Researching Violence

Methodology and Measurement

Raymond M. Lee, Elizabeth A Stanko, Raymond M. Lee, Elizabeth Stanko, Elizabeth A Stanko, Raymond M. Lee, Elizabeth Stanko, Elizabeth A Stanko

• Publisher: Routledge
• Year: 2014
• Print ISBN:
  9780415301329
• eBook ISBN:
  9781317834854
• Topic: Ciencias sociales
• Subtopic: Criminología

Violence is a research topic that is fraught with difficulties. A notoriously sensitive subject, and one that is presumed to be largely hidden, researchers have long struggled with the question of how to measure its impact and how to explore its incidence. Arising from the ESRC's Violence Research Programme, Researching Violence is a practical guide both to theses problems and to the obstacles encountered when negotiating this uneasy terrain. Comprising the reflections of researchers who have worked on diverse projects - from violence in the home to racial violence and homicide - this book demonstrates the ingenuity and at times courageous actions of researchers having to think on their feet. It also investigates the ethical and emotional issues arising from working with the victims and perpetrators of violence. This book will be indispensable for students and academics doing research projects on violence.

Part I

Documenting violence

Counting and accounting for violence

1 Researching violence in the past

Quantifiable and qualitative evidence

John E. Archer

This chapter proposes to examine and review historical studies and debates of English interpersonal violence, especially those focused on homicide, and for the purposes of this chapter homicide includes both murder and manslaughter. Such studies have had to address problems and issues which confront all historians of violence; not least, can we quantify violence over long periods? Or is the accuracy of the statistical data so questionable as to invalidate any conclusions drawn from quantitative analysis? In reviewing possible drawbacks to this statistical quantifying approach to studies both of homicide and of non-lethal forms of violence, this chapter will examine alternative approaches which draw on qualitative data. Such material provides context and detail surrounding violent incidents which quantitative evidence cannot. Historians can also glean some insights into conflicting attitudes and perceptions of violent behaviour which, though classified as illegal by the authorities, may well have been regarded as wholly legitimate and customary by significant numbers of people within a community. In reviewing possible drawbacks to the statistical studies, this chapter will draw on evidence collected in a recent study on violence between 1850 and 1914 in the north-west of England.¹ This industrial and urbanised region was thought by contemporaries to be one of the most violent areas of the United Kingdom but, according to official statistics, it experienced a decline in reported and prosecuted violence in the second half of the nineteenth century.

Historians of violent crime have until recently attempted to address and answer the ‘big question’: was the past more violent than the present? A corollary to this was the additional question: did crimes of interpersonal violence fluctuate, were there highs and lows, and if so, why? Answers to the question ‘was the past more violent?’ have relied on and been constrained, to a large extent, by what has left a trace or ‘footprint’ in surviving documents and records. Court records for the period and, since the early nineteenth century, judicial annual statistical returns collected by parliament have provided historians with most of their sources of evidence. Quantifying incidences of criminal behaviour prior to the nineteenth century presents historians with many problems. First, it is as well to note that crime, processed by the judicial system, and recorded in the form of assize and quarter sessions indictments and calendars held at the Public Record Office or in the local county record offices, has left the most voluminous records and even then some areas of England and Wales are better provided than other areas. Lower level crime such as common assault would have been dealt with, if at all, at the local magistrates’ or petty sessions courts. Records emanating from such forums of justice have, in most instances, not survived.² From 1810 data relating to the number of males and females committed for indictable offences in England and Wales were collected and published annually by parliament. Throughout the nineteenth century various modifications, improvements and additions were made to these annual statistics – not least the inclusion of the numbers of women and men brought before magistrates on summary jurisdiction for less serious offences, and the numbers of indictable offences known or reported to the police, whether or not they resulted in a prosecution. Finally, further changes were introduced in 1893 which can complicate historians’ tasks in creating long-time series for criminal offences of various kinds. The most important change concerns the year in which the data were recorded and counted. For example, until 1893 police returns had a year end of 30 September, and prison returns ended on 31 March. Both sets of returns were then standardised and ran from 1 January to 31 December. Historians have, therefore, at their disposal a reasonably consistent set of national data on crimes known to the police dating from 1857. Thus comparisons and trends can be made for most crimes of interpersonal violence for the past 150 years, although it would be unwise – perhaps even foolish – of historians to place too much trust, meaning or significance in changes over time for some of these crimes. This is especially true of violence relating to women and children, for example, which was and is notoriously under-reported.

