Ethnic Inequalities in Sentencing: Evidence from the Crown Court in England and Wales

Abstract

In recent years, there has been considerable policy and academic interest in the existence of ethnic inequalities in the Criminal Justice System. A large body of sentencing research has been dedicated to exploring whether ethnic minority defendants are treated more harshly than similarly situated white defendants. This paper extends this research utilizing Ministry of Justice linked criminal justice datasets and multilevel models to assess the effect of ethnicity and other defendant case and contextual factors on sentencing outcomes in the Crown Court. The analysis shows that legal characteristics such as plea, pre-trial detention, offence type and severity are important factors determining sentencing outcomes although they do not fully explain disparities in these outcomes between ethnic groups. Ethnic disparities in imprisonment persist and, in some cases, become more pronounced after controlling for defendant case and court factors. In contrast, ethnic disparities in sentence length are largely explained by legal factors, and after adjusting for other predictors of sentencing outcomes, observed differences between most (but not all) ethnic minority groups and the white British disappear.

INTRODUCTION

In recent years, there has been considerable policy and academic interest in the existence of ethnic inequalities in the way people are treated in society (Jivraj and Simpson 2015; Cabinet Office 2017; Byrne et al. 2020). Following a series of high-profile events such as The Black Lives Matter global protests, sparked by the killing of African American, George Floyd in the United States, ethnic disparities in policing and the Criminal Justice System (CJS) have come into the forefront of public and political debates. The 2017 Lammy Review, commissioned by two UK governments, presented evidence of stark ethnic inequalities at all stages of the CJS. From the point of arrest, through prosecution to custodial remand, sentencing and imprisonment, ethnic minority groups were shown to be both disproportionately represented and to experience disproportionately worse outcomes. Ethnic disproportionality in criminal justice outcomes has been documented in subsequent government reports with the most recent report (Ministry of Justice 2021) showing that ethnic minority defendants were between 4 and 28 per cent more likely to be remanded in custody and to have a consistently higher average custodial sentence length (ACSL), than white defendants. In 2020, over a quarter of prisoners were from ethnic minority groups, and a third of children in prisons were from black groups despite black prisoners accounting for just 13 per cent of the entire prison population. These observations are telling, but they are also insufficient in understanding the causes of ethnic disparities. This was the motivation behind the call to ‘explain or reform’ in the Lammy review (2017), that CJS agencies should provide evidence-based explanations for ethnic disparities or introduce reforms to address them.

For decades, scholars have been concerned with the causes of ethnic disparities in the CJS and whether these disparities arise from discriminatory treatment of ethnic minorities. In the large body of research on sentencing, no question has received more attention than whether ethnic minority defendants are treated more harshly than similarly situated white defendants (King and Light 2019, p. 366). There have been numerous studies examining racial and ethnic disparities in sentencing in the United States (Steffensmeier et al. 1998; Engen and Gainey 2000; Spohn 2000; Bushway and Piehl 2001; Johnson 2006; Feldmeyer and Ulmer 2011; Ulmer et al. 2011), and although the findings of these studies have varied, the available evidence suggests that ethnic minority defendants are more likely to receive harsher sentences than white defendants, even when legally relevant factors are taken into account. Evidence on the association, independent of other factors, between ethnicity and sentencing outcomes outside the United States is sparse. In the United Kingdom, for example, hardly any research has been carried out in sufficient depth to examine ethnic disparities in sentencing. While some studies (Hood 1992; Feilzer and Hood 2004; May et al. 2010; Hopkins et al. 2016) have examined the role of ethnicity in the decision to imprison or not, there is little evidence about the degree of unwarranted disparities at different stages of the sentencing process. Moreover, much of the focus of previous studies has been on sentencing differentials between aggregated and heterogeneous ethnic groups, namely the ‘white’, ‘mixed’, ‘Asian’, ‘black’ and ‘Chinese or other’ groups, which fails to acknowledge variations among constituent groups comprising these ethnic groups.

In light of this, the study uniquely examines the relationship between ethnicity and sentencing outcomes, considering imprisonment, sentence length and disaggregated self-identified ethnic groups. It investigates ethnic differences in sentencing outcomes whilst allowing for extra-legal factors (such as age, gender, deprivation) as well as important legal (case) factors (such as plea, remand status, offence type and severity). The research assesses how the effect of sentencing varies across ethnic groups after controlling for all other variables. It therefore examines the difference in the measured effect of sentencing between the various ethnic minority groups compared with the white British group once all other variables have been controlled. Unlike previous studies, which have overlooked contextual factors, this research examines between-court variation in outcomes and the influence of court contextual factors. Utilizing unique data on defendant appearances in magistrates’ courts and the Crown Court created through the Data First (DF) data linking programme led by the Ministry of Justice, the study aims to deepen understanding about the factors associated with sentencing outcomes in courts in England and Wales and determine the extent of ethnic disparities independent of other factors.

THE EFFECT OF ETHNICITY ON SENTENCING

There is a significant body of research suggesting that ethnic disparities in sentencing exist in the CJS. This research has been dominated by US studies which have shown that there are significant ethnic inequalities in terms of the severity of sentences received, particularly the likelihood of receiving a prison sentence and that these inequalities disproportionately affect Black African and Latino groups.¹ For example, Chiricos and Crawford (1995) reviewed 38 studies published between 1979 and 1991 and found substantial evidence of a direct impact of race on imprisonment, after controlling for other factors that could affect sentencing such as the severity of the offense and the defendant’s prior criminal record. The review did not find evidence that race impacts on sentence length. Spohn’s (2000) review of 40 sentencing studies in the 1980s and 1990s also evidenced a direct race effect on the likelihood of imprisonment highlighting that Black and Hispanic defendants were more likely to be sentenced to prison, and in some jurisdictions they also received longer sentences than their white counterparts. She found that race/ethnicity combined with other extra-legal factors such as age and gender to produce greater sentence disparities while process related factors such as pre-trial detention, plea and legal representation also accounted for ethnic disparities. Mitchell’s (2005) meta-analysis of incarceration and sentencing decisions across 71 studies also found a small but variable effect of race on sentencing decisions. He explained inconsistent findings across studies by suggesting that the variation in results was due to differences in study design and measures, the type of crime being considered and jurisdictional location. He concluded that even in studies that used more rigorous research designs, controlled for relevant defendant and case confounding variables, unwarranted racial disparities persisted.

