Locating the Judge within Sentencing Research

Research into sentencing is undertaken from a range of theoretical, disciplinary and methodological perspectives. Each approach offers valuable insights, including a conception of the judge, sometimes explicit, often implicit. Little scholarly attention has been paid to directly interrogating the ways in which different research traditions construct the judge in the sentencing process. By investigating how different research approaches locate the judge as an actor in sentencing, theoretically and empirically, this article addresses that gap. It considers key examples of socio‐legal scholarship which emphasise the judge as operating within experiential, emotional and social, as well as legal dimensions. This growing body of research offers a more social, relational and interactive understanding of the judge in sentencing, extending and complementing the valuable, but necessarily limited, insights of other research approaches about the place of the judge in sentencing.


Introduction
Research from many scholarly traditions contributes to understanding sentencing as a legal and social process. These various research approaches rely largely on outputs produced by judges, more directly when analysing specific judicial sentencing decisions, somewhat less directly when using statistics aggregating large numbers of judicial decisions. Although each approach offers a conception of the judge, sometimes explicitly, often implicitly, little scholarly attention has been paid to the ways in which different kinds of research construct the judge in the sentencing process. This article asks how these different research approaches locate the judge as an actor in sentencing, directly or by implication. It then examines key examples of research illustrating what sentencing research looks like when it focusses more directly on the judge as a social participant in an interactive sentencing process.
To investigate these constructions of the judge 2 the article first outlines conceptualisations of the judge in conventional legal sentencing research. This research tends to treat the judge as a conduit through which the law operates, albeit recognising a legally confined space allocated for judicial discretion. The second research approach discussed is quantitative social science research which draws attention to sentencing patterns. In this research tradition, the judicial officer becomes a mechanism through which social forces operate, often within the discretionary space available for judicial sentencing decisions. Each approach can be seen as regarding the judge primarily as an abstraction.
Next, this article considers key examples of socio-legal scholarship which explicitly identify sentencing as an interactive social process (for example , Booth 2016;Tata 2007;Travers 2007). This research treats the judge as an embodied human participant, rather than as an abstraction or construct, and recognises that the judge operates within a distinct context comprising experiential, emotional and social dimensions. Such research usually entails empirical methods investigating sentencing from the point of view of judges (individually or collectively) and considers the situational context and subjectivity of judging. It offers a more social, relational and interactive understanding of the judge in sentencing.
Careful examination of these different strands of research confirms the centrality of the judge as the nexus between law and the sentence actually imposed, though the construction of the judicial place varies. These constructions include the judge as a legal abstraction, as a conduit for law or larger social forces, and as an active human and social participant, emotionally connected within an interactional social process, while still constrained by law and committed to achieving legally appropriate outcomes.

The judge in conventional legal sentencing research
We begin by exploring how sentencing research in law constructs the judicial place in sentencing. In doing so, we consider both doctrinal and policy research, primarily from Australia, as well as from the UK and the USA.

