What a strange question. Why did they not simply ask, "Would Mr Robertson's homosexuality make him an unacceptable prime minister for you?" The fact they did not ask the obvious question suggests they knew it would produce a different result.
Perhaps they thought most people would answer it in the negative but only because not many people will admit to a prejudice even when they hold it. Pollsters may think they get a more truthful reading by inviting each person to assess the general view, but this is dangerous.
It asks people something they cannot possibly know and invites them to hazard a guess. Their answers can easily propagate a gross misrepresentation of public opinion, perverting the purpose of polls.
To illustrate their poll, TV3 ran a comment from an adviser to one of Mr Robertson's rivals, David Cunliffe. She had said, "It would be naive to imagine there would be no resistance to a gay prime minister at this point. I think some people might have a problem with it but I certainly wouldn't."
For that, Mr Cunliffe dropped her from his campaign. Her comment was said by MP Clare Curran to be, a "dog whistle" - an appeal to prejudice that dares not declare itself.
Mr Robertson may be the candidate preferred by most of Labour's caucus but if he is not declared the winner tomorrow the result will be explicable for many reasons, none to do with his sexuality. One reason would be that Mr Cunliffe has been more impressive, at least on television where leadership counts most at general elections.
It is hard to believe Labour's voters would let homosexuality count against a candidate even if they suspected the rest of the country would hold it against him. But they have no reason to think so. The best poll of modern attitudes was the gay marriage bill. It passed far more easily than its promoters had expected.
Public opinion has changed, prejudice is not what it was. A gay prime minister would just need to be good enough.