Drunk and Doubly Deviant? the Role of Gender and Intoxication in Sentencing Assault Offences

Abstract

Little is known about how alcohol intoxication impacts sentence outcomes. This study assesses whether intoxication differentially aggravates sentence outcomes for male and female defendants of assault offences. It does so by modelling the probability of custody and sentence severity using the Crown Court Sentencing Survey, including interaction terms to account for the gendered application of intoxication as a sentencing factor. The main finding is that the aggravation afforded to female defendants is twice than that afforded to males where intoxication in present and when controlling for relevant case characteristics. The study spotlights how cases of assault are processed through the criminal justice system and raises concerns with how gender equality is interpreted in sentencing practice with reference to alcohol intoxication.

Introduction

At the time of writing, the UK Government is due to produce a female offenders strategy. Underpinned by conceptions of social and distributive justice and a resounding consensus with the recommendations from the Corston report (Corston 2007), it is broadly recognized that responses to women’s offending ought to be gender-specific, respond sensitively to the needs of women (House of Commons 2017) and divert them away from custody (Ministry of Justice [MoJ] 2013). However, recent sentencing policy has forwarded gender-neutral guidance according with the principles of ‘procedural’ justice to promote consistency and ‘equality’ of outcome. Although these principles accord with a just-deserts-based model focused on criminal risk management, they run counter to conceptions of justice prioritizing rehabilitation or reparative approaches to social harm (see Player 2014). These paradoxical imperatives result in challenges for ensuring just and fair punishment for women in policy and practice.

In the context of drives to standardize the administration of justice through sentencing, there also exists a lack of practical direction on the extent to which alcohol intoxication should aggravate sentence outcomes, for whom and in which contexts (Lightowlers and Pina-Sanchez 2017). Despite ample evidence that women are judged more harshly for their alcohol intoxication than men (Plant 1997; Staddon 2015), within scholarship of sentencing there has been limited engagement with how alcohol intoxication interacts with gender to shape punishment for defendants. Nor has there been much explicit examination of how the presence of alcohol consumption or intoxication in a case serves to impact sentence outcomes, with only a few recent exceptions focusing on capital sentencing deliberations in the US (Bjerregaard et al. 2010) and defendants’ histories of alcohol abuse (Stevenson et al. 2010). This represents a significant gap in our understanding, given intoxication has been identified as one of the most contentious legal factors in sentencing (Dingwall and Koffman 2008; Padfield 2011; Irwin-Rogers and Perry 2015) and that gender has consistently been found to be one of the most powerful extralegal factors influencing sentence outcomes over and above others such as age and ethnicity (Steffensmeier et al. 1998; Spohn and Holleran 2000; Steffensmeier and Demuth 2000).

Using nationally representative data, this study advances insights into how alcohol intoxication impacts the punishment of women for assault offences processed by the Crown Court in England and Wales. In so doing, it offers conceptual and policy-relevant insights into the gendered administration of justice and lays the foundation for further critical engagement with issues of gender and intoxication in sentencing.

Theorizing gender and sentencing

Historically debates on gender and punishment have centred around two polarized standpoints: those suggesting women are treated with greater leniency (sometimes also referred to as ‘chivalry’) and those suggesting women receive more severe treatment as they are seen as ‘doubly deviant’ (Lloyd 1995) (for violating not only the law but also conventional social norms associated with womanhood) in a male-dominated legal and criminal justice system. Although both accept gender will impact sentencing, they see it as operating in different ways. Given several studies highlighting female offenders tend to receive more lenient treatment than their male counterparts committing the same crimes (Daly 1989; Daly and Tonry 1997; Spohn 1999), the former framework suggests notions of women being less threatening, dangerous and culpable than men and a paternalistic desire to protect and aid women (Daly 1989; Heidensohn and Silvestri 2012) results in a more lenient approach to female offending being adopted by (predominantly male) decision makers. The latter framework, sometimes also referred to as theories of ‘evil women’ (Bontrager et al. 2013), is at odds with the view that male leniency protects women in the criminal justice system. It suggests women receive harsher social judgement for their actions as well as harsher sentences (double jeopardy) given the courts’ view of them having transgressed legal rules as well as socially approved patterns of female behaviour (Heidensohn 2010; Heidensohn and Silvestri 2012).

In this latter framework, tougher sentencing of women can result (Hedderman and Gelsthorpe 1997) and punishment is likely to be amplified where intoxication is present in female offending, as both criminal behaviour and intoxication violate traditional hegemonic conceptions of womanhood. However, previous research points to a seeming universal leniency or ‘chivalry’, with females receiving fewer and often shorter custodial sentences when controlling for key case characteristics (Steffensmeier et al. 1993; Rodriguez et al. 2006; Bontrager et al. 2013; Doerner and Demuth 2014), although these studies have not explicitly examined the role gender plays in offences involving alcohol intoxication.

Hybrids of these opposing theoretical standpoints also exist. Nagel and Hagan (1983) suggested sentencing outcomes are likely to be offence specific, with gender operating as a contextual effect on sentencing decisions. That is, leniency for women is more likely to be applied in relation to less serious and property offences as opposed to serious and personal crimes given these behaviours adhere to traditional notions of ‘appropriate’ female behaviour. However, there exists limited robust evidence to support selective leniency dependant on crime type. Rodriguez et al. (2006) found whilst women were no less likely to receive a custodial sentence than men for violent crime (although they were less likely for acquisitive and drug-related crime), they did receive substantially shorter sentences than males. The authors credit the first finding to the reduced legal discretion available to sentencing practitioners for more serious (violent) offences, supporting the liberation thesis whereby less serious crimes are likely to be influenced by extralegal factors such as gender (Rodriguez et al. 2006). However, their findings also support the notion of selective leniency, whereby females will benefit from their gender only when the offence is seen as befitting females (e.g. acquisitive crime) as opposed to more masculine (e.g. violent) crimes (Rodriguez et al. 2006).