Doctrinal research
In legal doctrine, the judge is the central figure in sentencing. The judge, and only the judge, exercises legal authority to impose sentence after conviction for a crime (Campbell 2004;Lovegrove 1997;Mackenzie et al. 2010;Warner et al. 2002). Doctrinal research concentrates on existing legislation and case law as inputs to judicial decision making. Determining sentence in a particular case requires the individual judge to consider a range of legally applicable factors. This legal complexity is reflected in extensive doctrinal research aimed at establishing the law which the judge must consider when sentencing. This research is produced by academics, practitioners and occasionally judicial officers. It includes material such as annual reviews of sentencing legislation and case law (Bartels 2016;Warner 2013), textbooks (Ashworth 2015;Mackenzie et al. 2010), practice manuals (Freiberg 2014;Haines et al. 2015) and benchbooks or looseleafs IJCJ&SD 48 Online version via www.crimejusticejournal.com © 2017 6(2) (Campbell 2004;Sentencing Council for England and Wales 2016;Thomas 2015). These materials are directed primarily at the judiciary and the legal profession, to inform them about the legal and procedural requirements necessary for judges to impose legally correct sentences and reducing the likelihood of sentences that are legally incorrect.
Statute and case law identify general sentencing principles (for example, proportionality, just deserts), purposes (for example, deterrence, rehabilitation) and aggravating and mitigating factors specific to the offence (for example, nature/extent of harm) or to the offender (for example, contrition, cooperation with authorities). The judge must then apply and weigh these factors to decide the sentence in a specific case. In general, the law does not specify the relative weighting or hierarchy of the multiplicity of elements. Because of the need to consider so many factors, reflecting 'overlapping, contradictory and incommensurable objectives' (Spigelman 1999: 6), sentencing law acknowledges a wide range of judicial choice in determining a sentence (Mackenzie et al. 2010;McGarrity 2013;Roberts 2011;Tata and Hutton 2002). In Australia, this process is described as an 'instinctive synthesis'. In Markarian v The Queen Justice McHugh states: By instinctive synthesis, I mean the method of sentencing by which the judge identifies all the factors that are relevant to the sentence, discusses their significance and then makes a value judgment as to what is the appropriate sentence given all the factors of the case. Only at the end of the process does the judge determine the sentence. (McHugh 2005: 378) Australian judicial officers describe this as a balancing process (Mackenzie 2005), while research in the UK finds varying views among the judiciary as to whether sentencing is predominately a structured or intuitive process (Ashworth et al. 1984;Kapardis 2014;Millie et al. 2007).
The notion of instinctive synthesis is controversial (Bagaric 2015;Markarian v The Queen 2005). Instinct implies intuition or an innate quality, not subject to external investigation. Fox and Freiberg comment: '[t]he opaqueness produced by this "instinctive synthesis" method of arriving at a sentence' makes it difficult to unravel 'the individual threads of reasoning that support it ' (1999: 196). This emphasis on space for judicial discretion or choice suggests considerable judicial agency in sentencing, and these descriptions imply an undeniably human process, rather than a mathematical algorithm. 3 Nevertheless, they generate an abstract construction of the judge, rather than recognising a distinct and proper judicial humanity or individuality, exercised within an interactive process. The judge remains a conduit through which factors are weighed, the law is applied, and a sentence is rendered, akin to a 'black box' (Hogarth 1971: 35).
Consistency is an important principle of justice: similarly situated people who commit similar crimes should receive similar penalties (Ashworth 2012;Hili v The Queen 2010;New South Wales Law Reform Commission 2013;Tata 2013;Warner et al. 2002). Apparent inconsistency can generate concerns about judicial error in a particular case, whether undue lenience or harshness. Variability in outcomes suggests that, while the judge can still be understood metaphorically as a conduit, the transmission can sometimes be erroneous. Correction may be needed, through appeal to a higher court in a particular case or more generally perhaps through changes to the law.
One legal mechanism to address alleged legal error in sentencing in a specific case and to foster overall consistency is through appellate judgements on sentencing. Some of these may be recognised as guideline judgements, decisions made by a court on appeal setting out general principles of sentencing and the range of penalties that apply to a particular offence (Mackenzie et al. 2010;Millie et al. 2007). More radical legal reforms impose structured sentencing processes to promote consistency in sentencing, including determinate sentencing, mandatory minimum sentences, mandatory sentences, presumptive grids, guidelines or baseline sentencing (Ashworth IJCJ&SD 49 Online version via www.crimejusticejournal.com © 2017 6(2) 2012; Ashworth and Roberts 2013;Mackenzie et al. 2010;Ostrom et al. 2003;Quilter 2014;Roberts 2012).
In some jurisdictions, considerable legal statistical research about sentencing patterns is produced largely by government agencies, primarily designed to inform the judiciary and the legal profession about the appropriate range or 'tariff' for sentencing (Sentencing Advisory Council 2007;Summary Justice Review Committee 2004). This type of research provides a guide to the judge in making appropriate sentencing decisions at the outset and to the legal profession to frame arguments to support a particular sentence or sentencing range or in relation to appeals against sentence (Hutton 2013). This research also aims to promote consistency in sentencing (Judicial Commission of New South Wales 2015; see also Hutton et al., 1995). Further discussion of quantitative analysis of sentencing outputs to identify the nature or causes of disparity is presented below.

