The margin between yes and no was -7% in the provisional results. This may change in the final results, but we can do a quick compare with the six companies that did a public pre-election poll. I’ve calculated for each what they had as the margin between yes and no, once you remove undecideds.
- Colmar Brunton -11% (4% difference)
- Green Party +3% (10% difference)
- UMR +4% (11% difference)
- Reid Research -18% (11% difference)
- Horizon +5% (12% difference)
- Research NZ +7% (14% difference)
So only two companies correctly forecast no would win. However one of them (RR) had a loss by 18%, when it was only 7%. Both the Green Party internal poll and UMR were closer to the correct margin.
I will do a revised post when we have final results as that will probably change things.