Conceptualizing equality

Broader feminist debates surrounding the administration of justice have traditionally centred on notions of ‘equality’ and ‘social justice’ (c.f. Heidensohn 2010; Heidensohn and Silvestri 2012; Gelsthorpe and Sharpe 2015), highlighting the way in which women are often more disadvantaged and unequal before they even encounter the law and experience penal sanctions as harsher than men¹. Related critiques of the criminal justice system have highlighted how administrative practices are linked to deep-seated assumptions of gender based roles and behaviour, and masculine conceptions of justice based on legal and procedural equality (Heidensohn 2010). Feminist scholars also spotlight how more feminine conceptions of justice, such as ‘social’ and ‘distributive’ justice, are relatively absent in sentencing guidelines and wider criminal justice policy, given that white middle-class men remain overwhelmingly the guardians of justice². It is thus argued equality cannot be conceptualized merely by formalized definitions of procedural fairness and accountability (i.e. consistency in sentencing/equality of sentence outcomes), as these measures do not always do women justice (Heidensohn 2010). When it comes to sentencing, equality of outcome (e.g. sentence length and/or severity) on average between men and women is not necessarily the correct goal to strive for. Indeed, many favouring redistributive conceptions of justice would argue that to do so is inherently unjust (see Richards 1980 cited in Heidensohn 2010; Player 2014). Although equality and human rights legislation promote a form of distributive justice that recognizes and responds to gender differences in the pursuit of equal treatment (Player 2014), regrettably progress towards a distinct woman-centred approach (as advocated by Corston (2007) in the English and Welsh context) has been minimal (HM Chief Inspector of Prisons 2012) and an ideology characterized by procedural justice and desert-based principles prevails (Player 2014).

A focus on violent offending and alcohol intoxication

Sentencing of alcohol-related violent offending (as opposed to other forms of offending, such as acquisitive crime) makes for an interesting case in which the gendered administration of justice can result for several reasons. Firstly, as both alcohol consumption and violence violate traditional hegemonic conceptions of womanhood, these may serve to amplify punishment where intoxication is present in female offending. And, although earlier research into how intoxication determines sentences in practice has produced varied results, these have not explicitly examined the role of gender (Shapland 1981; Rumgay 1998; Dingwall 2006; Padfield 2011; Lightowlers and Pina-Sanchez 2017).

Player (2014: 281) highlights the tensions sentencers face when operating ‘within a broader institutional context that presents them with a number of competing messages from politicians, legislation and the media, about the use of custodial sentences and about gender equality’. This is also acknowledged by Hedderman and Barnes (2015) in relation to alcohol-related offending by women, as they found that many sentencers felt women’s violence under the influence of alcohol was rising and this was probably informed by personal experiences outside the court room as well as exposure to media representation of women’s drinking patterns and associated violence. They noted this might result in sentencers responding more severely to women offenders either because they saw alcohol-related violence as a growing menace, thus necessitating a deterrent response or perceived an amplified threat to public safety posed by female offenders (Hedderman and Barnes 2015).

Secondly, a focus on violent offending is warranted given its routine association with alcohol intoxication (Boles and Miotto 2003; Graham and Homel 2008; Lightowlers 2011, Lightowlers et al. 2014) and given that this behaviour is more commonly associated with men. However, in England and Wales, official figures for 2013 (the year of published data that corresponds most to the period under study here) show violence against the person offences account for the second largest number of convictions for female offenders (MoJ 2014), with an intriguing ‘gender neutrality’ in associated sentence outcomes in comparison with other offence types (see Figure 1). Trend data for the decade 2005–15 verify this more recent ‘gender neutrality’, with the average custodial sentence length (months) for females actually overtaking that of males in 2015 (which is the most recent year for which data are publically available; see Figure 2).

Fig. 1

Average custodial sentence length for indictable offences by sex, 2013 (reproduced from MoJ 2014 [Figure A.06 p13])

Fig. 2

Average custodial sentence length for violence against the person offences, 2005–15 (Source: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/524326/cjs-outcomes-by-offence.xlsx)

Finally, Sentencing Guidelines in England and Wales are offence specific and the first guideline released by the Sentencing Council related to assault offences (2011), covering everything from threatening words to a severe physical attack, with common assault at the lower end of harm and grievous bodily harm (GBH) at the upper end. Guidance was issued to promote greater transparency and consistency in sentencing and represents a move to limit judicial discretion. Judges are thus obliged to follow sentencing guidelines, only disregarding them where their application is believed to be ‘contrary to the interests of justice’ (Coroners and Justice Act, 2009 s128(1)(a)).

The direction prescribed in the Sentencing Council’s guideline for assault offences (Sentencing Council 2011) sets a precedent as to how alcohol should impact sentencing more generally. In its guidance, the Sentencing Council prescribes that (both alcohol and drug) intoxication should aggravate an offence on the basis of its seriousness, making an offender more culpable whilst under the influence of alcohol or drugs. This stance accords with a pervading ‘malevolent assumption’ about alcohol intoxication in socially disadvantageous events (Collins 1981; Dingwall 2006) and sets a precedent that intoxication cannot be cited as mitigation (Sentencing Council 2011).

The guidelines apply equally to males and females. Player (2014: 282) suggests that whist this approach ‘does not directly contradict the Government’s published strategy for women offenders, it does uphold a degree of universalism that can eclipse consideration of important gender differences’ in favour of committing to proportionality in sentencing. Moreover, despite the Sentencing Council having a duty under section 128 of the Coroners and Justice Act 2009 to monitor the effect of its guidelines, to date there has been no assessment of sentencing practice in relation to gender and/or intoxication.

Study aims

The purpose of this study is to test for differential effects of intoxication in determining sentence outcomes in Crown Court cases in England and Wales, based on the sex of the defendant (as a characteristic that influences social and cultural practices and as a proxy for gender). The study assesses (1) whether female defendants receive less severe and fewer custodial sentence outcomes for violent offences and (2) whether any such leniency in sentence outcome is still afforded to female defendants in offences in which intoxication is cited as an aggravating factor.

Methods

Data

The Crown Court Sentencing Survey (CCSS) was inaugurated on 1 October 2010 with a view to monitoring the operation and effect of its sentencing guidelines outlined in duties under section 128 of the Coroners and Justice Act 2009. It facilitates an exploration of sentencing based on the mitigating and aggravating factors considered by the judge in determining the final sentence and so enables the effect of specific legal factors to be quantified (net of other factors) as well as the assessment of significance of theoretically informed interactions between such factors (Roberts and Hough 2015). It was administered until 31 March 2015 when the survey was ended following external review. Data for 2015 are, however, not available at the time of writing. Between 2012 and 2014, the survey achieved average response rates between 58 and 64 per cent, with some Crown Court locations reporting higher and lower response rates (see Sentencing Council 2015). Although missing data potentially pose a threat to internal validity, the Sentencing Council examined non-response by comparing records to the case management system used by Crown Court and advise relatively robust conclusions may be drawn from data collected by the survey (see Sentencing Council 2013; Roberts and Hough 2015).

The Sentencing Council amended the assault guidelines in 2011. These revised guidelines came into effect for those sentenced on or after 13 June 2011 and set out a comprehensive list of factors (including both aggravating and mitigating) that could be considered in sentencing assault offences. The CCSS data used for this study were thus pooled to include all sentences from the second quarter of 2012 to the end of 2014, as these were known to have employed the new assault guideline. These publically available data include detail on sentencing factors as well as the age and sex of defendants. However, they do not include identifiers of the courts at which sentences were passed or the sex of the judge.

