The Herald poll of polls is wrong!

I am a big fan of polling averages and simulations. In fact I always say don’t pick just one poll, but look at the average, and the trend over time. So I was delighted when the Herald did Herald-Motu poll of polls. However whatever formula they are using is producing results that simply do not match the actual results it is meant to be averaging or simulating.

There are many ways you can do a poll of polls. I do one which just weights any polls within the last six weeks by size and recency. I don’t know what the Herald-Motu methodology is but it fails the most basic common sense test.

Their current poll of polls says “The Greens are on 12.1%, NZ First is on 11.8%, Opportunity is on 8.5%, Act is on 8.3% and Te Pāti Māori is on 2%.”

So they have Opportunity 0.2% ahead of ACT. Now let’s look at the actual poll results from the last few polls:

• Verian – ACT 11%, TOP 6%: ACT +5%
• Freshwater – ACT 9%, TOP 6%: ACT +3%
• Anacta – ACT 8%, TOP 7%: ACT +1%
• Curia – ACT 10.5%, TOP 4.5%: ACT +6%
• Roy Morgan – ACT 10.5%, TOP 9.5%: ACT +1%
• Reid Research – ACT 7.7%, TOP 5.3%: ACT +2.4%

So the last six polls, all from different firms, have had ACT ahead of Opportunity by margins ranging from 1% to 6%, and the Herald-MOTU poll of polls says that the poll of polls means Opportunity is ahead of ACT.

If this is what the formula produces, then bluntly the formula is wrong. This can’t be adjusting for house effects because all six companies have ACT ahead of Opportunity, yet the poll of polls says they are all wrong and actually Opportunity is ahead.

It can’t even be adjusting for the last election because in fact TOP got a lower election result than the polls had them.

I hate saying this, because I am a big fan of poll of polls. But the results being produced by the Herald should not be taken as an indication of the current polls. Once again, every polling company has ACT ahead of Opportunity, yet their poll of polls says Opportunity is ahead.