Policy and critique
Some primarily legal sentencing research is more extensively or explicitly focussed on sentencing policy and critique of existing law (Ministry of Justice (UK) 2015). In relation to judicial discretion, those who seek to restrict sentencing discretion see the capacity for judicial choice, exercised through an 'intuitive' or 'instinctive' 'balancing' process, as the source of inconsistency and inappropriate variation. To the extent that the judge is individualised or humanised, the judge is perceived as a source of fallibility in the process of applying the law to generate decision in particular case. Under this analysis, the judge as conduit must be further constrained by sentencing law, with the expectation that sentencing will become more consistent, predictable and less variable (see Tata and Hutton 2002).
The alternate view, valorising judicial discretion, appears to construct the judge differently. Arguments for judicial choice depict the judge as an instrument of law, compelled to apply a complex set of requirements to a complex set of facts, which requires a range of judicial choices from limited options, with the goal of imposing an appropriate sentence on the individual offender being sentenced (see Tata and Hutton 2002). This process will, inevitably, result in some variations in outcome, as the circumstances of offenders and their crimes vary widely, but this does not mean that the judge is acting freely or individually or without constraint. Variation within the range of legally valid sentences is not (necessarily) an inconsistency which needs correction (Hili v The Queen 2010).
Under either view, the judge is still essentially a conduit but with a narrower or wider scope for legally correct-or at least legally acceptable-outputs. The focus is on law rather than on the judge as a distinct actor; the judge is the means through which any rules and principles would be given effect. This strand of legal sentencing research is theoretical and jurisprudential, focussing more on abstract sentencing principles and purpose, and less on their application by the judiciary in practice (Ashworth, von Hirsch and Roberts 2009;Braithwaite 2002;Duff and Garland 1994;Tonry 1996Tonry , 2009). This literature tends not to consider the judge at all, or sees the judge as almost entirely constrained or 'locked into… sentencing arrangements' (Henham 2012: 17-18; see also Doak 2012).
Legal research into sentencing focuses on legislation and case law or sentencing principles derived from existing law. This material can be considered as inputs to judicial sentencing decision making, constituting sources on which judges must or may rely in formulating a legally correct sentence. Much legal research therefore analyses sentencing decisions to investigate whether legal inputs are used appropriately in sentencing decisions. Another important strand of legal research addresses conceptual and jurisprudential issues especially in relation to (in)consistency of sentencing outcomes, which may include investigating the policy or value preferences of judges exercising sentencing discretion (Tata and Hutton 1998 2009). However, under either research focus, the judge is largely-and perhaps necessarilyregarded or constructed as an abstract entity or conduit, which is, or should be, entirely constrained by positive law.

The judge within social science research: Sentencing patterns
This section asks how quantitative social science research locates the judge in sentencing. Sentencing research adopting a quantitative design tends to undertake complex modelling and regression analyses on large data sets assembled from officially collected court statistics, seeking to identify the extent and nature of any sentencing disparities and the mechanisms that produce them. (See Pina-Sánchez and Linacre (2016) for a comprehensive review of the methodologies used in such research.) Much of this research has been undertaken in the United States where there is a strong tradition of empirical, quantitative research in criminology (Adler and Simon 2014). For example, the US Sentencing Commission, which promulgates and updates the federal sentencing guidelines, collects information on all cases and offenders punished in US federal courts. This information is used to assemble longitudinal and cross sectional data sets (Light 2014 (Steffensmeier and Demuth 2001;Ulmer and Johnson 2004). Sometimes the research focus is broader and incorporates penalties other than imprisonment (Johnson and Di Pietro 2012) and/or departures from sentencing guidelines (Johnson 2005). An overriding interest is the ways in which social and economic inequalities are reproduced, even magnified, through the criminal justice process. Baumer calls this the 'modal' approach to studying race and sentencing (2013: 234).
Sentencing disparities in the US take place against the backdrop of a highly racialised society; the key independent, and explanatory, variable has been White/African American (Seron 2011). Over time, the independent race/ethnicity variables have been specified to include Hispanic (Steffensmeier and Demuth 2001), Native Indian (Franklin 2013), Asian-American (Johnson and Betsinger 2009) and, more recently, citizenship status (Wolfe et al. 2011). These studies adopt sophisticated multivariate modelling and detailed analyses to disentangle 'warranted and unwarranted disparity' (Bushway and Piehl 2001: 734). This kind of approach has been characterised as an input-output methodological design (Hogarth 1971), or 'jurimetric analysis' (Hunter et al. 2008: 78), whereby sentencing outcomes are attributed to the judge following identification of sentencing patterns, but without direct empirical investigation of individual judicial attitudes, experiences, accounts and/or practices (for exceptions see, Gibson 1978 andJagger 2006). The judge is again constructed as the mechanism for the reproduction, even magnification, of factors which are shown, statistically, to affect sentencing outcomes.
IJCJ&SD 51 Online version via www.crimejusticejournal.com © 2017 6(2) To enhance the explanatory power of models, some researchers incorporate measures of courtroom context, including judge level variables, and external environments, such as political forces (Bontrager et al. 2005;Johnson 2005;Myers and Talarico 1987;Ulmer 2012;Ulmer and Johnson 2004;Ward et al. 2009). Several researchers seek to isolate the effects of the demographic and racial composition of various actors in the criminal justice system, including judges (Spohn 1990;Tiede et al. 2010;Welch et al. 1988;Wooldredge 2010), prosecutors (Ward et al. 2009), and the local legal profession (King et al. 2010). Judge level variables including gender, race/ethnicity, age and time on bench serve as proxy measures of judicial decision making (Johnson and DiPietro 2012).
These measures of judicial officers' characteristics do not, however, provide measures of judicial decision making itself (Baumer 2013). The dearth of available measures of judicial decision making is a source of frustration among researchers in this tradition, who are often left to theorise or speculate on the role, behaviour or attitudes of judges in the production of sentencing outcomes (Johnson and DiPietro 2012). For example, 'the absence of minority representation in a profession may permit decision makers to unobtrusively act on stereotypes and assumptions about race and criminality [emphasis in original]' (King et al. 2010: 7; see also Ulmer and Kramer 1996). Similarly, a study of the use of non-custodial sentences concludes: 'These results are consistent with theoretical perspectives that emphasize increased perceptions [among sentencers] of culpability and dangerousness for young, male, and minority offenders' (Johnson and DiPietro 2012: 837). Johnson and Betsinger also conclude that their findings regarding the disparate treatment of Asian-American defendants 'lend credence' to theories of 'the situational meaning of race and ethnicity' in sentencing (2009: 1079). In other words, these findings regarding sentencing disparity offer a particular construction of judicial decision making, even though their data do not, and usually cannot, directly or completely test it; the characterisation of the judge is theoretical, implicit and abstracted.