There is some limited evidence from the United Kingdom on the relationship between ethnicity and sentencing outcomes suggesting that ethnic minority defendants are more likely to receive prison sentences than white defendants. Hood’s (1992) influential study of 2,884 individual defendants sentenced in West Midlands courts showed that after controlling for other variables that explain imprisonment, defendants from black ethnic minority groups had a 5 per cent greater probability of being sentenced to prison than defendants from the white group and that black defendants received longer sentences, on average, than white defendants for similar offences. The study also found that ethnic disparities in sentencing were more pronounced for certain types of offences, such as drug offences, and that the disparities were greater in some geographic areas than in others.

Feilzer and Hood’s (2004) study based on 17,054 case decisions from eight Youth Offending Team (YOT) areas found that minority ethnic young people were more likely to be referred to the youth justice system and to receive custodial sentences than white young people even after controlling for other characteristics. The study also found that minority ethnic young people were more likely to be subject to certain types of sanctions, such as being placed on a supervision order or being required to participate in community service.

May et al’.s (2010) study in twelve YOT areas also showed local variation in sentencing outcomes and ‘some evidence that at some stages of the youth justice system there may be discrimination against ethnic minorities’ (May et al. 2010, p. vi). A more recent study by the Ministry of Justice (Hopkins et al. 2016) based on 21,000 defendants convicted of indictable offences in the Crown Court, showed that the odds of imprisonment were 53 per cent, 55 per cent and 81 per cent higher, for defendants in the black, Asian and Chinese or other groups, respectively, compared to white defendants. The study also showed that within drugs offences, the odds of imprisonment were around 240 per cent higher for ethnic minority defendants, compared to those from a white background.

EXPLANATIONS OF ETHNIC DISPARITIES IN SENTENCING

The causes of ethnic disparities within the CJS are complex and multifaceted, and are commonly explored through theories of discrimination or judicial decision-making.

The dominant theoretical framework used by contemporary studies of ethnicity and sentencing is the focal concerns theory, which posits that sentencing decisions result from complex processes that weigh factors such as ‘offender blameworthiness’, ‘victim harm’, community protection and practical implications (Steffensmeier et al. 1998; Steffensmeier and Demuth 2000, 2001). Within the focal concerns framework, both legal and extra-legal factors play pivotal roles in influencing sentencing outcomes. Legal factors, including criminal history, offence type, severity and plea, act as aggravating or mitigating circumstances. For instance, the severity of the conviction offence and the defendant’s criminal history are considered aggravating factors, leading to higher rates of imprisonment and longer sentences (Steffensmeier and Demuth 2000, 2001). Conversely, mitigating factors, such as guilty pleas and the absence of prior convictions, tend to lessen the severity of sentences (Roberts 2011).

The focal concerns framework also incorporates discrimination theories, suggesting that ethnic and racial disparities may arise from discriminatory practices by law enforcement officials. Studies indicate that phenotypical features, such as darker skin, are associated with harsher sentencing outcomes, implicit biases and racial stereotyping among judges (Albonetti and Hepburn 1996; Rachlinksi et al. 2009; King and Johnson 2016). These stereotypes may result in certain ethnic groups being viewed as more dangerous and blameworthy, leading to disparate sentencing outcomes (Williams 2015; Franklin and Henry 2019). Stereotypes about race and crime can manifest at earlier stages of the CJS during character and moral assessments by probation and police officers, leading to higher pre-trial detention rates for ethnic minority defendants (Bridges and Steen 1998).

Focal concerns theory integrates findings from discrimination theories by proposing that discriminatory practices result from a ‘perceptual shorthand’ adopted by judges when faced with limited information about the culpability or dangerousness of defendants (Steffensmeier et al. 1998). Subjective assessments of focal concerns make it likely that stereotypes and biases based on race/ethnicity influence the sentencing process, signalling a greater degree of dangerousness and a higher risk of recidivism for specific ethnicity-age-gender groups (Demuth and Steffensmeier 2004; King and Light 2019).

Within the focal concerns framework contextual factors associated with courts and the broader communities in which they operate can also shape sentencing outcomes. The ‘court communities’ perspective proposes that the effects of race and ethnicity on sentencing may vary between courts due to distinctive and localized organizational, political and legal cultures, with studies showing that variations in sentencing depend on court caseload and sentencing severity (Johnson et al. 2008; Ulmer et al. 2011). The effects of race and ethnicity on sentencing may also vary between courts due to differences in the ethnic composition of the communities they serve (Feldmeyer and Ulmer 2011). This perspective aligns with the racial threat hypothesis, suggesting that the size of the ethnic minority population influences discriminatory practices, leading to inter-group conflict, prejudice and discrimination towards ethnic minorities (Blumer 1958; Blalock 1967). Sentencing decisions in areas with a high concentration of ethnic minority groups would be expected to be less favourable towards these groups compared to areas with smaller ethnic minority populations (Steffensmeier et al. 1998; King et al. 2010; Feldmeyer and Ulmer 2011).

Complementing focal concerns theory, alternative explanations, such as the disparate impact hypothesis, highlight the unintentional consequences of seemingly neutral policies or practices disproportionately affecting certain ethnic groups. In the United States, disparate impacts are associated with drug sentencing laws that impose more severe sentences for certain types of drugs offenses (e.g. crack cocaine), disproportionately affecting black defendants. In the United Kingdom, the application of sentencing decisions, such as Joint Enterprise, has led to unequal treatment of ethnic minority groups, often associated with the prosecution of suspected gang members (Lammy 2017; Williams and Clarke 2018). The plea-dependent (guilty/not guilty) sentence differential also disproportionately impacts ethnic minority defendants due to factors such as lower levels of trust in the justice system, a preference for jury trials over judge trials and poor legal representation (Kellough and Wortley 2002; Shute et al. 2005; Tata and Gormley 2016; Lammy 2017).

Moreover, pre-trial detention, which represents a fundamental denial of freedom for defendants not yet proven guilty (Kellough and Wortley 2002), can impact outcomes at plea bargaining and subsequent sentencing stages. Ethnic minorities may be overrepresented in the CJS due to other factors, including higher offending rates linked to socioeconomic disparities, concentrated poverty and poor social organization (Sampson 1987; Lymperopoulou and Finney 2016). These disadvantages increase the likelihood of ethnic minority defendants facing challenges such as failing to appear in court, inability to afford bail and higher rates of pre-trial detention (Phillips and Bowling 2003; Leslie and Pope 2018). It is also increasingly recognized that ethnic minority groups are subject to cumulative disadvantage, a process that involves the compounding effects of forms of disadvantage over time, as well as the accumulation of multiple and interconnected disadvantages which interact to deepen punishment for some groups (Kurlychek and Johnson 2019).

