When I get curious about a controversial political issue (okay, that ‘s redundant) one of the first things I do is check the public opinion polls. That’s because I’m a recovering opinion pollster.
One of the most satisfying jobs in my public relations career was as the opinion research director at Illinois Bell Telephone in the 1980s. Opinion polling is an important way to measure public relations work. Tabulating news coverage or web clicks does not really tell you whether you’ve raised awareness or changed opinions. Survey results are a reality check that may shatter assumptions (and I enjoyed being the corporate gadfly).
I had no academic background in opinion research but worked with top-notch outside experts, including an advertising agency research director and a market research professor. Every business lunch with these folks was a master class.
Opinion research can guide policy when it’s used effectively. My biggest project was when Illinois Bell sought regulatory approval to replace a patchwork of local telephone rates in the Chicago area with a single plan of usage-based rates. Because rate proposals like this had been defeated in other states by regulators and legislators, gaining public support was critical.
We conducted more than a dozen focus groups and a number of telephone surveys to determine the issues most important to customers and test their reactions to specific proposals. (I was in hog heaven because the hundreds of millions of revenue dollars at stake in the rate proposal justified a generous research budget.) Our survey results helped shape the final rate proposal and ease negotiations with consumer groups and regulators. The rate plan ultimately was approved.
A focus group is a facilitated discussion with a dozen or so people to elicit their opinions. Researchers observe the discussion behind one-way glass or on video. This is qualitative research that identifies issues and reveals a range of viewpoints but does not represent a majority opinion. It’s a good way to test products and TV commercials. You don’t want the boss observing a focus group, by the way. He’s bound to say: “See? I was right all along. That lady agrees with me!”
Major decisions require a two-step process: Focus groups to identify what the issues are, followed by a quantitative survey of hundreds of respondents to determine majority opinion. Focus group findings help formulate questions for these large-scale surveys. Without this step, you may miss something important when you write the questionnaire.
Once in a while focus groups can guide a clear decision. Illinois Bell’s rate proposal would have eliminated a premium plan with unlimited calling for a flat monthly rate. In the focus groups we noticed that the most outspoken consumers, who emerged as virtually every group’s natural leaders, subscribed to the unlimited plan and liked it. We quickly concluded that angering the people most likely to write their legislators was not a good strategy, and persuaded management to grandfather the unlimited plan for existing customers instead of eliminating it.
Public opinion often changes over time as events unfold, as we see with many political issues. During the run-up to the court-imposed breakup of AT&T in 1984, I conducted a tracking survey that asked the same questions in a series of surveys over several years. We found that customers overwhelmingly favored the breakup until it actually affected them. Then some were indignant: “You mean I have to get long distance phone service from a different company and pay more for local service?” When I presented my survey findings I described the sudden opinion shift as the “holy shit moment.”
We may be seeing a significant shift in public opinion on gun control in the wake of a spate of mass shootings. Congress is working on bipartisan legislation that reflects a strong public consensus for expanded background checks and “red flag” laws but avoids measures with less public support such as allowing teachers to carry guns and banning “assault-style” weapons.
Opinion survey findings are only as useful as their interpretation. When I worked with Unilever’s foodservice business I conducted a first-ever satisfaction survey of hundreds of employees who attended a lavish annual sales meeting. My survey respondents liked the presentations, hotel accommodations and recreational activities but were critical of the food. This worried the sales department. A month or so later I conducted a survey of another employee conference and found similar criticism of the food. Then I noticed that when I went out to lunch with the company’s managers they always critiqued the restaurant. I discussed this with my clients and we came to the same conclusion: This was a food company and the employees were foodies. Criticizing what they were fed was in their DNA and was not a management problem.
Academic research is important and mostly useful. I enjoyed partnering with college professors in surveys for the National Stuttering Association. But I’m inclined to trust corporate surveys a little more because businesses have a bigger stake in the outcome. An academic paper that’s later refuted is embarrassing, but a blunder like Coca-Cola’s New Coke costs millions of dollars and ends careers. Professional polling companies also have a lot of skin in the game.
One of the reasons I check polling data as part of my news consumption is that politicians and pundits not only fail to represent public opinion but often misrepresent it. On the contentious abortion issue, for instance, polls show that the majority of Americans favor a right to abortion in the early stages of pregnancy but only about 20% support late-term abortion and 10 to 15% want abortion to be completely illegal.
The end of Roe v Wade will force politicians to play for keeps on this issue for the first time by actually passing legislation. Opinion polls suggest that most Americans will support the kind of moderate laws most European countries have enacted. Yet some politicians — who apparently have not read the polls — are stampeding to the fringes by pushing for either completely unrestricted abortion or a total abortion ban. It will be interesting to see ow this plays out.
Election polling is particularly slippery. Deciding whom to survey requires guessing which groups of people are going to vote, which may not be the same mix of people who voted last time. I enjoy getting calls from local political pollsters during election season. Sometimes they test campaign arguments by asking my reaction to statements such as: Candidate X failed to pay his taxes in 2011. Does this make your opinion less favorable? My usual response is: No, I already know he’s a crook. Usually I can guess which candidate sponsored the survey.