Methods of Criminology and Criminal Justice Research

Mathieu Deflem, Derek M.D. Silva, Mathieu Deflem, Derek M.D. Silva

• Publisher: Emerald Publishing Limited
• Year: 2019
• Print ISBN: 9781787698666
• eBook ISBN: 9781787698673
• Topic: Social Sciences
• Subtopic: Criminology
• Index: Social Sciences

The eleven chapters in this volume of Sociology of Crime, Law and Deviance discuss a variety of issues of methodological significance in research in the fields of criminology and criminal justice studies. As scholarly work on various aspects of crime, deviance, criminal justice, and social control has progressed tremendously in recent decades, both in terms of scope as well as with respect to theoretical approaches, the employed methods of investigation have also broadened and advanced to be as sophisticated as those used in any other area of contemporary social-science inquiry. The authors in this volume demonstrate the methodological maturity and diversity of current empirical research in criminology and criminal justice in a number of areas, such as general trends of crime, criminal networks, violence against women, sex work, elder financial exploitation, school safety, immigrant detention, extremism on the internet, and human trafficking. 

Presenting a state-of-the-art overview of criminological and criminal justice methodologies today, this book is of interest to a wide range of scholars and students in the fields of criminology, sociology, justice policy, and criminal justice.

PART I

GENERAL PATTERNS AND TRENDS

CHAPTER 1

IS CRIME RISING OR FALLING? A COMPARISON OF POLICE-RECORDED CRIME AND VICTIMIZATION SURVEYS

Barak Ariel and Matthew Bland

ABSTRACT

Purpose – Statistics about the level of crime continue to attract public and political attention but are often presented in conflicting ways. In England and Wales, police-recorded crimes are no longer considered “national statistics” and, instead, the crime survey of England and Wales (CSEW) is used. However, it is not clear why partial population data (e.g., police-recorded crime) are considered less reliable or valid for measuring temporal crime trends in society than inferential statistical estimation models that are based on samples such as CSEW. This is particularly the case for approximating rare events like high-harm violence and specific harmful modus operandi (e.g., knife crime and firearms). In this chapter, the authors cross-reference victim survey and police-recorded data to determine similarities and contradictions in trends.

Methods – Using police data and CSEW estimates, the authors contrast variance and logarithmic trend lines since 1981 across a range of data categories and then triangulate the results with assault records from hospital consultations.

Findings – Change in crime rates in recent years is neither as unique nor extreme as promulgated in media coverage of crime. Moreover, analyses show conflicting narratives with a host of plausible but inconclusive depictions of the “actual” amount of crime committed in the society. The authors also conclude that neither source of data can serve as the benchmark of the other. Thus, both data systems suffer from major methodological perils, and the estimated crime means in CSEW, inferred from samples, are not necessarily more valid or accurate than police-recorded data (particularly for low-frequency and high-harm crimes). On the other hand police-recorded data are susceptible to variations in recording practices. As such, the authors propose a number of areas for further research, and a revised taxonomy of crime classifications to assist with future public interpretations of crime statistics.

Originality – There is much public and academic discourse about different sources of crime measurement yet infrequent analysis of the precise similarities and differences between the methods. This chapter offers a new perspective on long-term trends and highlights an issue of much contemporaneous concern: rising violent crime.

Keywords Rising crime; victim surveys; recorded crime; violent crime; crime statistics; police records

INTRODUCTION

Is crime going up or down in society? This question is frequently the subject of public debate, being of particular interest to politicians and police professionals alike (Mercer, 2019). Being able to answer this question accurately is also essential for policing scholars who seek to identify appropriate targets (Sherman, 2013), as well as for scholars interested in research methods and crime trajectory analyses (Farrington, 1979; Jennings, Gray, Hay, & Farrall, 2015). However, temporal patterns – that is, what are often referred to as “trends” in crime volume and type – are difficult to measure and interpret (Nguyen & Loughran, 2018). For example, is an arithmetic increase in crime volume during a particular time period a genuine upward crime trend? The answer is not simple, as the increase can be explained through other theories: crime-recording practices, crime reporting patterns, or changes in the definitions of crime, which are all plausible explanations vis-á-vis the case for explaining an arithmetic increase as an “actual” change in crime volume or type.

One major reason for the difficulty in interpreting crime figures is that the ways in which they are measured, or estimated, determine the outcome. From a methodological perspective, when crime levels are dependent on the data collection methodology, it raises suspicions about the integrity of claims about increases or reductions in crime levels over time. Put differently, if the measurement tool “causes” the fluctuations in the figures, we experience statistical noise, to a point that we may not trust the measurement at all.

Given these concerns, in England and Wales, police-recorded crimes were recently declared to be invalid national statistics of crime. Statisticians made the case that police-recording practices affect crime records to such an extent that we can no long trust them. This is a logical deduction, because police crime figures largely depend on the extent to which the police deal with crime and record it accurately. Instead, national crime statistics have been “replaced” by the Crime Survey of England and Wales (CSEW). At least for the Home Office, the CSEW is a superior instrument, because it directly surveys victims. Statisticians who use CSEW suggest that they are the best available tools to estimate from a large sample onto the populations of England and Wales. Therefore, the question about the rise or fall of overall crime in England and Wales is broadly answered based on the evidence presented in the CSEW.

