Home / Articles / The truth is (probably) in the numbers: the excel sheets that could topple giants

The truth is (probably) in the numbers: the excel sheets that could topple giants

The use of opinion polls in elections goes back to the early 19th century, though the current method of representative sampling was developed in the 1930’s. Franklin D. Roosevelt became the first presidential candidate to use polling in the election of 1932, but it was after John F. Kennedy successfully based his campaign on creative polling in 1960 that the polling industry truly boomed. Nowadays, polls are used consistently and constantly, and on a large number of different topics.

John F. Kennedy jumping out of a car during his presidential campaign in 1960, the first campaign to successfully use extensive opinion polling.
John F. Kennedy jumping out of a car during his presidential campaign in 1960, the first campaign to successfully use extensive opinion polling.

But there might be reason to question this thorough dominance of opinion polling. Do the polls actually measure what they say they measure and are the respondents actually representative of the group they’re said to represent? The response rates to public opinion polls have dropped from 90-something percent in the 1930’s (when pollsters used to go around knocking on people’s doors) to single digits today. This might not be a problem, though, as long as the polls are seen as a guide and not gospel. More often than not, however, polls are taken at face value, and sometimes that has real political consequences.

Gearing up for the 2016 American presidential election, the Republican party has so many contenders for the nomination that the debates have had to be split into two, one major debate and one for the candidates who didn’t qualify for the major one, sometimes referred to as the “kids’ table”. The determining factor in who gets to be on the main stage is an average of a number of national polls.

Fox Business Network and Wall Street Journal, the hosts of the latest Republican debate, the fifth in this primary election (held Wednesday Nov. 10), announced a couple of days ahead of the debate who qualified for their main stage. This was a big reveal because they hadn’t previously announced which polls they would be including in the determining average, and the selection of polls would determine the fate of a few of the candidates, as the results varied slightly.

The rules were different this time. In order to qualify for the main stage, a candidate needed to average at least 2.5% in the four polls, and at least 1% was needed for the kids’ table. New Jersey Gov. Chris Christie, who had qualified for the earlier debates where all you had to do was be among the top ten candidates, regardless of actual numbers, now found himself outside the main field. He averaged 2.25%, and was narrowly delegated to the smaller debate. Three candidates did not qualify for even the small debate, as they failed to reach the required 1%.

Of course, this has a very real impact on these candidates. While presumably George Pataki, former New York Governor and one of the candidates who failed to reach even 1%, was never going to be the nominee, Gov. Chris Christie has been deemed a competitive candidate, at least in the past. Either way, the election is still a year away from now, and the primary election is far from being determined by these early polls. But being excluded from the main debate – or in Pataki’s case, as well as two other candidates’, being excluded from a publicity opportunity altogether – is certainly not going to help. Especially for a candidate like Christie, whose ability to bluster loudly at other candidates is the source of much of his political clout.

And it’s all based on a very narrow interpretation of opinion polls. While the result is an average of four polls, the difference between 0% and 1% is still completely statistically irrelevant. Not to mention Christie’s case, where a difference of 0.25 percentage points – not a statistically significant margin by any means – may have doomed his entire campaign.

On an even bigger scale, the recent controversy in the Australian parliament has its base in opinion polls as well, at least on a rhetorical level. Malcolm Turnbull replaced Tony Abbott as Prime Minister and leader of the Liberal Party after forcing a leadership spill, a vote among the party’s representatives in parliament. This sounds less extreme put into the context it deserves – this is in fact the third time this has happened in only the past five years, and Turnbull now makes the fourth Prime Minister in two years. The last PM to serve a full term left office in 2007. And it’s all in the polls.

Before the 2013 election when the then-Abbott-led Liberal Party took power, Labor, now the opposition party, led the Australian Parliament. Kevin Rudd replaced Julia Gillard as Labor party leader and Prime Minister only a few months before the election, after forcing a leadership spill. Before anyone starts feeling too bad about that, though, let’s note that that’s exactly how Gillard got the job in the first place. She ousted Rudd in just the same way in 2010, when the conventional wisdom (that is, the wisdom of the polls) was that Labor would lose the next election if Rudd was at the steering wheel.

The attentive reader might be starting to recognize a theme here. Before this Turnbull-led coup, the Liberal Party criticized Labor for metaphorically throwing their leaders out the window as soon as they started looking a little bit unsteady. Obviously that’s been true in Labor, as bad polling was the rationale for a number of leadership votes called, even looking further back than just the swapping of seats between Rudd and Gillard. But now it was also the reasoning behind the toppling of Abbott, prompting criticism from within the Abbott wing that the Liberal MPs let “a few silly polls” throw them into a panic. While this may be true, the Abbott-led Liberal Party had been trailing Labor in polls for over a year, with Abbott consistently polling below opposition leader Bill Shorten, when Turnbull saw his opening.

And it seems to have worked, too. Immediately after Abbott’s ousting, Turnbull’s personal ratings have skyrocketed, and are well into 60% now, according to some polls. At the same time, there’s no discernible difference in the numbers for party preference, which begs the question if this switch was actually the saving grace for the Liberal Party that it was made out to be. Either way, given the way the revolving doors of the Prime Minister’s office seem to work, Turnbull is probably not going to get too comfortable in his new seat. Much like Governor Christie and his co-contenders in the Republican presidential race, he can’t afford to let the small statistical margins slip to his disadvantage.

Klara Fredriksson

Leave a Reply

Your email address will not be published. Required fields are marked *