A guest post by Jonathan Hoffman
(See important update to this post here)
The recent Yachad poll was widely covered in the MSM. Particularly cited was the result that 24% of British Jews would be prepared to “support some sanctions against Israel” if this would “encourage the Israeli government to engage in the peace process”.
It was widely noted that for people under the age of thirty, the figure rises to 41 per cent. Hannah Weisfeld, the Director of Yachad, was quick to use the poll to try to whip up opposition to Israel within the Jewish Community:
“Members of Anglo-Jewry who have previously been afraid to give voice to their concerns over Israeli government policy, should realise that they are in fact part of the majority.”
But the sampling methodology used in the poll leaves questions to be asked. Until they are answered, the probity of the results cannot be assured.
The methodology is set out in Appendix 1 of the report, on page 49. The total sample size is 1131.
Three methods were used to obtain these respondents: (i) sampling ‘distinctive Jewish names’ (DJN) from the electoral register, which generated 418 responses (ii) ‘snowball sampling’, identifying 72 initial ‘contacts’ (representative according to age, synagogue affiliation and location) and requesting each of these to invite their own contacts to participate; this generated 568 responses, or over half of the total sample (iii) using the Jewish members of Ipsos Mori’s market research panel, generating 145 respondents.
The flaw is in the ‘snowball’ second category. We are not told how the 72 ‘seeds’ were chosen. Indeed the report tells us the following:
“The snowball sample over-represented Jews with a left-leaning political stance and those with post-graduate qualifications; this would have produced a dovish bias.”
So more than half of the respondents had a dovish bias!
This would not matter if it was offset by a hawkish bias in the other half of the sample. But this was not the case. The panel sample reportedly had a hawkish bias but that only accounts for 13 per cent of the sample. The DJN sample is reported to be broadly unbiased.
The sample of 1131 therefore has an overall dovish bias. To what extent, we do not know – I am trying to find out.
UPDATE Nov. 17th:
I am told by Stephen Miller, Emeritus Professor of Social research at City University and the lead author of the Report, that the 72 were nominated by “our advisory group plus some other contacts”. The ‘advisory group’ – see page 53 – had ten members. Six are Yachad signatories. Three more are known leftists. I leave it to you to judge whether those 72 (and the 496 respondents they recruited) are likely to be representative in terms of their views on Israel
UPDATE Nov. 18th: Important additional update here.