The resultant pooled data set contains detail on 30,861 sentences for assault offences. This large sample size thus facilitates an examination of sentence outcomes conditional on sex as 9.25 per cent (2,855) of these were dispensed to female defendants. Moreover, the large sample size affords sufficient statistical power for the consideration of interactions between sex and sentencing factors. Nearly a quarter of all assaults cited intoxication as aggravation (24.17 per cent), and this proportion was roughly similar for both male (24.29 per cent) and female (23.01 per cent) defendants. The volume and proportion of sentences by offences type is broken down in Table 1, which highlights sections 47 (Actual Bodily Harm (ABH)) and 20 (Wounding/GBH) offences make up over half of the offences seen by the Crown Court, which the next most prevalent offence type being Affray (an instance of group fighting in a public place that disturbs the peace; 17.11 per cent) and ‘Other’ offences³.

Table 1

Offence type by sex (2012–14)

	Males	Females	Total
	n (%)	n (%)	n (%)
Affray	4,909 (17.53)	370 (12.96)	5,279 (17.11)
Common assault	2,295 (8.20)	272 (9.53)	2,567 (8.32)
Other	3,107 (11.09)	529 (18.53)	3,636 (11.78)
Wounding/GBH with intent (s18)	1,846 (6.59)	118 (4.13)	1,964 (6.36)
Wounding/GBH (s20)	5,833 (20.83)	595 (20.84)	6,428 (20.83)
Assault occasioning ABH (s47)	10,014 (35.76)	971 (34.01)	10,985 (35.60)
Total	28,004 (100)	2855 (100)	30,859a (100)

^aTwo cases were missing accurate gender information.

Measures

Detail on the sentence dispensed was used as the basis for creating two outcome variables to be modelled. First, a binary indicator of whether the offence had attracted an immediate custodial sentence or not (i.e. compared to a non-custodial or suspended sentence order—the latter are custodial sentences that are not immediately imposed and to be served in the community unless breached). Unfortunately, detail of the precise sentence length contained in the original CCSS was restricted for external users for the sake of anonymity by converting the sentence length (continuous measure) into intervals. Second, an ordinal indicator of the severity of the sentence. This emulates a measure previously adopted by Irwin-Rogers and Perry (2015) and serves to distinguish between the following five categories:

• a non-custodial sanction or suspended sentence order;

• below 12 months’ imprisonment referred to as ‘low severity’;

• 12 months’ up to 18 months’ imprisonment referred to as ‘medium-low severity’;

• 18 months’ up to 4 years’ imprisonment referred to as ‘medium-high severity’; and

• 4 years’ imprisonment and above referred to as ‘high severity’.

The distribution of incidents across the ordinal response variable (scale of severity) is displayed in Table 2.

Table 2

Ordinal measure of sentence severity

	n	%
Non-custodial sanction/suspended sentence order	14,859	48.15
Low severity	5,136	16.64
Medium-low severity	4,825	15.63
Medium-high severity	3,979	12.89
High severity	2,062	6.68
Total	30,861	100

Besides the sentence type and length, the CCSS data contain details of the offence type (see Data section and Table 1) as well as all relevant factors cited as influencing the decision. The offence type facilitates controlling for a distinction in the severity of the offence. Legal factors captured include whether the defendant entered a guilty plea, their previous convictions as well as mitigating and aggravating factors (including whether the court cited intoxication as an aggravating factor in the case), all of which can be controlled for as binary (presence/absence) covariates in a regression modelling framework (for a full list and descriptive statistics see  Appendix I and Sentencing Council (2011) alongside the defendant’s sex (as a characteristic that influences social and cultural practices and as a proxy for gender) and age group at the time of sentencing. Age and previous convictions are only provided in the CCSS as interval-censored variables in the data set and so take this form in the analysis.

The survey’s original measure of intoxication conflates both drug and alcohol intoxication, as both are prescribed to aggravate similarly in the sentencing guidelines. It is thus limited in its ability to tease out the explicit impact of alcohol intoxication. However, it is used as a proxy for alcohol intoxication here given (1) substantial evidence to suggest alcohol use is significantly associated with violence but less evidence to link drug use with violence (Parker and Auerhahn 1998) and (2) the widespread prevalence of alcohol use amongst general and ‘criminal’ populations (HMIP 2015; Flatley 2016), so affecting more cases brought before the courts.

Analysis

Binary logistic regression was employed to model the probability of receiving a custodial sentence (using the binary custody indicator as an outcome variable). In this model, the previously outlined sentencing factors as well as age, sex and offence type were included as explanatory variables. Doing so allows for an assessment of the contribution of each factor, all else being equal, and models the probability of a custodial outcome as a function of these demographics, case characteristics and sentencing factors. It can thus help to explain the relationship between the sentencing and the impact of explanatory variables on the probability of receiving a custodial sentence.

Ordinal logistic regression was used to model the severity of sentence outcomes (using the ordinal indicator of sentence severity). Ordinal regression is the appropriate specification of a regression model used to predict an ordinal outcome (dependent variable) as a function of specified explanatory variables—in this instance, the previously outlined demographic and case characteristics. This enables an assessment of which of these have a significant effect on sentence severity.

As well as controlling for the independent variables outlined earlier, a multiplicative ‘sex × intoxication’ interaction term was subsequently included in both binary and ordinal models, to assess the extent to which gender moderated sentence outcomes where intoxication was present. This facilitates an insight into how the presences of these two concomitant factors impact sentence outcomes, whilst reducing bias in resulting estimates (regression coefficients). In this study, a statistical interaction term is specified with the explicit aim of exploring the combined (multiplicative) effect of being female and intoxicated in the modelling of sentencing outcomes for assault offences, thus enabling an assessment of the extent to which gender moderates sentence outcomes where intoxication is cited as relevant to the case. Interactions terms for the crime type and presence of intoxication were also examined to further explore how intoxication serves to moderate sentence outcomes based on the seriousness of the offence.

Results

A ‘Malevolent Assumption’ Pervades: Intoxication Aggravates Sentence Outcomes

Results from the initial logistic and ordinal models (Model 1 in both Tables 3 and 4) indicate being ‘under the influence’ aggravates sentences outcomes, both in terms of the probability of receiving a custodial sentence (odds ratio [OR] = 1.29) and in terms of attracting a more severe sentence outcome (OR = 1.23). These findings confirm empirical evidence presented elsewhere (Lightowlers and Pina-Sanchez 2017) and are in line with the 2011 guidance issued by the Sentencing Council: the intoxication ought to aggravate sentences for assault offences.