Quantifying homicide: debates on long-term trends

Historians have for the most part argued that one particular crime of violence, namely murder, can be quantified or counted reasonably accurately and over a long period. Comparisons with the present are thus possible and conclusions may be reached on the relative violence of specific periods in history. Whether the crime of murder can be taken as a measure of a society’s propensity to violence is, however, a moot point. The initial and arguably the seminal study on crimes of homicide was that put forward by Gurr (1981). He argued that of all data relating to incidences of violent crime, those which relate to homicide are probably the most accurate and fully reported. The more serious the crime, the greater the likelihood that it will be reported. Murder, he contended, possesses the smallest ‘dark figure’ of unrecorded crime. Moreover, most homicides have been committed by persons known to their victims, and as a consequence, clear-up rates for murder were and are relatively high. This high degree of reportage was, it has been argued, better than that for other crimes of violence, as sudden deaths have been reported and investigated by coroners since the Middle Ages (Havard 1960: 11–27). Historically, dead bodies have been investigated. Thus murder and manslaughter are more likely to have left an historical trace in the records. And, as a bonus for historians, the legal definition of murder has been relatively stable in England and Wales since the sixteenth century, allowing like to be compared with like.

Drawing on regional county studies by Hanawalt (1976) and Samaha (1974) for the medieval period, and Cockburn (1977) and Beattie (1974) for the early modern, Gurr was able to identify a long downward trend in homicides since the mid-fourteenth century. The rate, which is measured as the number of homicides per 100,000 of the population, declined from about 20/100,000 in 1200 to about 1.5/100,000 in 1800.³ Using the more accurate national figures from the nineteenth century, and also the more accurate census returns on which to base the ratio of murders to population, the murder rate continued to drop throughout the nineteenth century and the first half of the twentieth, reaching an all-time low of 0.3/100,000 in both 1931 and 1951. Only in the last 30 years of the twentieth century has the rate begun to creep up to just 1.0/100,000 in 1981. Gurr has, as a result, concluded that English homicide rates have declined dramatically from about AD1200 and that rates of violent crime were ‘probably 10 and possibly 20 or more times higher’ in medieval and early modern England than in the twentieth century (1981: 312).

While all historians accept that the homicide rate has declined over the past 500 years or so, not all have been in agreement as to the timing of this decline. In a short interpretative essay in Past and Present and a rejoinder, Stone (1983 and 1985) argued that the decline was ‘especially rapid between 1660 and 1800’. In response Sharpe (1985) accepted the downward trend in homicides but introduced a note of caution in so far as the measurement of homicide rates per 100,000 fails to tell us about how violence was perceived. Quantitative studies cannot, in Sharpe’s opinion, answer the more profound and challenging questions and issues such as whether society was fearful or not, whether it was conscious of living in a violent age, and just ‘how violent is a “violent society”?’(Sharpe 1985: 215).

When Cockburn (1991) entered the scholarly affray between Sharpe and Stone he did so not as a peacemaker but as a ‘powder monkey’, providing ammunition for both parties. His paper went far beyond his modest and self-effacing claims as it came to represent one of the most complete studies of homicide over a long period. Taking the county of Kent as his focus, Cockburn was able to present a continuous homicide rate from 1559 to 1971, and add a number of refinements, not least the changes in methods of killing and the use of weapons over time. Moreover, he confronts difficult methodological problems which affect all historical studies of homicide. Do we, for example, count in infanticide and concealment of births? Should attempted murders be included in the later decades under scrutiny since improvements in medicine have led to better survival chances? And what of unlawful killings on the public highway? Some of these questions, according to Eric Monkkonen, are capable of resolution. He has estimated that 50 per cent of New York’s victims of homicide, for example, would have been saved from death nowadays due to improvements in medicine (2001b: 18). The upshot of this empirical research is that English and Welsh society by the turn of the millennium would appear to be indisputably less violent than it was, say, 300 years ago, if murder is taken as the yardstick to measure violence. Most historians would argue that the decline in murder in particular, and in homicide in general, which shows up in the surviving records, represents a real decline as many factors have existed since the nineteenth century which would have increased reportage of violent deaths. The introduction of the ‘new’ police from 1829 onwards, the growing ‘sensitisation to violence’ and the necessity of registering deaths from 1837 are just three examples which would have made reportage more probable.

