Poll books as historical evidence

What information do poll books contain, and what can historians do with the data? [5-minute read]

Beyond the names of the voters, and the candidates for whom they polled, the internal arrangement of poll books was never uniform – partly because the qualification to vote in England could vary widely from constituency to constituency, and partly because each compiler used their own format. In a county constituency, men possessed of freehold land or property valued at no less than 40 shillings per year were qualified to vote. A typical county poll book, therefore, might include information about each voter’s abode, the address of their freehold, what the freehold consisted of, and (if it was not the voter themselves) the names of the occupiers of the freehold.

A Copy of the Poll for the Knights of the Shire for the County of Norfolk. Taken at Norwich, March 23, 1768 (Norwich: J. Crouse, 1768).

For borough constituencies, the franchise was far more varied, and the information contained within poll books is similarly diverse. Many include each voter’s occupation or trade, as well as their abodes. Some provide details about the basis on which each voter was qualified to vote, and may distinguish members of the town’s corporation, the clergy, and burgesses. More rarely, nonconformist ministers are identified (as in Abingdon’s 1734 poll, and Bristol’s poll of 1754), or Roman Catholics (such as in Preston’s poll book of 1807). Sometimes Quakers can be identified by the fact that they would ‘affirm’ the oaths tendered to each voter, rather than swearing to them.

If you are lucky, a poll book compiler might have combined two consecutive elections, thereby making it possible to track an elector’s votes across both elections: the editor of the combined poll book for the Norwich general election of 1734 and by-election of 1735 wrote that, ‘we have found [it] to be a Work of much great[er] Trouble and Difficulty than we imagin’d’.[1] Equally useful for the modern historian are those poll books which record the order in which electors polled, allowing the changing tides of an election to be observed. In sum, then, poll books contain a vast amount of fascinating data which we can approach from several different angles.

It should be acknowledged, however, that poll books do not necessarily provide a completely accurate list of everyone who polled at an election. Errors or gaps could creep into the record at any point from the initial verbal declaration of a vote; to later transcription or printing errors; to physical damage caused by damp, vermin, or mishandling. Voting was a public and oral activity – voters were required to answer a series of questions posed by a polling clerk, often amongst a cacophony of noise and activity, and quite possibly following a fair volume of alcohol. It would be very surprising if these harried clerks did not occasionally misunderstand voter’s answers, or misspell their names and abodes over the course of a protracted poll. These errors could then multiply during the various copying stages: for example, in variant copies of one Cheshire poll we find a single voter named as ‘Astle Roylance’, ‘Astle Rylens’, and ‘Aston Rylance’.[2] In a printed poll book for the Hampshire election of 1710, the printer explained that, since the poll had been conducted ‘in the arbor’ – essentially an outdoor alcove – some of the names had been rendered illegible by rain.[3] All of this is to say that the recording of votes in the eighteenth century was not the meticulous process it is today – it often took place amidst a chaotic hubbub, open to the elements.

Although some scholars have framed the printing of poll books as simply another phase in which mistakes could be added (and this is true, to an extent), it should not be overstated. Compilers clearly had access to the original poll books and appear to have made every effort to produce an accurate record. In fact, several compilers used the printing of poll books as an opportunity to correct basic errors. Moreover, the vast majority of mistakes are misspellings, rather than outright incorrect information. Where comparisons can be made between variant editions of poll books, the degree of continuity is very high.

What can we do with poll book data?

Despite any mistakes they might contain, poll books remain excellent sources for local history, genealogy, electoral politics, and psephology. A poll book can be approached from a number of directions. Perhaps the most common approach, is to look for trends in voting behaviour. In two-member constituencies, voters had two votes (one for each seat), and could deploy them in different ways: they could vote ‘straight’ for two complimentary candidates; they could ‘split’ their votes between two candidates who were opposed to one another; or they could cast one of their votes for their preferred candidate, and discard the second altogether, in what was known as ‘plumping’. A number of scholars have analysed these voting patterns for evidence of partisanship amongst the electorate: are they voting in an ideologically consistent fashion? This is a more straightforward question when the election was a clear party struggle between two Tories on the one hand, and two Whigs on the other. In the Cornwall election of 1710, for example, 85% of the electors voted ‘straight’ for either the Tory or Whig pairings.[4] However, it can quickly become more complicated if party identities were less clear-cut, or the election was fought between, say, 3 or 5 candidates.

In addition to this voting behaviour, a poll book can present a range of other interpretative possibilities. If we know the order in which votes were cast, it can provide intriguing evidence about when certain groups voted during the often-protracted poll, and in what formations. During the Cornwall election of 1710, on two occasions over twenty clergymen were listed as voting sequentially, having attended the hustings en masse.[5] The geographical distribution of votes may also be significant. Which areas within constituencies, or outside of them, polled most heavily, and in what sequence? We can also assess the electorate’s socio-economic status. It can yield information about local industries, or the propensity for different trades or occupational strata to favour certain candidates. Poll books therefore offer a rich resource for the political, local, social, and family historian. A key aim of the ECPPEC project is to stimulate further study of this underused source.

[1] An alphabetical draught of the polls… (Norwich: W. Chase, 1735), i.

[2] Baskerville, et al., ‘Manuscript Poll Books’, 391.

[3] The poll at the election… for the county of Southampton… (London: Geo. James, 1714), 107.

[4] Richard G. Grylls (ed.), The 1710 Pollbook for Cornwall (Tring, 2002), 6.

[5] Grylls (ed.), The 1710 Pollbook for Cornwall, 8.