Table 3

Logistic regression models for the probability of custody

	Model 1		Model 2		Model 3
	Main effects		Main effects and interaction		Main effects and interactions
	Beta coefficient (standard error)	OR	Beta coefficient (standard error)	OR	Beta coefficient (standard error)	OR
Intercept	−0.246 (0.063)	0.782***	−0.238 (0.063)	0.788***	−0.199 (0.064)	0.819**
Female	−0.741 (0.055)	0.477***	−0.82 (0.064)	0.437***	−0.831 (0.064)	0.436***
Age 25–34 years	0.013 (0.034)	1.014 n.s.	0.013 (0.034)	1.013 n.s.	0.014 (0.034)	1.014 n.s.
Age 35–44 years	0.009 (0.043)	1.009 n.s.	0.009 (0.043)	1.009 n.s.	0.011 (0.043)	1.011 n.s.
Age 45–54 years	−0.158 (0.056)	0.854**	−0.159 (0.056)	0.853**	−0.155 (0.056)	0.857**
Age 54+ years	−0.41 (0.102)	0.664***	−0.409 (0.102)	0.664***	−0.407 (0.102)	0.666***
Crime type
Reference category ‘Other’
Affray	−0.328 (0.052)	0.724***	−0.325 (0.052)	0.722***	−0.391 (0.057)	0.676***
Common assault	−0.400 (0.063)	0.670***	−0.403 (0.063)	0.669***	−0.386 (0.068)	0.680***
GBH(s18)	4.261 (0.136)	70.864***	4.260 (0.136)	70.787***	4.121 (0.148)	61.608***
GBH(s20)	1.033 (0.053)	2.810***	1.031 (0.053)	2.805***	0.943 (0.057)	2.568***
ABH(s47)	0.209 (0.046)	1.233***	0.207 (0.046)	1.230***	0.178 (0.050)	1.195***
Guilty plea	0.153 (0.040)	1.165***	0.153 (0.040)	1.165***	0.153 (0.041)	1.165***
Previous convictions
Reference category 1–3
None	−0.524 (0.037)	0.592***	−0.524 (0.037)	0.592***	−0.524 (0.037)	0.592***
4–9	0.554 (0.048)	1.740***	0.555 (0.048)	1.741***	0.558 (0.048)	1.748***
10+	0.974 (0.077)	2.649***	0.972 (0.077)	2.643***	0.979 (0.077)	2.663***
Not answered	0.267 (0.114)	1.306*	0.266 (0.114)	1.304*	0.281 (0.114)	1.324*
Aggravating factors
Location of the offence	0.324 (0.039)	1.382***	0.322 (0.039)	1.380***	0.325 (0.039)	1.384***
Timing of the offence	0.083 (0.043)	1.086 n.s.	0.086 (0.043)	1.090*	0.086 (0.043)	1.090*
Ongoing effect on victim	0.679 (0.039)	1.973***	0.680 (0.039)	1.973***	0.673 (0.039)	1.961***
Offence against those in the public sector/service to public	0.508 (0.063)	1.661***	0.507 (0.063)	1.661***	0.521 (0.063)	1.684***
Presence of others	0.353 (0.038)	1.423***	0.353 (0.038)	1.423***	0.352 (0.038)	1.422***
Gratuitous degradation	0.824 (0.105)	2.280***	0.823 (0.105)	2.277***	0.832 (0.105)	2.298***
Victim compelled to leave home	0.489 (0.113)	1.631***	0.491 (0.113)	1.634***	0.495 (0.113)	1.640***
Failure to comply with court orders	0.904 (0.061)	2.468***	0.901 (0.061)	2.461***	0.917 (0.062)	2.501***
On licence	1.710 (0.129)	5.529***	1.709 (0.129)	5.525***	1.715 (0.129)	5.557***
Attempt to conceal/dispose of evidence	2.247 (0.285)	9.460***	2.241 (0.285)	9.401***	2.255 (0.286)	9.533***
Failure to respond warnings/ concerns	0.794 (0.123)	2.212***	0.796 (0.122)	2.216***	0.802 (0.123)	2.231***
Offender was under the influence of alcohol/drugs	0.255 (0.036)	1.290***	0.228 (0.037)	1.256***	−0.134 (0.122)	0.875 n.s.
Abuse of power/trust	0.677 (0.089)	1.968***	0.683 (0.089)	1.979***	0.681 (0.089)	1.975***
Exploiting contact arrangements	0.452 (0.335)	1.571 n.s.	0.447 (0.334)	1.563 n.s.	0.473 (0.335)	1.605 n.s.
Previous violence/threats	0.537 (0.068)	1.712***	0.537 (0.068)	1.711***	0.540 (0.068)	1.716***
Established evidence of community impact	0.265 (0.197)	1.304 n.s.	0.272 (0.197)	1.312 n.s.	0.320 (0.200)	1.377 n.s.
Steps take to prevent reporting/ assisting prosecution	0.816 (0.241)	2.262***	0.826 (0.241)	2.285***	0.815 (0.242)	2.258***
Taking offences into consideration (TICs)	−0.416 (0.653)	0.659 n.s.	−0.425 (0.649)	0.653 n.s.	−0.446 (0.651)	0.640 n.s.
Mitigating factors
No previous relevant convictions	−0.248 (0.044)	0.781***	−0.247 (0.044)	0.781***	−0.248 (0.044)	0.781***
Single blow	−0.687 (0.046)	0.503***	−0.687 (0.046)	0.503***	−0.690 (0.046)	0.501***
Remorse	−0.671 (0.035)	0.511***	−0.669 (0.035)	0.512***	−0.672 (0.035)	0.511***
Good character/exemplary conduct	−0.488 (0.052)	0.614***	−0.486 (0.052)	0.615***	−0.483 (0.052)	0.617***
Determination/demonstration to address addiction/behaviour	−1.332 (0.061)	0.264***	−1.334 (0.061)	0.263***	−1.335 (0.061)	0.263***
Serious medical conditions	−1.122 (0.107)	0.326***	−1.125 (0.107)	0.325***	−1.120 (0.107)	0.326***
Isolated incident	−0.662 (0.049)	0.516***	−0.662 (0.049)	0.516***	−0.665 (0.049)	0.514***
Age/lack of maturity affecting responsibility	−0.223 (0.060)	0.800***	−0.227 (0.060)	0.797***	−0.224 (0.060)	0.799***
Lapse of time (not fault of offender)	−0.859 (0.098)	0.424***	−0.859 (0.098)	0.423***	−0.863 (0.098)	0.422***
Mental disorder/learning disability not linked to commission of offence	−0.935 (0.087)	0.393***	−0.934 (0.087)	0.393***	−0.939 (0.087)	0.391***
Sole/primary carer for dependent relatives	−1.023 (0.093)	0.360***	−1.034 (0.093)	0.355***	−1.032 (0.093)	0.356***
Interaction(s)
Female × under the influence			0.353 (0.125)	1.424**	0.365 (0.126)	1.440**
Affray × under the influence					0.472 (0.145)	1.602**
Common assault × under the influence					0.026 (0.174)	1.027 n.s.
GBH (s18) × under the influence					0.926 (0.379)	2.523*
GBH (s20) × under the influence					0.548 (0.139)	1.729***
ABH (s47) × under the influence					0.314 (0.132)	1.369*
Model fit (akaike information criterion)	30,346		30,340		30,321

n.s. = not significant. Significance levels denoted by *p < 0.01, **p < 0.001, ***p < 0.0001.

