A Dodgy Dossier From a Partisan Professor

A widely-reported recent paper from Glasgow University claims to show that the ‘Vow’ had no impact on the outcome of the Independence Referendum. In fact, there’s little or nothing in the paper, the methodology of which is deeply flawed, to substantiate such a finding, while its lead author was a prominent ‘No’ campaigner: and academics can spin just as much as Malcolm Tucker.

That lead author is Professor Ronald MacDonald, who was a prominent campaigner for a ‘No’ vote (see, for example: http://www.scotsman.com/news/politics/top-stories/ronald-macdonald-the-currency-case-for-no-1-3506544). He’s an academic, and he’s entitled to his opinion, but he’s not ‘disinterested’ or ‘neutral’; rather, he’s a partisan member of the camp that has a vested interest in claiming that the ‘vow’ had no impact – because they can then try to claim that Westminster is under no obligation to deliver on it.

Partisan authorship aside, however, the whole basis of the Professor’s paper is fundamentally flawed. It’s based on Internet search volumes, as measured by Google Trends big data, to analyse the active information demand by referendum voters. It therefore ignores those who got most or all of their information about the referendum from sources other than the Internet.

And the major flaw in his analysis is that the demographic most likely to have been influenced by the ‘Vow’ is excluded from his analysis.

The ‘Vow’ was published in a newspaper (if the Daily Record can be described as such), and widely reported on TV. Research carried out by YouGov (http://www.pressgazette.co.uk/survey-reveals-importance-media-helping-scots-make-referendum-decision) for News UK found that the most popular source of information on the referendum was television and radio with 71 per cent, closely followed by 60 per cent of Scots who relied on newspapers and their websites for the majority of their information on the independence debate. This compared to 54 per cent who said they obtained their information from social media and only 44 per cent from the ‘Yes’ or ‘No’ camps directly.

Those demographic groups most likely to rely on traditional media and least likely to rely on the Internet were those that tended to vote “No’.

The Lord Ashcroft Poll conducted overnight on 18/19 September –
shows that those over age 65, and especially those over 75, were those who voted ‘No’ in far larger numbers than the rest of the population. Indeed, if the over 65 vote had been ignored, ‘Yes’ would have won. While it is one poll, the pattern of the demographic breakdown of support for ‘Yes’ and No“ – i.e. men more supportive of ‘Yes’ than women,  and older voters more supportive of the Union – is entirely consistent with just about every poll conducted throughout the referendum campaign.

And guess what else? The demographic most likely to vote ‘No’ is the same as that with least access to the Internet, i.e. the ones excluded from Professor MacDonald’s research. According to Ofcom’s Communications Market Report 2014 (http://stakeholders.ofcom.org.uk/binaries/research/cmr/cmr14/UK_4.pdf) while 82% of adults had access to the Internet in 2014, the proportion falls to two thirds for those aged 65-74, and to a mere 32% of those aged over 75.

That alone should undermine the credibility of the report. But there’s more. As the professor himself admits in the paper, there is no evidence to show that the complex methodology he uses can be applied to voting intentions. He says: “recent studies have demonstrated that the Google trends data can help explain the dynamics of financial markets and make more accurate economic predictions. However, there is no research using Google Trends to gauge political information demand and make projections on voting results (my emphasis), which is the novelty of the present paper.” Translated into plain language, that means “we’ve put some fancy algebra in the paper to make it look impressive, but we’ve actually no basis for claiming that the methodology we are using is valid for analysing voting intentions or not”.

Furthermore, in describing his methodology, Professor MacDonald admits that it relies on a series of assumptions; and he seems very happy to dismiss out of hand anything that doesn’t fit his opinion.

For example, at one point, the paper states: “The polling results from several companies did show some gains on the ‘Yes’ side after George Osborne rejected the idea of a currency union in February 2014. Some reports claimed that the ‘Yes’ campaign had gained momentum from the currency debates. However, such claims remain to be verified.” Not only does polling evidence support the idea that the ‘Sermon on the Pound’ boosted “Yes’; but shortly after it, I spoke to a previously staunch ‘No’ Tory voter who switched to ‘Yes’ because of what she described as George Osborne’s bullying’ – and there were hundreds and thousands of similar conversations in pubs and workplaces all over Scotland. “Cannot be verified’”? Sorry, Professor, but if it waddles like a duck and goes “quack”, you don’t need a quadratic equation to tell it’s a duck.

In short, this is on several levels a deeply flawed piece of ‘research’, the alleged conclusions of which are not supported by the main body of its published ‘evidence’, written not by an unbiased academic, but by a partisan with an agenda to promote.

And in the interests of transparency, I’ll declare my own interest: I’m a long-standing SNP activist, and I’m proud to say that I voted ‘Yes’.

The full report, so you can judge for yourself, is available here: http://www.gla.ac.uk/media/media_392985_en.pdf

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s