In line with other empirical studies to date this paper seeks to establish if there are significant ethnic disparities, net of other well-established predictors of sentencing outcomes (Baumer 2013). To this end, the research presented in this paper examines a range of extra-legal, legal and contextual factors and their association with sentencing outcomes and seeks to determine whether imprisonment and sentence length outcomes are different for ethnic minority defendants after controlling for other factors affecting these outcomes.

DATA AND METHODS

The study draws on magistrates’ and Crown Court datasets developed through the DF programme led by the Ministry of Justice.² All criminal cases in the United Kingdom start in a magistrates’ court which is presided by magistrates or a district judge. Magistrates’ courts deal mainly with summary (less serious) offences with more serious offences sent to Crown Court for trial or sentencing. The Crown Court is a single entity, operating from 71 courts across England and Wales and deals with serious criminal cases and appeals sent for trial or sentencing by magistrates’ courts.³

A common challenge in research focussing on ethnicity is having insufficient number of cases relating to ethnic minority groups. Small sample sizes and missing information on ethnicity affect the conclusions that can be drawn from data and have restricted previous analyses of ethnic disparities to aggregated, and often very heterogeneous, ethnic groupings (Cabinet Office 2017). While ethnicity reporting in the CJS started at the end of the 1990s in the United Kingdom it has been inconsistent since information on ethnicity has primarily been collected for defendants charged by the Police. For this reason, coverage of self-identified ethnicity tends to be less complete for less serious offences.⁴ To overcome some of these limitations, the analysis presented in this paper draws on four years of data (2017–2020) and excludes summary offences.

MEASURES

Dependent variables

The study examines two sentencing outcomes for defendants appearing in the Crown Court between 2017 and 2020. The first outcome variable is a measure of whether a defendant convicted in the Crown Court received a custodial sentence (imprisonment) (1 = yes and 0 = no). The second outcome variable is the sentence length (in months) for defendants with a custodial sentence. Since the sentence length variable is positively skewed this variable has been log-transformed.

Independent variables

Age and gender

The age of the defendant at the time of the appearance in Crown Court. Age is broken down into four age categories: under 20, 20–29, 30–50 and over 50. The models include a dummy variable for male defendants.

Ethnicity

Ethnicity is based on the 16 + 1 self-identified ethnic group classification from the 2001 Census. Defendants with unknown ethnicity have been excluded. Ethnic minority groups refer to ethnic groups other than the white British.

Income deprivation

Each defendant’s Lower Super Output Area (LSOA) of residence was matched to the income domain of the 2019 Index of Multiple Deprivation. LSOAs are Census geographies, commonly used as approximations of neighbourhoods, defined on the basis of homogenous population size (approximately 1,500 people) and household characteristics. This variable identifies whether a defendant lives in one of the 20 per cent most income deprived LSOAs in England and Wales.

Not guilty plea

The plea variable indicates whether a defendant pleaded not guilty based on the overall plea rank for the defendant’s case (1 = yes and 0 = no). If a defendant pleads guilty early on in a case, this often results in a discounted sentence length or a smaller likelihood of receiving a custodial sentence.

Offence type

This is derived from the 12-fold Home Office offence categorization and relates to the offence flagged as the most serious disposal excluding summary motoring and summary non-motoring offences.

Co-defendants

Defendants in cases with other defendants (1 = yes and 0 = no).

Remanded in custody

Defendants remanded in custody when they were committed to the Crown Court (1 = yes and 0 = no).

High severity offence

High severity offences have been identified using the 2020 Cambridge Crime Harm Index (CCHI) (Sherman et al. 2016) which is based on the number of days in prison suggested in England and Wales sentencing guidelines for the starting point sentence for an offence. 87 offences with a starting point sentence of three years (>1,000 CCHI score) have been classified as high severity offences (1 = yes and 0 = no).

Previous convictions

Previous convictions measure whether a defendant has been convicted in cases (other than for summary offences) in any of the years captured by the magistrates’ (2011–2020) and Crown Court (2013–2020) datasets (1 = yes and 0 = no).

Court case workload

Volume of cases (including appeals) processed across all years. Courts have been classified according to their case workload into three categories: high (greater than 0.5 standard deviations above the mean), medium (between 0.5 standard deviations above and below the mean) and low (0.5 standard deviations below the mean) workload.

Average case severity

The average case severity variable captures the average severity of offences processed by the court, and thus the caseload’s potential to produce severe sentences. This is measured by the per cent of convictions out of all cases processed in each court.

Ethnic density

Ethnic density is measured by the proportion of ethnic minorities in the local police force area of each court based on estimates from the 2011 Census.

In line with previous studies carried out in the United States over the last decade, the study utilizes a multilevel modelling approach (Hox et al. 2017) to examine ethnic disparities in the likelihood of imprisonment and sentence length. The modelling strategy involved estimating two-level models of sentencing to account for the clustered nature of data (individual defendants and cases nested within courts) and to investigate sources of variation within and between clusters. Multilevel modelling has emerged as a prevailing approach for investigating ethnic and racial disparities within focal concerns due to its capacity to analyse hierarchical data structures (Johnson 2006). It enables researchers to simultaneously consider contextual effects that shape disparities while controlling for relevant case characteristics which is ‘crucial for studying the emergence of unwarranted disparities across courts’ (Pina-Sánchez and Grech 2018; p. 530). This approach is beneficial for capturing the complexities of the legal system, where decisions are not only influenced by individual case attributes but also by broader contextual factors, legal norms and institutional practices as suggested by court communities’ perspectives.

Results from hierarchical logistic regression models of imprisonment and hierarchical linear models (HLM) of sentence length are presented which include defendant characteristics (Model 1), defendant and case characteristics (Model 2) and defendant, case and court contextual characteristics (Model 3). In light of the dearth of studies that investigate ethnic disparities in sufficient detail, particularly the lack of research examining disaggregated ethnic groups and given the detailed information on case characteristics afforded by the newly available criminal justice linked datasets, the focus of the analysis is on the direct effects of ethnicity on sentencing outcomes and the inclusion of additional controls of defendant, case and court characteristics. While acknowledging the importance of indirect effects of ethnicity and other case factors on sentencing, in this paper we prioritize the exploration of direct relationships to identify the factors influencing sentencing outcomes and determine which ethnic subgroups are subject to harsher sentencing in England and Wales.

The sample means and proportions of the data used in the estimation are shown in Table 1.

Table 1.