Table 4

Ordinal regression models for the probability of sentence severity

	Model 1 Main effects		Model 2 Main effects and interaction		Model 3 Main effects and interactions
	Beta coefficient (standard error)	OR	Beta coefficient (standard error)	OR	Beta coefficient (standard error)	OR
Sentence severity (reference case high)
Non-custodial/suspended sentence order	−0.055 (0.0532)	0.946 n.s.	−0.063 (0.053)	0.939 n.s.	−0.095 (0.055)	0.909 n.s.
Low severity	0.914 (0.054)	2.495***	0.907 (0.054)	2.476***	0.875 (0.055)	2.399***
Medium-low severity	2.127 (0.055)	8.390***	2.120 (0.055)	8.329***	2.090 (0.057)	8.082***
Medium-high severity	4.448 (0.067)	85.447***	4.440 (0.067)	84.809***	4.415 (0.068)	82.641***
Female	−0.557 (0.046)	0.573***	−0.645 (0.054)	0.525	−0.646 (0.054)	0.524***
Age 25–34 years	0.037 (0.028)	1.038 n.s.	0.037 (0.028)	1.037 n.s.	0.038 (0.028)	1.038 n.s.
Age 35–44 years	0.024 (0.035)	1.024 n.s.	0.024 (0.035)	1.025 n.s.	0.026 (0.035)	1.026 n.s.
Age 45–54 years	−0.104 (0.046)	0.901 n.s.	−0.105 (0.046)	0.900 n.s.	−0.102 (0.046)	0.903 n.s.
Age 54+ years	−0.366 (0.085)	0.693***	−0.365 (0.085)	0.694***	−0.365 (0.085)	0.694***
Crime type
Reference category ‘Other’
Affray	−0.017 (0.044)	0.983 n.s.	−0.018 (0.044)	0.982 n.s.	−0.048 (0.049)	0.953 n.s
Common assault	−0.657 (0.054)	0.518***	−0.659 (0.054)	0.517***	−0.647 (0.059)	0.524***
GBH (s18)	6.063 (0.080)	429.462***	6.060 (0.080)	428.380***	5.986 (0.088)	397.884***
GBH (s20)	1.432 (0.044)	4.186***	1.430 (0.044)	4.179***	1.346 (0.049)	3.843***
ABH (s47)	0.325 (0.039)	1.383***	0.323 (0.039)	1.381***	0.295 (0.043)	1.343***
Guilty plea	−0.168 (0.033)	0.845***	−0.169 (0.033)	0.845***	−0.168 (0.033)	0.845***
Previous convictions
Reference category 1–3
None	−0.396 (0.031)	0.673***	−0.396 (0.031)	0.673***	−0.397 (0.031)	0.672***
4–9	0.356 (0.036)	1.427***	0.357 (0.036)	1.429***	0.360 (0.036)	1.433***
10+	0.576 (0.051)	1.778***	0.575 (0.051)	1.776***	0.583 (0.051)	1.791***
Not answered	0.147 (0.090)	1.159 n.s.	0.146 (0.090)	1.157 n.s.	0.156 (0.90)	1.169 n.s.
Aggravating factors
Location of the offence	0.234 (0.031)	1.264***	0.233 (0.031)	1.262***	0.235 (0.031)	1.265***
Timing of the offence	0.041 (0.033)	1.042*	0.043 (0.033)	1.044 n.s.	0.040 (0.033)	1.041 n.s.
Ongoing effect on victim	0.615 (0.029)	1.850***	0.615 (0.029)	1.850***	0.609 (0.029)	1.838***
Offence against those in the public sector/service to public	0.240 (0.048)	1.271***	0.240 (0.048)	1.271***	0.255 (0.048)	1.290***
Presence of others	0.222 (0.030)	1.248***	0.222 (0.030)	1.249***	0.223 (0.030)	1.250***
Gratuitous degradation	0.612 (0.068)	1.843***	0.612 (0.068)	1.845***	0.620 (0.068)	1.859***
Victim compelled to leave home	0.440 (0.077)	1.552***	0.440 (0.077)	1.553***	0.445 (0.077)	1.561***
Failure to comply with court orders	0.347 (0.041)	1.415***	0.345 (0.041)	1.412***	0.356 (0.041)	1.427***
On licence	0.712 (0.066)	2.037***	0.712 (0.066)	2.037***	0.715 (0.066)	2.043***
Attempt to conceal/dispose of evidence	1.151 (0.137)	3.160***	1.146 (0.137)	3.147***	1.143 (0.137)	3.135***
Failure to respond warnings/concerns	0.301 (0.080)	1.351*	0.305 (0.080)	1.356**	0.313 (0.080)	1.367***
Offender was under the influence of alcohol/drugs	0.193 (0.028)	1.213***	0.170 (0.029)	1.186***	−0.087 (0.098)	0.917 n.s.
Abuse of power/trust	0.475 (0.065)	1.608***	0.479 (0.065)	1.614***	0.479 (0.065)	1.614***
Exploiting contact arrangements	0.036 (0.227)	1.037 n.s.	0.034 (0.227)	1.035 n.s.	0.035 (0.227)	1.036 n.s.
Previous violence/threats	0.400 (0.047)	1.491***	0.400 (0.047)	1.492***	0.403 (0.047)	1.497***
Established evidence of community impact	0.328 (0.143)	1.388 n.s.	0.332 (0.143)	1.394 n.s.	0.358 (0.148)	1.431 n.s.
Steps take to prevent reporting/ assisting prosecution	0.714 (0.147)	2.042***	0.719 (0.147)	2.052***	0.710 (0.147)	2.034***
TICs	−0.194 (0.446)	0.824 n.s.	−0.198 (0.445)	0.820 n.s.	−0.200 (0.445)	0.819 n.s.
Mitigating factors
No previous relevant convictions	−0.201 (0.038)	0.818***	−0.201 (0.038)	0.818***	−0.199 (0.038)	0.819***
Single blow	−0.685 (0.037)	0.504***	−0.685 (0.037)	0.504***	−0.687 (0.037)	0.503***
Remorse	−0.506 (0.029)	0.603***	−0.505 (0.029)	0.604***	−0.506 (0.029)	0.603***
Good character/exemplary conduct	−0.316 (0.044)	0.729***	−0.314 (0.044)	0.731***	−0.310 (0.044)	0.734***
Determination/demonstration to address addiction/behaviour	−0.889 (0.050)	0.411***	−0.892 (0.050)	0.410***	−0.891 (0.050)	0.410***
Serious medical conditions	−0.779 (0.088)	0.459***	−0.785 (0.088)	0.456***	−0.782 (0.088)	0.458***
Isolated incident	−0.504 (0.041)	0.604***	−0.504 (0.041)	0.604***	−0.504 (0.041)	0.604***
Age/lack of maturity affecting responsibility	−0.175 (0.049)	0.839***	−0.177 (0.049)	0.838***	−0.176 (0.049)	0.839***
Lapse of time (not fault of offender)	−0.637 (0.081)	0.529***	−0.638 (0.081)	0.528***	−0.640 (0.081)	0.527***
Mental disorder/learning disability not linked to commission of offence	−0.704 (0.073)	0.495***	−0.702 (0.073)	0.496***	−0.706 (0.073)	0.494***
Sole/primary carer for dependent relatives	−0.666 (0.075)	0.514***	−0.675 (0.076)	0.509***	−0.672 (0.076)	0.511***
Interaction(s)
Female × under the influence			0.336 (0.130)	1.399*	0.336 (0.103)	1.400*
Affray × under the influence					0.229 (0.117)	1.257 n.s.
Common assault × under the influence					0.005 (0.141)	1.005 n.s.
GBH (s18) × under the influence					0.437 (0.168)	1.547*
GBH (s20) × under the influence					0.429 (0.111)	1.536 n.s.
ABH (s47) × under the influence					0.229 (0.105)	1.258 n.s.
Model fit (akaike information criterion)	68,053		68,045		68,023

