The Secret Ballot Does Not Eliminate but Changes the Type and Timing of Election Violence: Evidence from Election Violence Deaths 1832-1914 in England and Wales

In this guest blog, Lydia Buckroyd explores the impact of the 1872 Secret Ballot Act on election violence and in particular on election violence deaths. Contrary to widely held views, the introduction of the Secret Ballot did not eliminate, but merely changed the type and timing of election violence deaths:

In the period between the Great Reform Act and Great War, British elections transformed from public, rowdy, and often violent events to the more private, tranquil occasions that we recognise today. The violent nature of elections during the nineteenth Century is starkly reflected in the number of fatalities. It was routine in the nineteenth Century to see several deaths per election, sometimes several killed in a single event. One of the worst examples of this was the Sheffield election of 1832, when five people were shot by military forces called in to disperse an election riot (Bell’s Weekly Messenger, 1832). But by the turn of the century, only the very occasional fatality was recorded.  Why did British electoral conduct change so dramatically?

Cartoon the introduction of the secret ballot to Britannia by the newly elected MP in Pontefract, the first by-election to be held under the Secret Ballot Act in 1872. Source: Punch, 24th August 1872.

This Punch cartoon offers an explanation that was widely agreed on by contemporary observers. It depicts the Pontefract by-election, the first election in the UK after the 1872 Ballot Act introduced the secret ballot to UK parliamentary elections. In it, ‘Little Boy Ballot’ is presented to Britannia by the successful Liberal candidate, Hugh Childers, with a comment on how successful he had been. The use of the secret ballot in Pontefract thought to promote tranquillity and peaceful conduct during polling, a substantial change from the rowdiness of earlier elections in general.

Up until 1872, polling was public. Electors were required to announce their votes, often in front of a gathered crowd, which were then recorded in the pollbook. Candidates and their agents therefore could compile an accurate record of vote choices, allowing them to target specific voters in upcoming elections via bribery, intimidation, and violence. But this state of affairs also allowed the disenfranchised to heckle and even attack voters whose allegiance they disagreed with, ‘punishing’ them for voting ‘wrong’. This rowdy behaviour was not solely limited to polling day. Elections pre-1872 also featured a Nomination Day prior to polling, where candidates would be nominated by their voting constituents. Non-voters would gather in crowds numbering into the tens of thousands and make their thoughts on candidates known through shouting and rioting (Lawrence, 2009, p. 45). The public nature of these events meant that elections often became a competition between the parties to see who could put on the best display for onlookers, from hiring circus acts to organising processions accompanied by marching bands (Crook and Crook, 2007, p. 452).

The public nature of these electoral contests often led to violence – and, at times, deaths. Richter (1971, p. 25) describes the fatal rioting observed as being the result of a ‘spontaneous outpouring of sheer ebullience’ brought on by the excitement and tension of crowded elections. But it was not only rioting that resulted in fatalities: one of the most shocking deaths recorded in the Victorian Election Violence Dataset (VEVD) occurred in Liverpool in 1837, when a Liberal supporter slit the throat of his 2-year-old daughter after finding her wearing the ribbon of the opposing candidate, gifted to her by a stranger (Exeter and Plymouth Gazette, 1837). The public nature of elections was blamed for producing these outpourings of overwhelming emotion, which was said to drive people to commit fatal violence. After 1868, following what was seen as a particularly and unacceptably violent election (Wasserman and Jaggard, 2006, p. 145), Gladstone’s government finally concluded that secret balloting was needed to curtail the public rowdiness of British elections.

This brings us back to the subject of the Punch cartoon. The Pontefract by-election was celebrated as a tranquil event that demonstrated the success of secret balloting. It was thought to have eradicated deplorable violence and fatalities from elections to the extent that residents hardly even knew an election was taking place (O’Leary, 1962, p. 86). At the following election in 1874, The Times asserted that the introduction of the secret ballot meant that ‘mobs, processions, favours, free fights, and punch-drinking have become for the most part things of the past’ (Hawkins, 2015, p. 285). Based on these observations, the answer should be clear: our elections became peaceful because of the introduction of the secret ballot, and removal of the crowded and festive atmospheres around nomination and polling day that had previously led to violence and fatalities.