Sample means and proportions, defendants in the Crown Court datasets, 2017–20

	Imprisonment	Custodial sentence length
Sentence length (months)*		29.9 (38.2)
Imprisonment	0.544
Remand
Ethnicity (ref:White British)	0.712	0.707
Indian	0.010	0.010
Pakistani	0.029	0.029
Bangladeshi	0.011	0.011
Other Asian	0.019	0.018
Black African	0.042	0.041
Black Caribbean	0.038	0.040
Other Black	0.035	0.036
White and Asian	0.003	0.002
White and Black African	0.004	0.004
White and Black Caribbean	0.017	0.018
Other Mixed	0.009	0.008
Chinese	0.002	0.002
Other	0.013	0.013
Other White	0.049	0.052
White Irish	0.008	0.009
Age (ref: under 20)	0.094	0.022
20–29	0.375	0.393
30–50	0.434	0.492
Over 50	0.096	0.093
Gender:Male	0.908	0.934
Income deprived	0.446	0.465
Offence (ref:Other)	0.434	0.443
Drug offences	0.206	0.192
Fraud offences	0.048	0.046
Possession of weapons	0.083	0.072
Robbery	0.041	0.053
Violence against the person	0.188	0.194
High severity offence	0.094	0.114
Previous conviction	0.585	0.677
Pre-trial remand	0.402	0.588
Co-defendants	0.145	0.144
Plea not guilty	0.115	0.150
Court workload (ref:low)	0.398	0.396
Medium	0.310	0.307
High	0.292	0.297
Court case severity (%)*	48.3 (4.1)	48.7 (3.9)
Ethnic density (%)*	20.8 (17.5)	20.7 (17.3)
Year (ref:2017)	0.290	0.295
2018	0.250	0.260
2019	0.244	0.241
2020	0.215	0.204
No of courts	70	70
N	170,428	92,030

*The standard deviations of continuous variables are shown in parenthesis.

RESULTS

Table 2 shows the logistic multilevel model results for imprisonment. The model is designed to predict the probability (odds) of a defendant with a given set of characteristics to receive a custodial sentence. All the explanatory variables are categorical (except court case severity and ethnic density) so the coefficients show the effect of being in a given group compared with the reference group for that factor. Odds ratios greater than 1 show increased odds of receiving a custodial sentence compared to the reference category.

Table 2.

Two-level logistic model of imprisonment

	Model 1			Model 2			Model 3
	b	SE	OR	b	SE	OR	b	SE	OR
Ethnicity (ref:white British)
Indian	−0.01	0.05	0.99	0.15**	0.06	1.16	0.15**	0.06	1.17
Pakistani	0.12***	0.03	1.13	0.19***	0.04	1.21	0.20***	0.04	1.22
Bangladeshi	0.14**	0.05	1.15	0.19***	0.06	1.21	0.20**	0.06	1.22
Other Asian	0.04	0.04	1.04	0.06	0.04	1.06	0.06	0.04	1.06
Black African	0.19***	0.03	1.21	0.09**	0.03	1.09	0.09**	0.03	1.10
Black Caribbean	0.32***	0.03	1.38	0.17***	0.03	1.19	0.18***	0.03	1.19
Other Black	0.28***	0.03	1.32	0.09**	0.03	1.09	0.09**	0.03	1.10
White and Asian	−0.07	0.10	0.93	−0.17	0.11	0.84	−0.17	0.11	0.84
White and Black African	0.38***	0.09	1.46	0.19*	0.10	1.22	0.20*	0.10	1.22
White and Black Caribbean	0.33***	0.04	1.39	0.16**	0.05	1.17	0.16**	0.05	1.18
Other mixed	0.05	0.06	1.06	−0.09	0.06	0.92	−0.09	0.06	0.92
Chinese	0.18	0.12	1.19	0.35*	0.14	1.41	0.34*	0.14	1.41
Other	0.08	0.05	1.08	0.03	0.05	1.03	0.03	0.05	1.03
Other White	0.16***	0.02	1.18	0.13**	0.03	1.13	0.13***	0.03	1.14
White Irish	0.19**	0.06	1.21	0.05	0.06	1.05	0.05	0.06	1.05
Age (ref: under 20)
20–29	2.21***	0.02	9.12	2.86***	0.03	17.54	2.86***	0.03	17.53
30–50	2.46***	0.02	11.70	3.12***	0.03	22.54	3.11***	0.03	22.52
Over 50	2.09***	0.03	8.05	3.08***	0.03	21.76	3.08***	0.03	21.74
Gender:Male	0.84***	0.02	2.32	0.60***	0.02	1.83	0.60***	0.02	1.83
Income deprived	0.18***	0.01	1.19	0.08***	0.01	1.08	0.07***	0.01	1.08
Offence (ref:Other)
Drug offences				0.02	0.02	1.02	0.02	0.02	1.02
Fraud offences				0.08**	0.03	1.09	0.08**	0.03	1.09
Possession of weapons				−0.32***	0.02	0.73	−0.32***	0.02	0.73
Robbery				0.85***	0.04	2.34	0.85***	0.04	2.35
Violence against the person				−0.28***	0.02	0.76	−0.28***	0.02	0.76
High severity offence				0.94***	0.02	2.57	0.94***	0.02	2.57
Previous conviction				0.70***	0.01	2.01	0.70***	0.01	2.01
Pre-trial remand				2.02***	0.01	7.53	2.02***	0.01	7.53
Co-defendants				0.19***	0.02	1.21	0.19***	0.02	1.21
Plea not guilty				1.17***	0.02	3.23	1.17***	0.02	3.23
Court workload (ref:low)
Medium							−0.01	0.05	0.99
High							−0.02	0.05	0.98
Court case severity (%)							0.05***	0.00	1.05
Ethnic density (%)							−0.01*	0.00	0.99
Court var	0.06			0.09			0.02
AIC	216,701.70			176,099.40			176,015.80
BIC	216,952.90			176,451.00			176,407.60
N = 170,428

All models include controls for year. The intercept is supressed.

*p < 0.05; **p < 0.01; ***p < 0.001.

The results from Model 1 which includes defendant demographic and socioeconomic characteristics shows that the likelihood of imprisonment is higher for male than female defendants, those who live in deprived areas and for defendants in older age groups compared with those in the under 20 group (Model 1). Defendants from nine ethnic minority groups including those in white ethnic groups are more likely to be sentenced to prison than white British defendants after controlling for other defendant characteristics.