n.s. = not significant. Significance levels denoted by *p < 0.0, **p < 0.001, ***p <.0001.

In these models, all mitigating factors are significant and in the expected negative direction, decreasing the probability of a custodial or more severe sentence. All aggravating factors (including intoxication) are in the expected positive direction (increasing the probability of a custodial or more severe sentence) although a few were not statistically significant, namely the timing of the offence, where the defendant had exploited established contact arrangements, where there was evidence of the offence having impacted on the community and whether other offences had been taken into consideration (TIC)⁴.

Further, the effect of age increases at first and then tails off in older age in accordance with expectations based on the well-established age–crime curve (dating back to Quetelet 1831/1984). The age–crime curve relates to well-established findings about the relationship between age and criminal behaviour that point to an increase in offending behaviour during early adolescence (from around the age of criminal responsibility), which peaks in teenage years before declining from the early 20s (Loeber and Farrington 2014). This relationship is also approximately mirrored in the ‘drinking arc’, which follows a similar trajectory (Maggs and Schulenberg 2004).

Effects associated with different offence types are in line with expectations, being most influential factor in determining the outcome, and relating to the severity of the offence. The effect of previous violent offending increases with the number of previous offences. Thus the less serious the offence and the fewer previous convictions will result in lesser sentence severity and fewer custodial sentences as we might expect. If a defendant enters a guilty plea, this reduces sentence severity on average, but not necessarily probability of custody, which may in fact be slightly increased. This is presumably due to sentencing practitioners having limited discretion in determining the appropriateness of a custodial sentence where there is an admission of guilt for more serious offences.

First impressions: Females less likely to attract a custodial or severe sentence

Model 1 (Table 3) presents the main effects of the explanatory variables and highlights females are less likely to receive a custodial sentence, controlling for the relevant case characteristics (OR = 0.48). Results in Model 1 Table 4 show females are also less likely to attract a more severe sentence (OR = 0.57), when only the main effects of these covariates are considered. On initial inspection then this would offer strong support for the leniency hypothesis. However, to assess with more nuance the contribution of these factors in combination, further multiplicative interaction effects are necessary to account for the effect of being both female and intoxicated. This is assessed in the second set of models discussed later (Model 2 in Tables 3 and 4).

On second glance: Leniency is less pronounced where intoxication features

In both logistic and ordinal models, when controlling for all other relevant factors, a significant positive interaction term between being female and intoxicated suggests that the effect of intoxication is amplified for female defendants compared to when only examining main effects, both in their probability of attracting a custodial sentence (OR = 1.42; Model 2 Table 3) and in terms of sentence severity (OR = 1.40; Model 2 Table 4)⁵. However, the magnitude of this aggravation is still not on par with outcomes associated with males who are intoxicated (see Table 5).

Table 5

Probability of custody for ABH

Sentencing factors present in case	Probability of custody
Male (reference case)	0.492
Female	0.297
Male and Intoxication	0.549
Intoxication, female and interaction	0.431

To illustrate the magnitude of the effect of sex on the probability of sentence, severity for varying scenarios for an offence of ABH are displayed in Table 5, calculated based on the formula p = exp (B₀ + B₁x)/(1 + exp (B₀ +B₁x)), where p is the probability of the outcome for a reference category of an ABH offence with no other aggravating or mitigating factors asides the presence of intoxication, B₀ is the intercept, B₁ is the regression coefficient or ‘slope’, and x the value of the covariate under consideration, which converts the log odds to probabilities. From these values, it is possible to see when considering the presence of intoxication and being female as well as the interaction between these factors the probability of a more severe outcome rises (to 0.431 compared to 0.347 when only accounting for main effects) although it is still lower than for males perpetrating this offence type with or without being intoxicated (0.492 and 0.549, respectively). In sum, intoxication does not have the same impact for male and female defendants. Any lenient effect is effectively halved when intoxication features in female offending once an interaction is accounted for. Indeed, the aggravation for intoxication applied for females is over twice that applied to male defendants.

Table 6 displays the probabilities for different severity outcomes based on the presence of the main effects and interaction term. It highlights more severe outcomes are probable for female defendants where intoxication features when an interaction term to account for the concomitant effect of being female and intoxicated is included. This figure also seems to suggest where an offence warrants a more severe sentence the role of intoxication has less influence in determining the sentence outcome.

Table 6

Probabilities of sentence severity

Probability	No intoxication (male)	Intoxication (male)	Intoxication + female	Intoxication × female
p (non-custodial/suspended sentence order)	0.484	0.527	0.369	0.450
p (low severity)	0.713	0.746	0.606	0.683
p (medium-low severity)	0.893	0.908	0.841	0.879
p (medium-high severity)	0.988	0.990	0.981	0.987

Interaction terms in both the logistic and ordinal models indicate increased punitive sanctions for female defendants who are intoxicated (compared to when simply considering main effects of these isolated factors). Whilst the magnitude of these increases does not lend sufficient support for the doubly deviant hypothesis in either model, the magnitude of any effect of the leniency hypothesis is not as strong as it first appears.

