Sir Ivor Crewe

Sir David Butler - he was both pleased and mildly embarrassed at being knighted - was the energetic and fertile founder of British election studies. He pioneered two new fields of study in the UK - elections as core democratic institutions and elections as arenas for voting behaviour - and set an enduring standard and agenda for younger generations of electoral analysts who stood on his shoulders.

To the wider public he was known through his election night appearances as a slightly nerdy expert on election statistics and local constituencies, like a cricket nut who could recite the test match scores of individual batsmen (which he did as a schoolboy) He was called a ‘psephologist’ (after ‘psephos’, the small pebble used for voting in Ancient Greece), a term he disliked for trivializing the subject. But he was equally uncomfortable with ‘political scientist’, being suspicious of the over-quantification and elaborate statistical modelling of political life and underwhelmed by the tepid conclusions it frequently produced.

Nor was David Butler very interested in ‘theory’ or even ideas at the grand abstract level. He once told me that he considered himself an academic but not an intellectual. His compulsion was with facts, verifiable concrete facts, about elections in their full dimension: not only the results, but the local and national campaigns, the media’s role, campaign funding, candidate selection, the production of the manifestos and the impact of the opinion polls, so that a complete, granular account - descriptive as well as analytic - could be given of an election as an historic event.

He thought of himself, at least partly, as a modern-day chronicler, establishing his facts not from archives but from the contemporary media, from an unceasing round of interviews, and from official electoral data. The interviews were particularly important. Some were in London with senior politicians, but most were with the infantry and NCOs on the electoral battlegrounds across the whole country: the local candidates, agents and party officers, provincial journalists and electoral returning officers. To understand the electoral process, one needed to listen to what those involved at street level said about it.   

This was the aim of the series of ‘Nuffield election studies’. He was present at their birth as a 21-year old research assistant for R.B. McCallum and Alison Readman’s study of the 1945 election and taking over from 1951. He continued, with a succession of co-authors, for over 50 years until 2005, his fourteenth study, and Nuffield studies have continued to this day. Similarly structured from election to election, with separate chapters on each facet of the election, and the indispensable Appendix by Michael Steed and Sir John Curtice analysing the results, these studies taken together present a unique picture of the historical evolution of British general elections as an institution and process - there is nothing remotely comparable elsewhere.

Critics occasionally dismissed the Nuffield Studies as descriptive and narrative. But they were the inspiration of numerous analytic concepts and general propositions that stood the test of time for many decades such as ‘uniform national swing’ and its offspring, the ‘cube law’, the mainly absent ‘personal vote’ and its cousin the ‘new incumbent effect’, and the inverse relationship between the plausibility of a single poll result and its promotion by the media.

 In 1969 Butler published, with Donald Stokes from the Institute of Social Research, Michigan, Political Change in Britain. It was a path-breaking and transformational study of British voters, unmatched in its originality, rigour and sophistication. Previous studies had relied on local community surveys or constituency-level analysis at risk of the ‘ecological fallacy’ of falsely inferring individual voting patterns from aggregated data. Using national panel surveys for the first time, Butler and Stokes made seminal discoveries that upended habitual assumptions about British voters. It turned out that the traditionally modest swing of the pendulum between the Conservative and Labour parties at elections was the product of a very substantial but largely self-cancelling volatility among fickle voters dipping into and out of voting.

It turned out that that most voters did not think in terms of Left and Right or indeed evaluate parties in terms of their policies, as the rational-liberal model of electoral choice assumed. It turned out that the deep-anchored class alignment of party support was a generational phenomenon that was gradually weakening. It turned out that vote choice owed more to the political equivalent of religion than consumption: a voter’s accumulating party identification (Butler’s term was ‘partisan self-image’) generally determined their policy and leader preferences rather than vice versa, and thus their vote. It turned out that the issues on which parties largely campaigned rarely had the necessary features to shift many votes. Our understanding of British voters - and the meaning of election results - was forever altered and set the framework, and rationale, for the survey-based British Election Studies that have been funded ever since.

David Butler was a model academic political scientist, even if he flinched at the term. In print and on air he was a consummate communicator to a wide audience, explaining quite complicated ideas in plain, muscular, accessible language, without trace of technical jargon, and imparting his evident enthusiasm for the subject. He was a generous patron of opportunity to the young. His young co-authors, who first cut their teeth on a Nuffield Study, went on to do outstanding studies of their own, including Dennis Kavanagh, Anthony King, David Marquand, and Richard Rose. Others at the very start of their careers (such as myself) were given the opportunity to spread their wings and develop specialisms with chapter contributions to his books.  

A chronicler of partisanship, he was scrupulously impartial and always respected as such. It was alleged that he never voted, but this was untrue: he told me (in 1970) that he had voted for all three parties. But no-one could have guessed when he voted for whom. He was non-partisan for parties but cared deeply for democratic integrity - for fair elections, accountable government, responsible and honourable politicians, and the public interest. He will have been dismayed by the last few years.  



Sir Ivor Crewe, Honorary Fellow of the University of Oxford and former President of the Academy of Social Sciences

Photograph: Jane Bown