Model 2 shows that defendants convicted for robbery and for high severity offences were more than twice as likely to receive a custodial sentence than those convicted of other offences. A not-guilty plea is associated with a 223 per cent increase in the chance of imprisonment. The likelihood of imprisonment is also higher for defendants in cases with multiple defendants.

Defendants with previous convictions are twice as likely than defendants without convictions to be sentenced to custody while defendants who were remanded in custody at committal to Crown Court were 7.5 times more likely to sentenced to custody.

After adjusting for individual and case characteristics, defendants from ethnic minority groups are shown to be more likely to receive a custodial sentence (imprisonment). A custodial sentence is 41 per cent more likely for Chinese defendants, and between 16 and 21 per cent more likely for defendants from Asian groups, compared with white British defendants. Similarly, a custodial sentence is between 9 and 19 per cent more likely for defendants in the black groups, and 22 per cent more likely for white and black African defendants than white British defendants after adjusting for other characteristics. The model suggests that around 2 per cent of the variation in imprisonment is between courts rather than within courts after adjusting for individual defendant and case factors.

Model 3 in Table 2 shows that a higher court custodial rate in courts (which indicates higher average severity in sentencing) increases the likelihood of imprisonment, but court case workload is not associated with custodial sentencing decisions. On the other hand, defendants appearing in courts with a higher ethnic minority population have a lower, rather than higher, probability of imprisonment, a finding which is not in line with the racial threat hypothesis.

The model results shown in Table 3 show the two-level Hierarchical Linear Model of sentence length. The dependent variable is log-transformed so the interpretation is in terms of the exponentiated coefficients. Model 1 shows that all ethnic minority groups except for the white Irish and those in the white and Asian and the white and black African groups receive longer sentences than white British defendants after adjusting for gender, age and income deprivation (Model 1). Model 2 shows that differences in sentence length are largely explained by the legal (case) factors included in the models. Defendants convicted for drug offences receive 100 per cent longer sentences, while those convicted for robbery and offences with a high severity score receive between 250 and 260 per cent longer sentences than those convicted of other offences. Similarly, defendants who were remanded in custody prior to trial or sentencing receive 30 per cent longer sentences than those who had a different remand status. Defendants who pleaded not guilty had 95 per cent longer sentences than those who pleaded guilty while those with co-defendants also received 56 per cent longer sentences.

Table 3.

Two-level HLM of sentence length

	Model 1		Model 2		Model 3
	b	SE	b	SE	b	SE
Ethnicity (ref:white British)
Indian	0.18***	0.04	0.02	0.04	0.02	0.04
Pakistani	0.38***	0.03	0.05*	0.02	0.06**	0.02
Bangladeshi	0.40***	0.04	0.11**	0.04	0.11**	0.04
Other Asian	0.27***	0.03	−0.01	0.03	−0.01	0.03
Black African	0.25***	0.02	−0.01	0.02	−0.01	0.02
Black Caribbean	0.27***	0.02	0.04*	0.02	0.04*	0.02
Other Black	0.23***	0.02	0.03	0.02	0.04	0.02
White and Asian	0.16	0.09	0.02	0.07	0.02	0.07
White and Black African	0.13	0.07	0.00	0.06	0.00	0.06
White and Black Caribbean	0.18***	0.03	0.03	0.03	0.03	0.03
Other mixed	0.18***	0.05	0.00	0.04	0.00	0.04
Chinese	0.29**	0.10	0.00	0.09	0.00	0.09
Other	0.18***	0.04	−0.07*	0.03	−0.07*	0.03
Other White	0.10***	0.02	−0.16***	0.02	−0.15***	0.02
White Irish	−0.10*	0.05	−0.06	0.04	−0.06	0.04
Age (ref: under 20)
20–29	−0.18	0.03	0.08**	0.03	0.08**	0.03
30–50	−0.22	0.03	0.11**	0.03	0.11**	0.03
Over 50	0.07	0.03	0.32***	0.03	0.32**	0.03
Gender: Male	0.24	0.02	0.25***	0.01	0.25**	0.01
Income deprived	−0.05	0.01	−0.02*	0.01	−0.02*	0.01
Offence (ref:Other)
Drug offences			0.69**	0.01	0.69**	0.01
Fraud offences			−0.15***	0.02	−0.15***	0.02
Possession of weapons			−0.24***	0.01	−0.24***	0.01
Robbery			1.25***	0.02	1.25***	0.02
Violence against the person			−0.01	0.01	−0.01	0.01
High severity offence			1.28***	0.01	1.28***	0.01
Previous conviction			−0.37***	0.01	−0.37***	0.01
Pre-trial remand			0.26***	0.01	0.26***	0.01
Co-defendants			0.44***	0.01	0.44***	0.01
Plea Not Guilty			0.67***	0.01	0.67***	0.01
Court workload (ref:low)
Medium					−0.01	0.03
High					−0.01	0.03
Court case severity (%)					0.01**	0.00
Ethnic density (%)					0.00	0.00
Court var	0.02		0.01		0.01
Individual var	1.67		1.22		1.22
AIC	308582.9		279850.6		279850.5
BIC	308828.1		280190.1		280227.7
N = 92,030

All models include controls for year. The intercept is supressed.

*p < 0.05; **p < 0.01; ***p < 0.001.

Turning to ethnicity, the results show that defendants from most ethnic minority groups are no more or less likely to receive longer sentences than white British defendants after controlling for individual, case and court characteristics. The exceptions are those from Pakistani, Bangladeshi and black Caribbean groups who are shown to have worse sentencing outcomes than the white British even after adjusting for all other factors, receiving between 4 and 11 per cent longer sentences than the white British.

Overall, the model suggests that only 2 per cent of the variation in sentence length is between courts rather than within courts after adjusting for all factors. Contextual factors are not associated with the sentence length given to custodial sentences, apart from court average sentencing severity, which slightly increases defendants’ sentence length.

Variations in sentencing within ethnic groups

Tables 4 and 5 examine variations in sentencing outcomes within selected ethnic groups with different legal and extra-legal characteristics (Model 2). The tables show that for ethnic minority groups the direction of the main effects is the same as for the white British. There are, however, differences in the magnitude of the effects of the factors examined across ethnic groups. Table 4 shows across ethnic groups males have higher odds of receiving a custodial sentence than females but within the Bangladeshi group there are no statistically significant differences in the likelihood of imprisonment between males and females. It can also be seen that both gender and age differentials are larger within the white and black Caribbean and Pakistani groups compared to other ethnic groups. Living in a deprived neighbourhood increases the chance of imprisonment only among white British and other white defendants. Among Bangladeshi defendants those convicted of very serious offences were 5.2 times more likely to receive a custodial sentence than those in less serious offences. In comparison, white defendants convicted of serious offences were 2.5 times more likely to receive a custodial sentence. While the likelihood of imprisonment is higher for defendants from ethnic minority groups convicted for drugs offences compared to those convicted of other offences, for white British defendants the likelihood of imprisonment is lower if convicted for drugs offences. Similarly, the effect of pre-trial detention is larger for defendants in the other white group than any other ethnic group. Pleading not guilty increases the chance of imprisonment for all ethnic groups but there are larger differences by plea proposal within the other white, Pakistani and the white British groups.

