Farrar on the poll divergence

I trust David Farrar’s polling, I use his polling for INCITE: Politics.

He knows polling inside out and he has noticed that there is a divergence now between Colmar Brunton and all other polling companies, and presumably his own.

Tonight’s One News Colmar Brunton poll is very different to the recent Newshub Reid Research poll. We seem to be seeing a significant divergence between the two polling companies as it is not just two  that are different about all of their recent .

The five recent polls from the two companies are:

  • 30 August Colmar Brunton Nat 41%, Lab 43% = Lab +2%
  • 30 August Reid Research Nat 43% Lab 39% = Nat +4%
  • 6 Sep Colmar Brunton Nat 39% Lab 43% = Lab +4%
  • 11 Sep Reid Research Nat 47% Lab 38% = Nat +9%
  • 13 Sep Colmar Brunton Nat 40%, Lab 44% = Lab +4%  

The differences between the results for National and Labour are outside the sampling margin of error. Now both companies are very good companies that I often recommend to clients I can’t work for (I recommend UMR also). So it is puzzling that they have diverged not once (which can happen randomly), but consistently in the last few weeks.

I believe that this poll is out of sync. We can assume that because Labour ignored the results of this when Jacinda Ardern flip-flopped on tax issues. They know from their own polling precisely what they need to do. So it is safe to assume that Colmar Brunton, normally a reliable pollster, has done something to give them wonky results.

Next week will be fascinating as either the two companies will converge with their polls (I expect one more poll from both of them), or they will remain divergent – in which case the election results will be good for one of them and bad for the other.

I will say that at this stage Colmar Brunton seems to be in a lonelier position. If UMR were showing Labour ahead by 4% I’m sure we would have heard all about it (Labour consistently reveal their internals when they are doing relatively well). And while historically Curia’s polls have aligned very closely to Colmar Brunton’s, I have found significant divergence in the last few weeks also. It is puzzling as there has been no methodological changes by either of us.

Now being in the minority doesn’t mean you are wrong. We saw that with the UK elections. So I guess we’ll see next week what the final pre-election polls say, and of course the result. Plus of course a lot can change in the last nine days. Worth noting the CB poll is done half over the weekend and half Monday to Wednesday so may not have captured the last two to three days fully, and their next one could be different.

In the meantime people will be saying, well what do we believe? My consistent advice is not to cherry pick the poll you like but look at the average of the polls. And taking this latest poll into account the size and time weighted average of the polls is:

  1. National 42.3% 52 seats
  2. Labour 41.1% 51 seats
  3. NZ First 7.2% 9 seats
  4. Greens 5.6% 7 seats

I expect a Roy Morgan poll out next week also. That is done over a two weeks period so will not be as up to date as the others which are done over four to five days.

Roy Morgan isn’t worth anything.

 

-Kiwiblog

 


THANK YOU for being a subscriber. Because of you Whaleoil is going from strength to strength. It is a little known fact that Whaleoil subscribers are better in bed, good looking and highly intelligent. Sometimes all at once! Please Click Here Now to subscribe to an ad-free Whaleoil.

59%