However, the data unearthed by the Victorian Electoral Violence Project raises doubts. In fact, there were more fatalities in the 1880 election (7 in total) than had occurred in most of the General Elections before secret balloting. Even Pontefract, which had no history of electoral violence pre-1872 according to the VEVD, saw violent incidents at each of the next five elections after 1872. In 1874, miners and collieries damaged buildings and property; in both 1880 and 1885 the Conservative Club was stormed by a Liberal mob armed with sticks; in 1886 there was widespread window breaking after the poll was declared; and 1892, a tranquil election from a national perspective, saw numerous assaults on local police. Even if violence did decline nationally in post-1885 elections, this evidence makes it hard to argue that the secret ballot led to any decisive change in Pontefract. The successes that The Times and Punch attribute to it seem an exaggeration.  If crowds, candidates, and their agents could no longer identify voters’ intentions, then it should follow that targeted electoral violence aimed at influencing the outcome of the poll became unfeasible. So why did violence continue to occur?

I provide an answer by systematically counting and examining electoral fatalities. The Punch cartoon is not wholly wrong in implication: overall deaths on polling day did decline, and fatal rioting did become substantially less common. But the timing, location, and type of deaths occurring simply adapted to the new restrictions that the 1872 Ballot Act imposed. A substitution took place with deaths shifting from mass events to individual attacks, rather than there being any absolute decline in fatalities.

Figure 1: Proportion of mass versus individual violence fatalities over time

Figure 1 illustrates the substitution between deaths due to mass and individual violence around the introduction of the Ballot Act in 1872. ‘Mass violence’ deaths constitute fatalities due to rioting, or untargeted public disturbances. ‘Individual violence’ fatalities include targeted deaths perpetrated by an individual or small group, including deaths due to hired roughs and political quarrels. This difference is best exemplified by the 1868 Blackburn election. In the lead-up to polling, Thomas Whittaker was beaten and killed by a group of Radical Irishmen after shouting his support for the Conservative candidate while walking down the street (Preston Herald, 1868). The targeted nature of the attack and the small number of participants makes this an instance of individual violence. But the rioting caused by Whittaker’s death resulted in the killing of Patrick Gallagher, who was hit over the head by a policeman trying to settle the riot (Bristol Times and Mirror, 1868). Since this happened as part of a riot, Gallagher’s killing is recorded as being borne of mass violence.

Figure 1 uses election year as the measure of time on the x-axis, and percentage of overall fatalities as the quantity measure on the y-axis. The left-hand graph shows the proportion of overall fatalities that are recorded as mass violence for each election, whereas the right-hand graph shows the same for individual fatalities. Obviously, the 1872 Ballot Act triggered a change from predominantly mass violence to predominantly individual violence fatalities. Mass violence fatalities dropped by around 50 percentage points between 1868 and 1874 and remained low for the rest of the time period analysed. The increase in individual violence mirrors this pattern, creating a substitution effect in the cause of deaths.

This is mainly attributable to the fact that the secret ballot did not tackle the underlying causes of violence. Candidates employed underhand tactics, including violence, to increase their chances of winning: this is evident from the continued levels of bribery, treating, and other illegal practices observed (Hawkins, 2015, p. 285). Abolishing public nominations may have removed the large-scale gatherings and excitement that had previously triggered fatal riots, but it also meant that there was a need for a new form of public hustings that would let people (particularly the disenfranchised) question candidates. Parties rose to fill this gap by organising multiple smaller hustings around the constituency (Lawrence, 2009, p. 55). These meetings provided opportunities for more targeted acts of violence on a small scale, including the involvement of hired roughs. These people had long been relied on to perpetrate violence and intimidation against opposing voters. But since it was now impossible to reliably determine party allegiance on polling day, these roughs were more effectively deployed to monitor attendees and votes at explicitly partisan meetings and attack those supporting the opposition (Lawrence, 2009, p. 59). It is no coincidence that the only recorded death of a candidate in the dataset happened after the introduction of the secret ballot, when roughs stormed Raymond Lluellyn’s party meeting, arguably killing him in the process (Dundee Courier, 1886).

The shift in the use and focus of violence can also be observed in the timing of deaths.

Figure 2: Election Timepoint of fatalities over time

Figure 2 shows that while polling day fatalities declined after 1872, deaths during the campaign stage of elections became much more common. The x- and y-axis are the same here as in figure 1, but this time the graphs track the timing of the fatality during the election period. Whether it occurred in the month before polling, during polling, or in the month after.

