You are using an older browser version. Please use a supported version for the best MSN experience.

Donald Trump takes poll lead over Hillary Clinton – is it time to panic?

The Guardian The Guardian 23/05/2016 Mona Chalabi Guardian US data editor
Replay Video

For the first time, Republican Donald Trump seems to have edged ahead of Democrat Hillary Clinton in presidential polling. But only just. What does it mean and should those opposed to Trump be worried?

The polling data site RealClearPolitics (RCP) takes an average of national polls that ask Americans who they would choose in a contest between the two candidates (a scenario that now looks inevitable). On Sunday, RCP updated their numbers to show that Trump is now, on average, 0.2 percentage points ahead of Clinton.

That gap might be narrow, but it has still led some (including a senior elections analyst at RCP) to conclude “it’s probably time to panic”. I’d disagree. For those concerned about the prospect of a Trump victory, it’s been time to panic for a while – zooming out from a single statistic shows it.

Real Clear Politics presidential polling average

Delayed panic?

Although this might be the first time that Trump has come out on top in the RCP polling average, he has come very close to doing so on two prior occasions – in September and December of last year. Since September, Clinton’s lead has fluctuated significantly from being as large as 11 percentage points to as narrow as 0.6. In other words, there have been many other periods when Trump’s opponents should have been worrying before now.

Just as it took pundits a while to wake up to the fact that Trump was a sufficiently popular candidate to win the Republican nomination, it seems that they have also been slow to switch focus to his chances of winning the White House.

Part of the reason why they didn’t panic before was a belief that measuring American public opinion a long time before a national election is a bad predictor of voting patterns. Why? Because people change their minds. November is still six months away. The alternative view is that preferences may be becoming hardened now, making voter behavior less likely to shift.

Looking backward, rather than forward, results from the last six presidential elections suggest that 31 states are “safe” – based on the fact that the same party has consistently won them. Unsurprisingly, polling companies are investing their time and resources in states they think might “swing” and therefore determine electoral outcomes nationally – states like Arizona, Florida, North Carolina, Ohio and West Virginia. So far, polling suggests neither candidate has a significant lead in those states which could provide extra cause for concern for those worrying about a Trump win.

Hillary Clinton (L) and Donald Trump (R) © Getty Images Hillary Clinton (L) and Donald Trump (R)

Polling on a pedestal

All that said, the RCP polling average is not the perfect political barometer it is so often held up to be.

The polling average has become the go-to number for those trying to make quick sense of the state of the presidential race. That’s problematic in itself; this election can’t be predicted on the basis of just one number, not least because a candidate’s vote share does not neatly translate to their chances of winning the White House. Each state wields a different number of electoral votes, so it matters how a candidate’s support is distributed throughout the country.

Democratic victories in highly populous states, which have more electoral votes to award, could cause the performance of the Democratic candidate in the electoral college to outstrip her or his performance in the popular vote. In 2008, Barack Obama won 68% of the electoral college vote with only 53% of the popular vote. Clinton may be buoyed in the 2016 election cycle by traditional Democratic strength in populous states such as New York, New Jersey, Illinois and California.

There’s another, much bigger problem with the RCP average, though: it’s only as good as the individual polls that make it up. RCP always takes a straight average of the five most recent polls that were conducted. Once you peek under the hood, you’ll see that an average of 0.2 percentage points flattens out some very different findings. These are the five latest polls RCP used to calculate Trump’s narrow lead:

ABC News/Washington Post: Trump leads by 2 percentage points
NBC News/Wall Street Journal: Clinton leads by 3 percentage points
Rasmussen Reports: Trump leads by 5 percentage points
FOX News: Trump leads by 3 percentage points
CBS News/New York Times: Clinton leads by 6 percentage points

These polls have different results in part because they use different methodologies for assessing public opinion. Although they were all conducted during a relatively similar period of time (which is essential when trying to get a snapshot of what the public thinks now), the polls by NBC/WSJ and Fox News only spoke to registered voters, Rasmussen interviewed “likely voters”, while the ABC News/Washington Post poll simply spoke to adults (82% of whom were registered voters).

All polls claim to speak to a “nationally representative” sample of adults, but virtually none will publish data showing where their respondents were based. Perhaps Rasmussen Reports had to adjust their findings to account for the fact that they only managed to speak to two people in Iowa – we just don’t know. Languages matter too. Two of the polls mentioned above conducted their interviews in Spanish and English – it seems that the others did not.

Methodologies matter. They’re one reason why the footnotes in each poll mention slightly different degrees of accuracy. Take the ABC News/Washington Post poll for example – those numbers have a margin of error of +/- 3.5 percentage points. That means that Trump’s reported lead of 2 percentage points could have actually been as high as 5.5 percentage points. Or it could be that Hillary Clinton actually led by 1.5 percentage points.

Polling is far from perfect. To understand what will happen when the country votes this November, pundits would do better to look at the electoral map, demographics and – most importantly of all – listen to the concerns of voters.

Tom McCarthy contributed to this article.

More from The Guardian

image beaconimage beaconimage beacon