This conclusion, based on statistical data, that the present is less violent than the past, may seem somewhat commonplace to today’s students. However, Gatrell, in one of the most influential and thought-provoking studies for 20 years observed that ‘it has become one of the more widely accepted axioms of our age that an increasing crime rate is the invariable price of material progress’ (1980: 238). Through the analysis of official statistics issued between 1857 and 1914 Gatrell has concluded that all forms of interpersonal violence, sexual violence excepted, declined in the second half of the nineteenth century (1980: 286). This conclusion also held good for property crime, and consequently historians have come to regard the second half of the nineteenth century as experiencing an ‘English miracle’. It would appear that ‘modernisation’ or progress need not necessarily result in increased criminality.

In what has come to represent the most profound challenge to the reliability of quantitative data, Howard Taylor, in ‘Rationing Crime’ (1998) has offered a radical revision of the accepted historical consensus on the declining homicide rate of England and Wales. In examining the official statistics of the higher courts, Taylor noted a peculiar pattern. After a huge increase in the number of trials for indictable crime between 1805 and 1842, the growth stopped thereafter until the twentieth century. Offences known to the police, which were collected from 1857, remained constant, varying by only 20 per cent either side of the norm, until 1925 when they began to move upwards. This stagnation in crime, in both recorded and prosecuted crime, occurred in the context of national population growth, which meant that in effect crime and prosecutions were actually falling. This was all the more remarkable given that the country had police forces in every town and county whose job it was to record and investigate crime.

How does this relate to the study of homicide? Taylor found a similar pattern for murder in England and Wales. Between 1880 and 1966 murders averaged 150 a year with a variation within 20 per cent either side of the average. This pattern would appear to hold good with one or two exceptions back to the mid-1860s (1998: 569–71). Explanations for this surprising and, some would argue, shocking pattern lay, according to Taylor, in the costs of investigation and prosecution. He goes so far as to argue that murders were ‘among the most strictly rationed of all crimes’, and that consequently ‘most murders and suspicious deaths went uninvestigated’ (1998: 588). Apart from calling into question the veracity of official statistical data on even the most reported of all crimes – murder – Taylor’s thesis also brings to our attention the way in which official criminal statistics were constructed and how homicides and sudden deaths were examined and decided upon. An initial reading of Taylor would suggest that some conspiracy involving chief constables, coroners and the office of the Director of Public Prosecutions was involved. Other questions arise out of his argument: not least, was the rise in suicides from 1,314 in 1856 to 5,054 in 1939 related in any way to the contemporaneous stagnation or decline in murders? Was a finding of suicide a disguised murder or secret homicide? He cites a number of medical practitioners who certainly thought this to be the case. And what is so significant about this figure of 20 per cent either side of the average? Surely such an enormous variation actually allows for wide differences in the amount of reported crime. The press certainly become very excited nowadays if crime rises by just 5 per cent over a previous year, and the government is quick to claim credit for single-figure percentage falls in crime.

In one of the first responses to Taylor, Robert Morris, in an article which plays on Disraeli’s aphorism for its title, ‘Lies, Damned Lies and Criminal Statistics’ (2001), picks up on the 20 pe...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Figures and tables
• Contributors
• Introduction: methodological reflections
• Part I: Documenting violence: counting and accounting for violence
• Part II: Enhancing data on violence
• Part III: The impact of institutional contexts for the study of violence
• Index