Deaths on polling day had previously been dominated by rioting and mass violence because of the large crowds that gathered. This meant that minor violent incidents had a tendency to escalate into large-scale and fatal disturbances. By legislating to remove festivities from elections and prevent any ‘spectacle’ by making them private affairs, the Ballot Act largely removed the crowds from polling day and so removed the tendency for fatal mass violence to break out. But because of the new party-specific events and campaign-stage strategic violence discussed, fatal violence in the campaign stage of elections only became more common. Evidently, other factors were affecting the scale of the transition. This is clear from the fact that both campaign and polling stage deaths seem to occur at a similar rate post-1872, rather than matching the dramatic shift seen in Figure 1. Arguments for this are discussed in more depth in the research, but primarily come down to challenges in implementing legislation.

Thus, while contemporary commentators and legislators were convinced that the Ballot Act of 1872 would curb illegal practices, it only shifted the type and timing of electoral violence. The Pontefract by-election of 1874 was an exception rather than the standard: the tranquillity and peace experienced were not reflective of typical general elections after 1872. Instead of the public spectacles of Nomination and Polling Day, which offered fertile ground for triggering mass riots and disturbances, post-1872 elections saw increased violence at campaign events involving parties and candidates. Roughs used these as an opportunity to easily identify and subsequently target opposing voters. Without removing the motivations related to the use of violence during elections, the election period remained as dangerous and fatal as ever.

Lydia Buckroyd is an incoming third-year International Relations student at Durham University. She has worked on 19th-Century election-related deaths as an alternative to her planned year abroad, which was canceled due to the pandemic. More on her research into 19th-Century election violence deaths can be seen in the lecture on 19th-Century election violence she gave (together with Dr. Patrick M Kuhn) in the Royal Armouries Museum’s Winter Lecture Series at


British Newspaper Archive. Bell’s Weekly Messenger, 17th December 1832 [URL: Accessed 08/06/2021]

British Newspaper Archive. Bristol Times and Mirror, 7th December 1868 [URL: Accessed 18/06/2021].

British Newspaper Archive. Dundee Courier, 28th June 1886 [URL: Accessed 01/06/2021].

British Newspaper Archive. Exeter and Plymouth Gazette, 5th August 1837, [URL: Accessed 01/06/2021].

British Newspaper Archive. Preston Herald, 28th November 1868 [URL: Accessed 18/06/2021].

Crook, M and Crook, T 2007 The Advent of the Secret Ballot in Britain and France, 1789-1914: From Public Assembly to Private Compartment. History, 92(308): 449-471.

Hawkins, A 2015 Victorian Political Culture: Habits of Heart and Mind. Oxford: Oxford University Press. pp. 273-286.

Hoppen, T 1994 Grammars of Electoral Violence in Nineteenth-Century England and Ireland. The English Historical Review, 109(432): 597-620.

Kam, C 2007 Four Lessons about Corruption from Victorian Britain. In: Trabalho apresentado no workshop Corruption and Democracy, Vancouver, BC, pp. 1-18.

Lawrence, J 2009 Electing Our Masters: The Hustings in British Politics from Hogarth to Blair. Oxford: Oxford University Press.

O’Leary, C 1962 The Elimination of Corrupt Practices in British Elections 1668-1911. Oxford: Oxford University Press.

Parliament, House of Commons 1870 Report from the Select Committee on Parliamentary and Municipal Elections; Together With the Proceedings of the Committee, Minutes of Evidence, and Appendix. HC 1868-69, 352-I.

Richter, D 1971 The Role of Mob Riot in Victorian Elections, 1865-1885. Victorian Studies, 15(1): 19-28.

Richter, D 1971 The Role of Mob Riot in Victorian Elections, 1865-1885. Victorian Studies, 15(1): 19-28.

Salmon, P 2013 Elections before the secret ballot. The Victorian Commons blog.

Seymour, C 1970 Electoral Reform in England and Wales: The Development and Operation of the Parliamentary Franchise 1832-1885. Devon: David and Charles Reprints.

Wasserman, J and Jaggard. E 2006 Electoral violence in Mid-Nineteenth Century England and Wales. Historical Research, 80(207): 124-155. DOI:

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: