In early October, Expressen published its opinion poll from Demokskop, where the Swedish Democrats are clearly identified as Sweden's second largest party, only 0.2 percentage points away from the Social Democrats.
On Thursday, SVT Nyheter in turn published its latest opinion poll from Novus, where 5.3 percentage points differentiate between S and SD.
It may seem strange that the numbers differ so significantly, but according to Dan Hedlin, professor of statistics at Stockholm University, there is an explanation.
- Every institution has different methods, so it is not strange that it becomes different. That this difference is noticeable on SD is also not strange, since the party has quite a lot of dropout in the investigations - so that many who sympathize with SD do not want to participate in investigations. This is something the institutes have to adjust for, and these adjustments are made differently, he says and continues:
- Then different ways are used to collect the answers. Some use, for example, self-recruiting web panels, which we know give uncertain results, even this is something that you have to adjust for.
"Scientifically correct"Torbjörn Sjöström is CEO of Novus.
- I can't comment on other people's results. But the Novus study meets our requirements and SVT's demands for a scientifically correct study, he says.
What are the requirements?
- Very simplified, it is a random selection that is based on all voters in Sweden.
"May wait until the next election"But how should one navigate the surveys to get the best possible picture of reality? According to Dan Hedlin, there is no simple answer to the question.
- If you look at the total amount of surveys and see big differences, then you can probably imagine that the answer is probably somewhere in between. But if you want exact answers, you have to wait until the next election, he says.