Proponents of politics as a science were left scratching their heads and checking statistical models on Sunday after election results that upturned the status quo.
Though analysts got the narrative right, they vastly underestimated the extent of the changes, which came as a surprise both to the image of Switzerland as a political precision-clock and to the tendency of pollsters to call the results accurately.
The margin of error on Sunday went well beyond the ±1.4% which the Sotomoexternal link institute factored in, spurring politicians from winning and losing sides to raise questions.
The group carries out “election barometers” on behalf of the Swiss Broadcasting Corporation, swissinfo.ch’s parent company, the last of which was published 10 days before the elections.
Speaking on Swiss public radio SRF, Adrian Amstutz of the (losing) right-wing People’s Party called for a ban on public broadcaster money being spent to finance such polls, while Balthasar Glättli of the (winning) Greens wondered if there shouldn’t be a cut-off point in the build-up to an election, beyond which polls shouldn’t be published.
Social Democrat Nadine Masshardt also spoke of “surprising and alarming” inaccuracy and said that her party could have put more effort into mobilising their electorate, had they been better aware of the potential extent of their losses.
This argument about the indirect effect of polls on voter mobilisation was also alluded to by political scientist Georg Lutz in the Neue Zürcher Zeitung newspaper on Tuesday, when he tried to explain another figure pollsters got wrong – the turnout rate, which rather than climbing as expected, actually sunk a few points to 45.1%.
Whereas in 2015, the expected gains of the right and a charged electoral context spurred more citizens to get out and vote, the gloomier prospects of the People’s Party this time around may have dampened the sense of urgency, Lutz said. Neither its supporters, previously motivated by the spectre of immigration and the European Union, nor its opponents, previously motivated by stopping them, turned out in the same numbers as four years ago, and Greens capitalised.
How wrong did they get it?
In its final election barometer on September 22, the Sotomo research group predicted that Greens would gain 3.6%, the Liberal Greens 2.7%, while the People’s Party would lose 2.1%. Most other parties would also drop compared with 2015, the poll indicated, including the Social Democrats, the second-biggest group in the country, who would lose 0.6%.
Similarly, and even more cautiously, a September surveyexternal link by the LeeWas agency for Tamedia reckoned that Greens and Liberal Greens would gain 3.1% and 2.6% respectively, while the People’s Party and the Social Democrats would lose 1.5% and 0.8%.
On Sunday, however, the forecasted wave became a full-on tsunami. The Greens picked up over 6% – the single biggest increase in Swiss political history – the Liberal Greens jumped 3.1%, and the People’s Party took a hit of almost 4%, shifting the parliamentary balance and spurring talk of a new era in Swiss politics.
Even the Social Democrat losses, at 2% not massive in numerical terms, amounted to the biggest setback for the party in a century, and could mark another significant step in the European trend of suffering centre-left parties.
Expect the unexpected
What this means for pollsters and their reputation is unclear. After all, in Switzerland they’re usually quite accurate with their predictions, which they also reckon are useful for the public sphere. Meanwhile, the government’s position is that a ban on such surveys would run against freedom of expression and freedom of the press.
For his part, Sotomo Director Michael Hermann doesn’t claim that polls are an exact science.
Speaking to swissinfo.ch before the elections, he also made the point that people don’t seem to get so irritated when polls at least predict the “general trend” right; it’s rather when there’s a shock upset that questions come, he said – the election of US President Donald Trump and Brexit being recent examples.
Finally, for the extent of the Green wave that took the country by surprise, in an interview with the Aargauer Zeitung last week Hermann said that the party’s surge would be “clear”, though he didn’t offer any insight into how clear it would be. “It depends on whether [the Greens] exceed or fall short of poll projections,” he simply said.
Interestingly, however, he did then go on to refer to recent elections in Austria, where a host of polling companies (by contrast, there are only a handful in Switzerland) underestimated by some 4% the extent of the political shifts, he said – almost the exact same amount by which his own poll underestimated the Greens.