Crime-specific punishment?

To further explore how sentence outcomes might be moderated by the presence of intoxication as an aggravating factor conditional upon crime type, both the logistic and ordinal models were also run to include interaction terms between these two contextual variables (Model 3 in both Tables 3 and 4). In these models, the main effect of intoxication is no longer significant, but several of the interaction terms with crime type are. Interpretation of these interaction terms is not straightforward. Although there is evidence of a differential influence on intoxication for different crime types, it is not simply a case of the more serious an offence, the more intoxication aggravates. Indeed, only section 18 offences displayed a significant interaction term in the ordinal model (Model 3, Table 4) and all but common assault were significant in the logistic model (Model 3, Table 3). This ‘noise’ may be down to unobserved confounding factors such as other extralegal factors not accounted for or differential practices by judges, which cannot be accounted for in the current model. Of note, however, is the significance of the sex × intoxication interaction in both models (Model 3 in both Tables 3 and 4), continuing to point to the strength of the gendered impact of intoxication in determining sentence outcomes, regardless of the offence type and/or severity.

Discussion

There has been little attention paid to how alcohol intoxication operates in determining sentences for female defendants—not least due to limited ongoing data captured on female offenders (Corston 2007; House of Commons 2014; MoJ 2014; Hedderman and Barnes 2015). In offering insights into the role gender and intoxication play in shaping sentencing outcomes, this study adds to the literature on gender disparities in sentencing as well as scholarship of how cases of alcohol-related violence are processed through the criminal justice system. Its main finding is that intoxication increases sentence severity more so for women than for men (both in terms of the probability of custody and severity of the sentence dispensed). This represents an important contribution, as to date little was known until now about how alcohol intoxication shapes sentence outcomes differentially for males and females.

Is a ‘Malevolent Assumption’ justified?

In line with guidance issued by the Sentencing Council (2011), intoxication increases both the probability of custody and the severity of the sentence for both males and females. However, any rationale or justification for intoxication as aggravation is surprisingly absent when it comes to sentencing guidance (Dingwall and Koffman 2008). There is no guidance as to how and when intoxication ought to aggravate or indeed by how much. Although such guidance may be welcomed in relation to the principle of ‘legal certainty’ and by those favouring a degree of predictability and equivalence in sentencing, there remain tensions with practitioners who favour flexibility with which to tailor a sentence to a specific case (Padfield 2011), as these are key to responding to the complex needs of female or other vulnerable defendants. Although sentencing guidelines are aimed at reducing bias in sentencing, they may indirectly standardize practice in a way that is contrary to gender equality and the ideals of justice, as criminologists have previously noted the adoption of determinate sentencing structures brings about a corresponding ‘equalization’ of justice resulting in higher female incarceration rates (Daly and Tonry 1997).

Is chivalry dead?

On first inspection of the CCSS data, exploring only main effects in a regression framework, one could be forgiven for assuming a pronounced leniency between male and female defendants. This would accord with expectations of what we know about female offending more generally, it is often less serious (MoJ 2014), often supports male offending of a more serious nature (Broad 2015), and female routes into offending are often the result of continual cycles of abuse (Chesney-Lind and Pasko 2004).

However, on second inspection—when accounting for an interaction between gender and intoxication in sentencing practice—the extent of this leniency is, although still present, reduced. Any leniency afforded to female offenders in determining sentence severity is effectively halved where intoxication features. Although these findings offer support for the presence on leniency for female defendants in the case of assault offences, the extent of leniency afforded appears contingent on the presence of intoxication in their offending behaviour.

These findings likely represent outcomes shaped by normative views of femininity and intoxication and thus a gendered interpretation of role of intoxication in offending in which females are viewed as more ‘deserving’ of punishment and potentially more dangerous given their intoxication (drunk and thus doubly deviant). Also evidenced by the fact that several of the interactions between offence type and intoxication were also significant. Punishment is thus being dispensed unfairly where intoxication features in female offending (doubly jeopardy). This is of concern in its own right, but also because women coming to the attention of the courts are likely those who are more socially and economically deprived and so likely face multiple disadvantages, which social and distributive conceptions of justice seek to redress (Corston 2007). Indeed, it may be because of this very fact they are seen as the ‘type’ of women more deserving of punishment by the courts (see Carlen 1998). That is, their ‘troublesome’ as opposed to their ‘troubled’ status is foregrounded (Gelsthorpe and Loucks 1997) and serves to further disadvantage them and control their drinking.

Including interaction terms to account for the differential application of sentencing factors based on the sex of the defendant, in this case intoxication, is key to determining the presence of gendered administration of sentencing and represents an effort to model the inherent social complexity of the gender–alcohol–punishment relationship. Earlier studies employing interactions with gender have only done so based on the crime type (e.g. Farnworth and Teske 1995; Koons-Witt 2002; Rodriguez et al. 2006). Whilst many scholars are necessarily limited by their study designs (including sample sizes), not accounting for the moderating effect of sex on sentence outcomes represents a general problem of mis-specification, and so a lack of methodological precision. As a result, studies are more likely to reflect spurious associations and overestimate any ‘leniency’ afforded to women. Pioneering this approach has enabled a more specific enquiry into the gendered administration of justice where alcohol intoxication was present.

Study strengths and limitations

In the absence of longitudinal data, this cross-sectional study unpicks some of the complexity inherent in the gendered administration of justice in Crown Court cases. Many studies of gender disparities in sentencing employing quantitative methodologies tend to focus on ensuring equality of outcome, glossing over any consideration of how gender plays out in the courtroom. A key strength of this study, is that it gives due consideration to the likely gendered processes by which outcomes are shaped, in relation to intoxication. It also engages critically with the notion that equality of outcome is ‘just’ or symbolic of an absence in sexism within sentencing.

It is worth noting whilst a binary distinction based on the defendants’ sex was employed here as a proxy for gender identity, gender is necessarily a social construct and not just a binary statistic variable (Heidensohn and Silvestri 2012). This measure thus necessarily limits the extent to which the nuances of gendered relations in the context of power and patriarchy can be explored in the Crown Court setting. This is further limited by the absence of contextual detail on sexual orientation, relationship status or prior victimization. It is therefore not possible to glean insights into the complexities of how norms held about relationship status and homosexuality, amongst others, are influencing decisions.

