Pollsters enter August Twilight Zone

Multi-candidate political party primaries are the most volatile election campaigns a pollster faces. Without an anchor like party identification to tether voters, their allegiance floats from candidate to candidate, especially early in the going. In late summer, we are about to enter a Twilight Zone of presidential politics in which pollsters will have a hard time determining what’s real and what is illusion in poll results.

My first big multi-candidate primary was the 1986 GOP race for Florida governor. Previously, my experience was limited to polling high-interest statewide general election races for newspapers and broadcasters. Mostly, we just asked voters the standard ballots: Will you be voting for John Doe, the Republican, or Jane Smith, the Democrat? It was easy work if you could get it. Most every voter knew the candidates. Most voters chose their party’s nominee, except for the occasional outbreak of ticket-splitting.

The Florida primary was entirely different. The candidates were generally unknown statewide before the race. So tracking this race was crazy. Every time a candidate aired a new TV ad, pleasantly addressing voters for the first time, he would be the pack leader for a week. But then another candidate would air a new commercial and voters would leap to that bandwagon.

That erratic election campaign taught me that, as a pollster, I needed better diagnostic tools than a simple trial-heat ballot test to figure out what’s happening in a primary. Although broadcast TV doesn’t play exactly the same decisive role today in shaping races, it’s still important. But its influence has been diluted by other media that keep the process even more volatile: cable, talk radio, the Internet, bloggers, micro-targeted mail and phone banks. All these constantly threaten to turn a race upside down virtually overnight. So how does a pollster cope?

Insider campaign pollsters have learned to adjust to all this, but public pollsters could take a tip or two from us to survive the roller coaster ride ahead. Every private pollster has his or her bag of tricks for primaries. Here’s mine.

Try asking unaided ballots before you ask the traditional trial heat. Don’t read the names. Simply ask: Have you picked a
candidate yet for the Republican nomination for president? This helps you measure the size of a candidate’s core support. If they are part of his base vote, they shouldn’t need to be reminded of his name.

In a similar vein, ask voters why they prefer a particular candidate. If they mention an issue or anything substantive, they might stick, but if they say, “I saw his TV ad and it was pretty,” mark them down as “leaning,” at best.

Ask about preference shifts — past or future. After asking the traditional ballot and hearing a voter’s current preference, ask in a follow-up question whether they have ever in the past preferred one of the other candidates or might someday switch to one of the others. This produces fascinating results. Voters readily confess to erratic behaviors, which explain the aggregate volatility seen in tracking.

Ask about the likely outcome. Which candidate do voters think will win? Inside a single party, where ideological differences are modest, most voters could support several candidates. For example, a social conservative Republican today could support any GOP candidate except Rudy Giuliani. So if you want an alternative to Giuliani, why not pick the candidate most likely to succeed? That’s voter logic. Go with a winner.

Use open-ended questions to discover the race’s “concrete puzzle solution,” the issue or imagery that is central in voters’ minds. Elections turn on core questions like “Who is most like Reagan?” “Best for the ticket?” or “Likely to beat Hillary?” Once you know what an election is about, you’ll be closer to knowing who’ll win.


Hill is director of Hill Research Consultants, a Texas-based firm that has polled for GOP candidates and causes since 1988.