The numbers game

To some extent, all politics is a numbers game: People running for office need the highest number of votes, and once in government they need the highest number of seats to enact their legislative programme.

So maybe it makes sense that the news media are fixated on numbers, too. The polling season has now begun for the year.

Yesterday, the Sunday Star-Times published the results of the first Colmar-Brunton poll for TV One News. It was headlined “Key’s big leap in PM poll,” but an NZPA version published online today (on both Stuff and the NZ Herald websites) was headlined differently: “Labour gains support but trailing Nats”.

The difference is telling. The Star-Times story began “Just three months after taking the helm of the National Party, John Key has leapt to within a few percentage points of Helen Clark as preferred prime minister.” While Key more than doubled his polling (from 11% in October to 27% now) Key is still well behind Prime Minister Helen Clark (32%, down 1). In fact, he’s still behind the highest poll position of his predecessor, Don Brash (31%). The “Preferred Prime Minister” poll is actually pretty meaningless, little more than a “beauty contest,” since we don’t elect prime ministers, just their parties.

The Star-Times story then reported the current poll numbers for National and Labour, but didn't make clear that National is down 3 points and Labour is up 3 points. These are the numbers that matter because the make-up of Parliament is determined by the percentage of the popular vote each party gets.

After giving only the poll totals and their 6-point gap between National and Labour, the Star-Times claimed “If the poll was translated into an election result, National could easily lead a government unless Labour signed up the Greens and another minor party to give a majority in the house.”

Well, no, it doesn’t mean that at all. Today’s NZPA story correctly said, “If the results were translated into an election with United Future, the Maori Party, Act and the Progressives holding on to their electorate seats both National and Labour would need the support of the Maori Party to govern.” Given the poll results, this is the most likely outcome. The article also begins with the real focus, the relative change in party support.

In the last election, the Colmar-Brunton poll was the least accurate of the main polls, when compared to the actual results. They also completely missed the impending victory of one candidate (Rodney Hide) because of flawed questioning.

By contrast, the TNS Poll conducted for TV3 News was credited as the most accurate overall. They released a poll around two weeks ago that showed only a 3 point gap between Labour and National, as well as slightly more support for the Greens and less support for the Maori Party.

Taken together, these polls tell us—well, very little, really. All polls are nothing more than snapshots, and like a photo of a moving object, they’re a bit blurry is places. At the moment the public and the media are having a love affair with John Key, but as that wanes his personal poll numbers will settle.

The more important number—the relative positions of the various parties—will be changing constantly between now and the election next year. But the truth is, at the moment the polls show that very little has really changed. It’s all in the numbers.