Table 4.

Two-level logistic model of imprisonment by ethnic group

	Black Caribbean			Black African			White and Black Caribbean			Pakistani			Bangladeshi			White Other			White British
	b	SE	OR	b	SE	OR	b	SE	OR	b	SE	OR	b	SE	OR	b	SE	OR	b	SE	OR
Age (ref: under 20)
20–29	2.63***	0.11	13.92	2.85***	0.10	17.28	2.98***	0.16	19.71	2.95***	0.15	19.04	2.81***	0.22	16.53	2.76***	0.13	15.87	2.87***	0.04	17.57
30–50	2.91***	0.12	18.29	3.10***	0.11	22.30	3.26***	0.17	26.10	3.18***	0.15	24.11	3.05***	0.23	21.17	2.97***	0.13	19.57	3.14***	0.04	23.00
Over 50	2.65***	0.14	14.21	3.30***	0.19	27.07	3.42***	0.39	30.71	3.51***	0.24	33.42	2.57***	0.38	13.07	3.16***	0.17	23.69	3.10***	0.04	22.11
Gender:Male	0.66***	0.13	1.93	0.73***	0.12	2.08	0.92***	0.17	2.51	0.84***	0.21	2.32	0.22	0.34	1.25	0.82***	0.10	2.27	0.57***	0.02	1.77
Income deprived	0.06	0.06	1.06	0.01	0.06	1.01	0.18	0.10	1.20	0.08	0.08	1.08	0.04	0.11	1.04	0.19**	0.06	1.21	0.07***	0.01	1.08
Offence (ref:Other)
Drug offences	0.08	0.08	1.09	0.36***	0.08	1.43	-0.05	0.12	0.96	0.47***	0.09	1.60	0.35*	0.14	1.41	0.26***	0.07	1.29	-0.08***	0.02	0.92
Fraud offences	-0.10	0.17	0.90	0.40***	0.12	1.49	-0.58	0.29	0.56	-0.04	0.15	0.96	0.32	0.23	1.37	-0.01	0.12	0.99	0.08**	0.03	1.08
Possession of weapons	-0.22*	0.11	0.80	0.12	0.11	1.12	-0.25	0.17	0.78	-0.30*	0.14	0.74	-0.67**	0.24	0.51	-0.57***	0.10	0.56	-0.34***	0.03	0.71
Robbery	0.50**	0.16	1.64	0.59***	0.15	1.81	0.74**	0.22	2.11	1.03***	0.19	2.79	0.11	0.36	1.12	0.39*	0.19	1.48	1.02***	0.05	2.77
Violence against the person	-0.03	0.11	0.97	-0.12	0.11	0.89	-0.38*	0.16	0.68	-0.30**	0.12	0.74	-0.41*	0.19	0.66	-0.34***	0.09	0.71	-0.28***	0.02	0.75
High severity offence	0.86***	0.14	2.37	1.26***	0.14	3.52	1.15***	0.21	3.15	1.23***	0.16	3.43	1.66***	0.26	5.25	1.07***	0.12	2.92	0.90***	0.03	2.47
Previous conviction	0.54***	0.07	1.72	0.63***	0.06	1.87	0.51***	0.11	1.67	0.44***	0.07	1.56	0.54***	0.12	1.72	0.56***	0.06	1.76	0.74***	0.01	2.09
Pre-trial remand	1.83***	0.07	6.24	1.81***	0.06	6.12	1.97***	0.10	7.19	1.95***	0.08	7.00	1.82***	0.13	6.19	2.33***	0.06	10.30	2.02***	0.02	7.51
Co-defendants	0.18*	0.09	1.20	0.12	0.08	1.12	0.05	0.13	1.05	0.31***	0.08	1.36	0.34*	0.15	1.41	0.23**	0.08	1.25	0.19***	0.02	1.21
Plea Not Guilty	1.05***	0.09	2.86	0.95***	0.08	2.58	1.05***	0.16	2.86	1.25***	0.11	3.50	0.81***	0.15	2.25	1.11***	0.09	3.04	1.26***	0.03	3.52
Court var	0.09			0.09			0.04			0.04			0.05			0.13			0.08
AIC	6628.00			7352.00			2885.80			5052			2003.50			8354.70			125908
BIC	6763.30			7489.40			3005.70			5182.1			2114.30			8495.30			126102
N Courts	66			70			68			65			70			70			70
N	6,417			7,118			2,964			4,919			1,879			8,348			121,367

All models include controls for year. The intercept is supressed.

*p < 0.05; **p < 0.01; ***p < 0.001.

Table 5.

