You can of course criticize that, but Forsa exists since 1984 and they are not known for publishing unrealistic polls. The polls are not the problem here. The people turning off their brains and voting for the far right are the problem.
They could, for example, ask for demographic information of the sampled population and weight that to the general population (let’s say that 50% of the people who answered are males from 70+; but they just represent 10% of the general population, then that answers are weighted to represent only 10% of the final results).
So it is a poll of people who answer phone numbers they don’t recognize.
You can of course criticize that, but Forsa exists since 1984 and they are not known for publishing unrealistic polls. The polls are not the problem here. The people turning off their brains and voting for the far right are the problem.
There’s a lot more than that, check the sampling methods on this Wikipedia article:
https://en.wikipedia.org/wiki/Sampling_(statistics)
They could, for example, ask for demographic information of the sampled population and weight that to the general population (let’s say that 50% of the people who answered are males from 70+; but they just represent 10% of the general population, then that answers are weighted to represent only 10% of the final results).
There’s a lot of theory behind sampling.
I’m sure it is probably fine. I’m just being flippant.