It would have also been useful to explore other dimensions, such as ethnicity and socio-economic status—given these are known to impact sentencing outcomes (Albonetti 1997, 2002; Steffensmeier and Demuth 2000; Mustard 2001; Everett and Wojtkiewicz 2002; Pasko 2002; Stacey and Spohn 2006) and an extant literature pointing to how intersectionality between these dimensions and gender shapes experiences of violence (Crenshaw 1991; Bograd 1999; Sokoloff and Dupont 2005; Walby et al. 2016; Nowacki 2017). However, further enquiries of these kind were not possible as the CCSS data contained no measure of the defendants’ ethnicity or socio-economic status.

Noting the earlier limitation associated with the intoxication measure outlined in the Methods section, this measure is also hampered in its ability to accurately measure the presence of alcohol intoxication/whether the defendant had been drinking immediately prior to a case, only being cited in instances where the sentencer believes the defendant’s intoxication to constitute aggravation and was taken into account when determining the final sentence. This necessitates several judgements on behaviour of pre-sentence report writers and sentencing practitioners as to the role any alcohol consumption played in the offending behaviour as well as whether any alcohol consumption is deemed to have resulted in a state of ‘intoxication’.

Despite its shortcomings, the CCSS is the most comprehensive data source in England and Wales on sentencing practices as recorded by the sentencing practitioner themselves, providing the opportunity to control for an unprecedented number of sentencing factors impacting outcomes in statistical analyses. It also offers the novel benefit of enabling statistical analysis of female offenders processed by the Crown Court (whose engagement criminal activity is less prevalent in official data). This is afforded predominantly by is wide national coverage and large sample size.

Conclusions: The ambiguous role of alcohol intoxication remains

Pervading norms around alcohol consumption and gender shape sentencing decision making. Findings from this study suggest intoxication remains a contested sentencing factor, as its influence does not uniformly aggravate male and female offending. With this in mind, the Sentencing Council may wish to consider monitoring the use of the aggravation of intoxication in sentencing practice and issuing further guidance as to how this ought to be applied in determining sentence outcomes for both male and female defendants.

Given that ongoing collection of data about sentencing practices is key to monitoring the operation and effect of sentencing guidelines (as required by the Sentencing Council under section 128 of the Coroners and Justice Act 2009), it is a shame the CCSS was ended, favouring instead bespoke data collection in both the Crown Court and magistrates’ courts to inform the development of specific guidelines. The revised approach represents a loss of transparency in monitoring how sentencing impacts upon minority and disadvantaged groups and limits insight into changing practices over time (e.g., convergence in sentencing outcomes for males and females). It also limits the ability to look at the role of intoxication across a range of crime types. Both of these limitations are inopportune given findings suggest the gendered administration of justice is an ongoing consideration in sentencing when it comes to the administration of intoxication as a sentencing factor. There is clearly a need to devote further conceptual consideration and empirical research to the issue of alcohol intoxication in sentencing practice so as to further unpick how it impacts punishment in different contexts and for whom.

Acknowledgements

I would like to thank colleagues Jose Pina-Sánchez, Emma Wincup, Louise Ellison and Beverley Clough at the Centre for Criminal Justice Studies (CCJS) at the University of Leeds as well as Sandra Walklate from International Criminological Research Unit (ICRU) at the University of Liverpool for their comments on earlier drafts of this article.

Appendix I: Descriptive statistics: extralegal and legal factors (independent variables)

Independent variables	n	%
Sex
Male	28,004	90.75
Female	2,855	9.25
Age group
18–24 years	11,732	38.02
25–34 years	10,766	34.89
35–44 years	5,060	16.40
45–54 years	2,564	8.31
Over 54 years	739	2.39
Guilty plea	26,302	85.23
Previous convictions
None	16,409	53.17
Between 1 and 3	8,492	27.52
Between 4 and 9	3,987	12.92
10 or more	1,488	4.82
Not answered	485	1.57
Aggravating factors
Location of the offence	12,335	39.97
Timing of the offence	7,055	22.86
Ongoing effect on victim	6,707	21.73
Offence against those in the public sector/service to public	1,719	5.57
Presence of others	6,676	21.63
Gratuitous degradation	826	2.68
Victim compelled to leave home (domestic violence in particular)	635	2.06
Failure to comply with current court orders	2,274	7.37
On licence	771	2.50
Attempt to conceal/dispose of evidence	229	0.74%
Failure to respond warnings/concerns	576	1.87
Offender was under the influence of alcohol/drugs	7,460	24.17
Abuse of power/trust	976	3.16
Exploiting contact arrangements	64	0.21
Previous violence/threats	1,787	5.79
Established evidence of community impact	195	0.63
Steps take to prevent reporting / assisting prosecution	177	0.57
Taking offences into consideration (TICs)	22	0.07
Mitigating factors
No previous relevant convictions	7,551	24.47
Single blow	4,535	14.69
Remorse	10,416	33.75
Good character/exemplary conduct	4,731	15.33
Determination/demonstration to address addiction/behaviour	2,608	8.45
Serious medical conditions	868	2.81
Isolated incident	4,857	15.74
Age/lack of maturity affecting responsibility	2,400	7.78
Lapse of time (not fault of offender)	891	2.89
Mental disorder/learning disability where not linked to the commission of the offence	1,137	3.68
Sole/primary carer for dependent relatives	1,167	3.78

Footnotes

1

The explanations for this have been extensively discussed elsewhere and include reasons such as worse treatment and amplified stigma, poorer prisons and greater loss in terms of separation from children and breakdown of family life (see Fawcett Society 2006; Corston 2007; Heidensohn 2010; House of Commons 2014). Women also experience custody more harshly than men, thus representing a more severe form of punishment for female offenders (Heidensohn 2010).

2

Indeed, only 30 per cent of professional judges in England and Wales are female (Bowcott 2016; Council of Europe 2016).

3

This category includes intent to resist arrest, assault on a police constable, cruelty/neglect of a child, harassment, other assault and public order, S.4 Public Order Act (POA), S.4A POA, S.5 POA and violent disorder offences.

4

The court has discretion as to whether or not to take other offences into consideration where an offender admits the commission these in the course of sentencing proceedings. (https://www.sentencingcouncil.org.uk/wp-content/uploads/Definitive_guideline_TICs__totality_Final_web.pdf).

5

Inclusion of the interaction improved the model fit in both instances (logistic: x2 = 7.8822, df = 1, p = 0.005 ordinal: x2 = 10.578, df = 1, p = 0.001) and does not dampen the effects of the individual covariates.

Author notes

Dr Lightowlers is a senior lecturer who employs a variety of secondary sources with which to conduct research on issues of alcohol, crime and justice; in particular examining the sociocultural dimensions of the alcohol–violence relationship and the processing of such cases through the criminal justice system