaign managers who are far too myopic in their thinking about why they want to do a poll.
A lot of campaigns make this mistake of being too shortsighted in their thinking.
You do not conduct a survey simply to find out how well-known your candidate is and/or what the head-to-head ballot looks like at that point in history.
These are just some fringe benefits you gain from a political poll.
There is so much more important information to be gained for very little additional cost.
My rule of thumb has always been that a campaign should not bother with research if they are not going to receive information on their survey which will be of value to them throughout the duration of the campaign.
If you are surveying voters without making every effort to understand the issue environment in your area and/or without trying to determine which messages have the potential to cause voters to support your candidate, then something is seriously wrong.
The big data books my clients receive after I have completed a survey on their behalf often times become a playbook of sorts, remaining close at hand for easy access throughout the campaign.
Some clients even keep them around for use in future campaigns taking place in the same area.
Polling Mistake #2: Polling Too Early or Too Late
“The essential ingredient of politics is timing.” — Former Canadian Prime Minister Pierre Trudeau
Most campaigns do not realize how important the timing of their research can be.
If your local campaign only has the finances available to conduct one poll, you must be very careful about the timing of that survey.
If you conduct your political survey too early (before the filing deadline), you may have a ballot question which does not include all the candidates running for your office, or you might not have all the candidate titles exactly in line with what the voters will actually see when they look at their ballots.
Such items may seem trivial to you now, but every little detail matters, especially in local elections.
Conversely, if you conduct your political survey too late, you may not have time to make proper use of your findings.
Your campaign can have the most on-target, motivational message in the world, but if you do not fine tune that message in time to use it in your stump speeches, your mail, and your social media efforts, it will not do you any good at all.
Stephen Kinney, my mentor in this business, always said, “A poll is just a snapshot in time.”
He then went on to talk about how major events (usually earthquakes since he lived most of his life in Southern California) or political scandals have the potential to change the entire dynamic of a campaign overnight.
There is very little you can do in such circumstances, especially on a tight budget local campaign. All you can do is plan as wisely as possible.
In this regard, I often advise my clients to conduct their research as soon as possible after candidate filing has closed and hope there are no surprises.
Polling Mistake #3: Polling People Who Won’t Be Voting
“Nobody will ever deprive the American people of the right to vote except the American people themselves and the only way they could do this is by not voting.” – President Franklin D. Roosevelt
Who are you interviewing in your poll? If they are not registered to vote, they do not matter to your campaign.
If they have never voted in a city election in the past, the odds are strongly against them suddenly doing so now.
You would be surprised how many local campaigns waste valuable resources talking to people who either cannot or will not cast a ballot in their race.
What voter information is available to your campaign varies from state to state, but we live in an age where a list of people who you can know for a fact are likely to vote is easily available to your pollster.
In California, where I have done the majority of my campaign research, I can even choose from all kinds of different likely voter universes which are roughly representative of various levels of turnout.
Ask your pollster about where he or she acquired their voter list and what steps are being taken to make sure your results represent the opinions of people who are at least somewhat likely to cast a ballot.
Polling Mistake #4: Not Testing A Candidate’s Weaknesses
“Success is achieved by developing our strengths, not by eliminating our weaknesses.” — Guinness Record Holder for highest recorded IQ Marilyn vos Savant
Most candidates are very hesitant to explore voter reaction to their own mistakes or weaknesses.
I know people who make a very nice living nowadays digging into the backgrounds (public records and beyond) of candidates running for office.
If your opponent’s campaign is doing its due diligence, they will find out about some items in your past and try to use them against you.
It could be as simple as a vote you once made while serving on a city planning commission or a lunch meeting you once had with someone who was later found to be corrupt.
It could be a tax lien once placed on your home or business (even if it has long since been taken care of) or a bankruptcy in your past.
It could be the fact that you have not always exercised your right to vote in past elections.
If there is anything to be found, chances are your opponent will find it.
You absolutely need to know how such items will impact your campaign if your opponent chooses to try and use them against you so you can be prepared with a response.
There are ways to test these type of issues discretely. You do not need to tie your own name to such questions.
I often set up a generic series on my political surveys which asks voters how finding out “x” about “a candidate” would impact their vote.
If you have deep, dark secrets which could tear your life apart and/or cause you to face criminal charges, maybe running for public office is a bad idea to begin with.
This mistake can be almost as bad for your campaign in reverse as well.
You also need to know everything there is to know about your opponents, and you need to test that information (both positive and negative) in the same manner you are testing information about yourself.
You cannot know how an election is likely to turn out if you have not taken a look at the best and worst case scenarios.
Let me be clear about one thing here: I do not advocate and will not ever personally be involved with so-called “push polls.”
My surveys will always test the positives and the negatives of all candidates equally or not test them at all.
If you have valid reason to believe the information you are testing to be true and are fair and balanced in the way you go about the testing in relation to your own faults, it is not a “push poll.”
Polling Mistake #5: Asking Useless Questions
“We must beware of needless innovations, especially when guided by our logic.” – Former British Prime Minister Winston Churchill
Why are you asking a particular question on your survey? What is the campaign going to gain from the results?
A lot of campaigns ask virtually useless questions on their surveys just because they have a personal interest in seeing the results.
This point may be a little more subjective than my earlier points.
Certain consultants have certain questions they always like to ask. Some of these questions may appear useless to the outsider, but chances are the consultants involved have a very valid reason for asking them.
They often have a long track record of asking a certain question and have developed their own methodology for interpreting the results of that question in a way that can be helpful to your campaign.
Pollsters are the same way.
There are questions I ask on virtually all my surveys which some people initially view as useless information and a waste of their money.
I am always happy to take the time and explain to them exactly how I plan to use a certain question or questions to their benefit. In most cases, the client ends up seeing my logic and allowing those questions to be asked.
Some of my biggest issues with unnecessary questions over the years have been in relation to demographics.
Unless you are conducting a micro-targeting campaign and interviewing thousands of people, there are only a handful of demographic questions which need to be asked on every survey.
Age, ethnicity (in many areas), geography (if you are not working from a list which includes this information), gender, and political affiliation are questions you always want to ask on a campaign survey.
Other questions such as church attendance, education level, homeowner versus renter, household income, and political ideology often have very valid uses for a campaign as well if you can draw conclusions from them which will help in your campaign targeting.
My point here is that you should not be afraid to ask your campaign consultant or your pollster why a certain question is being asked and what you can expect to gain from it.
If it is important, someone will be able to explain why.
Likewise, put your own question ideas to the same test, and do not be afraid to ask others about them.
A large component of how surveys are priced is how long it takes to complete an average interview.
I am happy to ask whatever questions you like, but you need to know that every additional question has some impact on pricing in relation to a telephone survey.
Matthew Jason is a professional pollster and owner of Candid Research Solutions, a full service market research firm which specializes in determining what people truly believe, understanding the motivations behind those beliefs, and harnessing those motivations to the advantage of its clients. Mr. Jason can be contacted at candidresearch.us.