However, the CSEW is not infallible either. The survey systematically excludes crime categories and subgroups of participants who are likely to be the most vulnerable to become victims or commit crimes: the young, the homeless, the transient, and the non-respondents. It also excludes crimes against businesses and those where there is no apparent personal victim. The CSEW is, after all, a survey of a sampled subset of households in society. By definition, a survey requires sampling techniques, estimation models, and statistical correction to account for various sampling errors and “design effects”. Therefore, the CSEW population estimates miss much of the low-frequency but high-harm crimes. Violence estimates are a case in point. These and other methodological issues underscore the problems that the CSEW faces, and for at least some crime categories, it ought not to be blindly accepted that a national survey of 35,000 households is the “best practice” for understanding crime patterns. Further, it is not necessarily the appropriate yardstick against which the validity and reliability of police-reported crime statistics ought to be evaluated.

One solution is to triangulate data with a third source of information. For example, National Health Services (NHS) hospital records can be contrasted with the two data sources (police records and CSEW). It is often argued that the NHS data are not affected by police crime-recording practices, nor are they susceptible to victim survey methodology flaws (Sivarajasingam et al., 2010). Therefore, hospital admission records give indications about the validity of the data sources, as well as the direction of crime trends. Naturally, hospital population data may only be useful for against-person crime figures that resulted in injury, but the data also provide additional evidence on crime patterns in society.

With these contemplations in mind, this chapter has a modest aim: to compare publicly available, population-level data from both the police as well as the CSEW, and to then contrast these “data narratives” with NHS records. Particular emphasis will be given to the argument that all crime statistics fail to reflect the actual amount of crime committed in society.

The chapter has the following structure: first, the study of temporal crime patterns is introduced, with a presentation of the two major data sources for the study of crime trends: police records and victim surveys. Unofficial versions of data collection instruments – localized surveys of offenders, victims, places, etc. – are purposely excluded, as these do not establish national crime policies or lead the political discourse on the necessary response to crime like official statistics do. For each source, the methodologies used to collate data are explored, with particular attention paid to the concerns of which source is exposed (reliability and validity). Next, the procedures used in this chapter are laid out, with our focus being on temporal trend analyses and comparisons between the datasets. The analysis is then followed by a discussion of the results in the context of policy, as well as research recommendations for future research.

POLICE-RECORDED CRIME VERSUS ESTIMATES FROM VICTIM SURVEYS

There are two major sources for crime count statistics: victim surveys and police-recorded crimes (Maguire, 2002, chapter 3; Skogan, 1984; Wittebrood & Junger, 2002). For a broader review of these two instruments, see Newburn (2018). At least in England and Wales, statisticians have argued for the superiority of the CSEW over police-recorded data (Home Office, 2018b); as discussed in more detail below, the UK Statistics Authority made the case that police records do not provide accurate estimates of crime reported to the police. On the other hand, it is not immediately clear why we accept this argument unconditionally (Baumer & Lauritsen, 2010). Relying on surveys creates at least as many challenges to the validity and reliability of the data as are in police records (Home Office, 2018a). Thus, we review the prominent difficulties of these data sources below.

Methodological Issues with Police-recorded Crimes

There are a number of hazards to police-recorded data protocols, and we will focus on two primary issues – recording practices and insufficient differentiation between police-generated and victim-generated crimes.

Police records are notoriously susceptible to changes in recording practices (MacDonald, 2002). The Greater Manchester Police (GMP) is a case in point; for a certain period of time, the GMP –the second largest force in the country – was excluded from national police records due to challenges with their crime-recording practices (ONS, 2017). Therefore, recorded crime statistics are affected by recording practices – which makes them more vulnerable to threats of validity and reliability.

To increase the reliability – that is, the consistency of recording – the Home Office Counting Rules (HOCR) were introduced to govern crime classification practices. The HOCR was established in 1998 and later supplemented by the National Crime Recording Standard (NCRS) in 2002. These guidelines are aimed to increase the consistency of crime recording between English and Welsh forces. They also sought to establish a victim-oriented approach in crime statistics, in which every victim of a crime event was allocated an individual crime record – a policy colloquially known as “one victim, one crime”. The result was greater numbers of crimes recorded for multiple victim events and a rise in overall crime, accordingly.

Still, Her Majesty’s Inspectorate of Constabulary and Fire & Rescue Services (HMICFRS) conducted reviews into the way police forces recorded crimes in 2009 and 2014, and while there were improvements over the years, recording issues were nearly always detected. Consequently, the UK Statistics Authority (House of Commons Public Affairs Select Committee, 2014) assessed police records, and concluded that the “underlying data on crimes recorded by the police may not be reliable.” Thus, the “National Statistics” quality badge was removed from police-recorded crime data, with the most concerning issue being the under-recording of certain crime categories by the police (HMICFRS, 2014): more than 20% of crimes that received police attention did not get recorded as such (for instance, violence with or without injury – 33%). Immediately after the 2014 report, the HMICFRS announced that it would conduct an “unannounced programme of rolling inspections of crime recording on an ongoing basis.” The results of these inspections are periodically published on the HMICFRS’s website.

Furthermore, a clear distinction between police-generated and witness/victim-generated records is required, but th...

Table of contents

• Cover
• Title
• Introduction: Measuring Crime and Criminal Justice
• Part I. General Patterns and Trends
• Part II. Special Groups and Problems
• Part III. Crossing Boundaries
• Index