How to rig a poll

One of many reasons that I distrust political polls is the ease with which the results may be skewed to justify a perspective that the poll-taker wants to support. A good case study is explained in an article in the UK Spectator, just published.  Australia has announced that it is considering moving its embassy in Israel to Jerusalem, following the lead of President Trump. The Australian Broadcasting Corporation, as reflexively progressive/anti-Israel as its UK counterpart, the BBC, commissioned a poll that got the result it must have wanted.

Peter Wertheim explains the problem:

A Roy Morgan SMS survey undertaken on December 14-15, 2017 ... It asked the question: Do you support or oppose President Trump’s recent decision to declare Jerusalem the capital of Israel? The survey found 76 per cent of Australians opposed the Trump announcement and 24 per cent supported it...

The wording of the question suffered from several defects. It linked recognising Jerusalem as Israel’s capital with Trump, who on any view is a polarising personality. It also mis-characterised Trump’s decision. Trump did not ‘declare’ Jerusalem to be the capital of Israel; the US recognised it as already being Israel’s capital...

Photo credit: David Holt

My organisation, the Executive Council of Australian Jewry, was keen to test the veracity of the Roy Morgan survey. We commissioned YouGov/Galaxy to conduct a poll asking: In 1949, Israel designated Jerusalem to be its capital city, and has its parliament there. Do you think Australia should recognise Jerusalem as the capital of Israel? The survey was conducted in February among 1,205 Australians....

The results paint a very different pictureto the published Roy Morgan findings. A key finding of the YouGov survey was that when the question of Jerusalem was framed as one of whether to ‘recognise’ (rather than ‘declare’) Jerusalem as Israel’s capital, and was asked without mentioning Trump or the US, Australians supported recognition by a margin of almost two to one (40 to 21 per cent).

Hat tip: Andrew Bolt, John McMahon

One of many reasons that I distrust political polls is the ease with which the results may be skewed to justify a perspective that the poll-taker wants to support. A good case study is explained in an article in the UK Spectator, just published.  Australia has announced that it is considering moving its embassy in Israel to Jerusalem, following the lead of President Trump. The Australian Broadcasting Corporation, as reflexively progressive/anti-Israel as its UK counterpart, the BBC, commissioned a poll that got the result it must have wanted.

Peter Wertheim explains the problem:

A Roy Morgan SMS survey undertaken on December 14-15, 2017 ... It asked the question: Do you support or oppose President Trump’s recent decision to declare Jerusalem the capital of Israel? The survey found 76 per cent of Australians opposed the Trump announcement and 24 per cent supported it...

The wording of the question suffered from several defects. It linked recognising Jerusalem as Israel’s capital with Trump, who on any view is a polarising personality. It also mis-characterised Trump’s decision. Trump did not ‘declare’ Jerusalem to be the capital of Israel; the US recognised it as already being Israel’s capital...

Photo credit: David Holt

My organisation, the Executive Council of Australian Jewry, was keen to test the veracity of the Roy Morgan survey. We commissioned YouGov/Galaxy to conduct a poll asking: In 1949, Israel designated Jerusalem to be its capital city, and has its parliament there. Do you think Australia should recognise Jerusalem as the capital of Israel? The survey was conducted in February among 1,205 Australians....

The results paint a very different pictureto the published Roy Morgan findings. A key finding of the YouGov survey was that when the question of Jerusalem was framed as one of whether to ‘recognise’ (rather than ‘declare’) Jerusalem as Israel’s capital, and was asked without mentioning Trump or the US, Australians supported recognition by a margin of almost two to one (40 to 21 per cent).

Hat tip: Andrew Bolt, John McMahon