Two-level logistic model of sentence length by ethnic group

	Black Caribbean		Black African		White and Black Caribbean		Pakistani		Bangladeshi		White Other		White British
	b	SE	b	SE	b	SE	b	SE	b	SE	b	SE	b	SE
Age (ref: under 20)
20–29	0.21*	0.1	0.12	0.09	0.01	0.13	−0.02	0.11	0.01	0.18	0.30**	0.11	0.04	0.03
30–50	0.18	0.1	0.18	0.1	−0.07	0.14	0.03	0.11	0.03	0.18	0.36***	0.11	0.09**	0.03
Over 50	0.22	0.12	0.16	0.13	0.04	0.21	0.03	0.15	0.36	0.27	0.60***	0.12	0.31***	0.04
Gender: Male	−0.05	0.11	0.25*	0.1	0.17	0.13	0.37*	0.15	−0.24	0.22	0.02	0.07	0.27***	0.02
Income deprived	0.01	0.04	0.11*	0.04	0.05	0.06	−0.01	0.05	−0.07	0.06	0.03	0.03	−0.03**	0.01
Offence (ref:Other)
Drug offences	0.91***	0.05	1.06***	0.05	0.72	0.08	1.05***	0.05	1.09***	0.08	0.55***	0.04	0.61***	0.01
Fraud offences	−0.18	0.13	0.03	0.09	−0.28	0.19	0.25*	0.1	0.59***	0.13	−0.06	0.07	−0.20***	0.02
Possession of weapons	0.07	0.07	0.1	0.07	−0.14	0.1	0.02	0.09	0.02	0.15	−0.29***	0.07	−0.30***	0.02
Robbery	1.45***	0.09	1.33***	0.09	1.31***	0.11	1.24***	0.09	1.47***	0.2	1.21***	0.09	1.24***	0.02
Violence against the person	0.18*	0.07	0.23**	0.07	0	0.09	0.04	0.07	−0.10	0.1	−0.07	0.05	−0.02*	0.01
High severity offence	1.42***	0.08	1.60***	0.07	1.45***	0.11	1.40***	0.07	1.75***	0.12	1.36***	0.06	1.21***	0.02
Previous conviction	−0.28***	0.05	−0.30***	0.04	−0.25***	0.07	−0.30***	0.04	−0.26***	0.07	−0.26***	0.03	−0.41***	0.01
Pre-trial remand	0.39***	0.04	0.24***	0.04	0.28***	0.06	0.38***	0.04	0.26***	0.07	0.25***	0.04	0.25***	0.01
Co-defendants	0.42***	0.06	0.47***	0.05	0.51***	0.08	0.38***	0.05	0.36***	0.09	0.44***	0.04	0.44***	0.01
Plea not guilty	0.66***	0.05	0.63***	0.05	0.52***	0.08	0.76***	0.05	0.48***	0.08	0.63***	0.04	0.69***	0.01
0.00		0.01		0.02		0.03		0		0.02		0.01
1.46		1.28		1.26		1.09		1.06		1.12		1.2
11805.50		11605.5		5039		7977.4		3071.4		14216.3		196735.7
11935.80		11736.3		5152.2		8101.4		3071.4		14352.2		196926.5
N Courts	66		68		66		63		65		70		70
N	3658		3741		1623		2702		1046		4785		65071

All models include controls for year. The intercept is supressed.

*p < 0.05; **p < 0.01; ***p < 0.001.

As shown in Table 5 with the exception of the other white group, there are no age differentials in sentence length within ethnic minority groups after taking into account of other factors. Pre-trial detention has a greater effect on sentence length among Pakistani and black Caribbean defendants than other defendants. In comparison to the white British group, defendants from the Bangladeshi, black African and Pakistani groups convicted of drugs offences, receive harsher sentences than defendants convicted of other offences. Similarly, black African and black Caribbean defendants convicted in cases with multiple defendants are more likely to receive longer sentences compared to those without co-defendants.

Taken together the results show that there are statistically significant differences in the chance of imprisonment and sentence length between and within ethnic groups, after having controlled for defendant and case and court characteristics.

DISCUSSION

In response to increasing evidence of ethnic inequalities in the CJS in England and Wales, analyses presented in this paper sought to examine whether ethnic minority defendants are sentenced more harshly than white British defendants, independent of legally relevant factors and determine the extent of disparities in sentencing between disaggregated ethnic groups. The research was provoked by the sparsity of studies examining the extent and drivers of ethnic disparities at different stages of the sentencing process. Unlike prior research, this analysis employs multilevel modelling, considers a wide range of case and contextual characteristics and incorporates disaggregated ethnic categories, setting it apart from earlier empirical studies examining ethnic disparities in sentencing in the United Kingdom.

The research found, consistent with the premises of focal concerns theory, a close association between sentencing outcomes and legally relevant (case) factors. Pleading not guilty, pre-trial detention and offence severity are associated with an increased likelihood of imprisonment and a longer custodial sentence. Among these, plea is shown to have a strong effect on sentencing outcomes with those entering a not guilty plea being three times more likely to be imprisoned and receiving nearly double the sentence length. In a similar vein, pre-trial detention holds a strong association with the likelihood of imprisonment with defendants remanded in custody prior to sentencing in the Crown Court being 7.5 times more likely to receive a custodial sentence. The association between poorer plea bargaining and pre-trial detention outcomes and harsher sentencing outcomes, combined with the higher not guilty pleas and pre-trial detention rates amongst ethnic minority groups (Uhrig 2016) suggests that ethnic minorities may be subject to ‘cumulative disadvantage’ (Wooldredge et al. 2015). From this it follows that tackling ethnic disparities in sentencing will require a greater appreciation by sentencers of the existence of ethnic disparities at earlier stages of the CJS (Roberts and Ashworth 2022). This may entail a review of processes leading to the overrepresentation of people from ethnic minority groups in pre-sentencing outcomes such as plea and pre-trial detention, and evaluation of guidelines which may contribute to harsher sentencing outcomes for ethnic minority groups. For example, court officials’ assessments of mitigating factors such as early admissions at the pre-trial stage in cases involving ethnic minority defendants can include considerations that they are less likely to benefit from this ground of mitigation and plead guilty at a later stage, because of distrust in the CJS (Roberts and Ashworth 2022). Greater transparency and accountability of pre-trial detention processes and improved information in pre-sentence assessments regarding the reasons for pre-trial detention is also needed to enable better assessments regarding risks posed by defendants.

In contrast to research in the United States which has drawn attention to the disparate impacts of race-neutral policies there is little understanding about the effect of policies such as the plea-dependant differential on the disadvantage faced by ethnic minorities in England and Wales. There is a necessity for future studies to elucidate the ramifications of such policies on bias to help identify and address systemic issues, ensuring that policies are effective in reducing ethnic disparities.

Although the research presented in this paper offers support for the importance of legally relevant factors in explaining sentencing differentials, it demonstrates that such factors, do not fully explain ethnic disparities. The results show that there is a consistent independent association between ethnicity and the likelihood of imprisonment after controlling for other well-established predictors of imprisonment. In contrast, disparities in sentence length between most, but not all, ethnic minority groups and the white British disappear after controlling for legally relevant factors such as offence type and severity. In other words, legally relevant factors explain, to a large extent, observed ethnic differentials in sentence length. Taken together these findings offer compelling evidence that ethnicity plays an important role in imprisonment decisions, but less so for sentence length decisions. These findings are in line with US studies that have shown that ethnic disparities are more pronounced for the decision to imprison or not than for sentence length (Chiricos and Crawford 1995; Spohn 2000; Baumer 2013) indicating that people from ethnic minority groups are treated more equally at later stages of the sentencing process. Although the likelihood of custodial sentencing varies within ethnic minority defendants according to differences in observed characteristics in a similar way that it does for white British defendants, sentencing differentials cannot be attributed solely to compositional factors. Future research could extend the analysis to better understand differentials in sentencing outcomes between individual ethnic minority groups and the white British using decomposition analyses which can identify the contribution of different legal and extra-legal factors in explaining ethnic gaps in sentencing.

