Last week, I published a blog post showing that North London constituencies bucked the national trend by swinging less heavily towards Labour than might otherwise have been expected. I have since repeated the analysis, looking at changes in votes for the Labour Party, Conservative Party, and Liberal Democrats in the twenty British constituencies with the highest Jewish populations. The findings of that analysis are consistent with the view that many British Jews who had previously voted Labour turned away from the party in response to its continued mishandling of its ongoing antisemitism crisis, but — contrary to some indications that ‘the Tories were likely to take huge swathes of Jewish votes’ — they did not turn towards the Conservative Party. Instead, such voters seem more likely to have voted Liberal Democrat.
Because somebody always asks ‘But is it statistically significant?’, I must now point out that – mathematically – the trend away from Labour in areas with high Jewish populations would be considered highly statistically significant, while the trend towards the Liberal Democrats in areas with high Jewish populations would not. That doesn’t mean that one is real and the other isn’t. It just means that, if the data came from a random sample, there would be less than a one in a thousand chance of getting a result like the trend away from Labour seen in the chart above, and a nearly one in five chance of getting a result like the trend towards the Liberal Democrats, through sampling error alone. But we aren’t dealing with a random sample, we’re dealing with the actual votes cast by a whole population, so sampling error is out of the question anyway. In other words, these are the actual trends. How we interpret them is another matter.
Given the lack of any apparent relationship whatsoever between the size of the Jewish population and the Conservative vote, my interpretation of the figures is as follows:
- Many Jewish voters very probably turned away from the Labour Party between 2015 and 2017
- There is no indication that these lost voters voted for the Conservative Party (though some may have on an individual level)
- Those voters seem rather more likely to have voted for the Liberal Democrats
In the interests of replicability, I am providing the R code for this analysis in the form of a downloadable notebook. You can download the data in the form of a tab-separated text file (also containing links where you can verify the voting figures).