The research also reveals significant disparities in sentencing outcomes among ethnic groups, particularly within black, Asian and mixed categories. Chinese, Bangladeshi, Pakistani and White and Black African groups, in particular, face greater disadvantages in sentencing. Notably, even within white groups, the ‘other white’ subgroup often subsumed in the white category in studies, faces a higher likelihood of imprisonment compared to their white British counterparts. This highlights the inadequacy of combining ethnic groups into broader categories, as it may lead to misleading conclusions about the extent of disparities. The findings emphasize the need to disaggregate ethnic data for a nuanced understanding of inequalities in the CJS. To address ethnic disparities effectively, criminal justice agencies should report and monitor outcomes in sufficient depth, recognizing the diverse experiences within ethnic minority groups rather than treating them as homogeneous.

The persistence of ethnic differentials which remain after adjusting for other well-established predictors of sentencing outcomes may be attributed to discrimination in the CJS or other factors including mitigating and aggravating factors not captured in the DF datasets. Although it is difficult to attribute these unexplained gaps solely to discrimination, the lower extent of ethnic disparities in sentence length compared to imprisonment is likely to be indicative of differences in biased use of discretion in decisions taken at different stages of the sentencing process. Sentencing guidelines dictate that the type of sentence imposed should be based on assessments regarding the seriousness of the offence, the culpability of the ‘offender’ and the harm caused to victims (Sentencing Council 2017). In choosing an appropriate sentence the judiciary are required to assess the relevance and weight of different sentencing factors, and in doing so, exercise discretion in varying degrees depending on the level of information available to them regarding the seriousness of the offence at time of sentencing. In line with focal concerns, in cases where there is ambiguity about the appropriate level of punishment, the judiciary rely on ‘perceptual shorthands’ and exercise more personal discretion on sentencing decisions within which biases towards racial and ethnic minority groups may arise (Steffensmeier et al. 1998). The sources of biases are also likely to be deeper and more systemic than bias in the way sentencing decisions are made by individual sentencers (Clair and Winter 2016). This level of discretion may not be exercised in decisions regarding sentence length, that is, after the judiciary have determined imprisonment as the most appropriate sentence. The reason is that sentence length is primarily determined by the maximum penalty for the crime allowed by law including mandatory minimum sentences passed by Parliament. In this way, mandatory minimum sentence guidelines may serve to limit biased judicial discretion in sentence length decisions. Although these are plausible explanations, there is a need for qualitative research to explore how ethnicity-based focal concerns and biases affect decisions at different stages of the sentencing process.

In the realm of empirical investigations into racial bias, a small but growing body of research in the United States has shed light on the additive and multiplicative effects of legal and extra-legal factors in shaping sentencing outcomes (Steffensmeier et al. 1998, 2017; Doerner and Demuth 2010). While the present analysis focuses on the direct effects of ethnicity using disaggregated ethnic categories, it is crucial to recognize that ethnicity can interact with other characteristics such as gender, age and offence severity, to produce unequal outcomes and punishment. There is a need for future research to delve into the complexities of these interactions in order to provide a more comprehensive exploration of the nuanced factors influencing biases in the CJS.

This research found that court disparities in sentencing are relatively small, which can be interpreted as evidence of some degree of consistency in sentencing between courts (Pina-Sánchez and Grech 2018). Contrary to expectations established in previous US studies (Johnson et al. 2008; Ulmer et al. 2011), the research found limited evidence of an association between court contextual factors such as court case workload and sentencing outcomes although higher average conviction rates in courts were shown to increase the likelihood of imprisonment and sentence length. Similarly, the ethnic composition of the areas in which courts operate was only found to be associated with the likelihood of imprisonment, but its effect was in the opposite direction than predicted by the racial threat hypothesis, a finding which echoes some previous US based studies (Feldmeyer and Ulmer 2011). While this study tested whether the size of the ethnic minority population has a direct effect on sentencing, it hasn’t explored whether ethnic density conditions the relationship between ethnicity and sentencing outcomes, a hypothesis that warrants attention in future research. Taken together, these findings appear to challenge the court community perspective which postulates that the local legal environment and community factors play a significant role in shaping sentencing decisions. However, variations in the implementation of sentencing guidelines and judicial procedures between the United States and the United Kingdom are probable explanations for the lower variation in sentencing outcomes between courts and the lesser significance of court community factors. The finding that court contextual factors hold less relevance in explaining sentencing variations highlights the need to acknowledge that sentencing research must consider the influence of unique legal traditions, practices and legislative frameworks within the study settings.

ACKNOWLEDGEMENTS

The author is grateful to Patrick Williams and Jon Bannister at Manchester Metropolitan University, members of the Ministry of Justice Data First team and ADR UK for their support of this research.

DISCLAIMER

This work was produced using administrative data accessed through the ONS Secure Research Service. The use of the data in this work does not imply the endorsement of the ONS or data owners (e.g. MoJ and HM Courts and Tribunals Service) in relation to the interpretation or analyses of the statistical data. This work uses research datasets which may not exactly reproduce National Statistics aggregates. National statistics follow consistent statistical conventions over time and cannot be compared to Data First linked datasets.

FUNDING

This work is supported by the Economic and Social Research Council (ES/V015613/1).

Footnotes

1

A comprehensive review of the literature is beyond the scope of this paper. Instead, a brief overview of existing reviews of US studies is presented, alongside findings on the effect of ethnicity on sentencing from studies in the UK.

2

The datasets provide information on defendant appearances in magistrates’ (2011–2020) and Crown courts (2013–2020) extracted from the court management information systems Libra and Xhibit, respectively. The datasets include a linking dataset that enables the linking of defendants and cases across the magistrates’ and Crown Court datasets. Whilst the analysis examines sentencing outcomes in the Crown Court the linking of Crown Court and magistrates’ court datasets enables identification of additional defendant characteristics such as previous convictions.

3

https://www.judiciary.uk/you-and-the-judiciary/going-to-court/.

4

Ethnicity coverage is less complete in magistrates’ courts and much of the unrecording of ethnicity relates to summary offences which make up most of the cases in magistrates’ courts. For example, across all years, around three-quarters of defendants appearing in magistrates’ courts for cases involving summary offences had missing ethnicity information.