Mechanisms of sentencing disparity
This quantitative approach to analysing sentencing patterns raises the question of the mechanisms by which these variables alter sentencing outcomes (Steen et al. 2005). A key mechanism for the continuing disparity appears to be judicial discretion (Henham 2012;Seron 2011). Discretion is cast as an opportunity for a judge to insert personal or political preferences into sentencing, aligning with the image of the judge as the fallible conduit. Presumptive sentencing guidelines (mandatory and voluntary) have been implemented to limit that discretion and so reduce unwarranted sentencing disparities. Studies of sentences that depart from guidelines or recommendations aim to isolate the sentencing disparity attributable to the sentencing judge (Bushway et al. 2012;Bushway and Piehl 2001;Fischman and Schanzenbach 2012). For example, using data from the US state of Maryland, Bushway and Piehl find 'that African Americans have 20 per cent longer sentences than whites, on average, holding age, gender, and recommended sentence length from the guidelines constant ' (2001: 761). This continuing (unwarranted) sentencing disparity is attributed to the extent of discretion remaining with the sentencing judge.
However, neither the scope of judicial discretion nor individual judicial characteristics provide a complete explanation, as other researchers have noted. In the 1980s, Frazier and Bock (1982) argued that multivariate modelling may not adequately explain decision making in sentencing, and that research needs to re-situate the judge, and account for the many social dimensions of sentencing. Researchers acknowledge that variations in sentencing outcomes represent the actions of a number of participants in the criminal justice system, not just the sentencing judge (Bushway and Piehl 2001). Court outcomes can mask subtle cumulative inequalities embedded in other processes such as police contact, arrests, prior imprisonment (Zatz 1987)  courtroom decision-making processes [that] still presents important opportunities for variation in judicial sentencing to occur' (Johnson 2006: 267). Because presumptive sentencing guidelines 'are filtered through individual courtroom actor interpretations, however, and because they are coloured by informal, locally varying courtroom norms, it is not surprising that they have failed to eliminate judge and court variation in sentencing' (Johnson 2006: 291). As Tata et al. note, the discourse of judicial ownership of sentencing is strong, and judicial perceptions can be 'shifting target[s]', their dialogue continually moving between 'reference to abstract principles and the particular unique case' (2008: 850).

Judicial subjectivity
Some quantitative research on sentencing patterns attributes unacceptable disparities in large volumes of cases to judicial reliance on stereotypes or bias, which may be unconscious or unintentional (Albonetti 1991;Albonetti 1997;Hagan 1974;Steen et al. 2005;Steffensmeier and Demuth 2001;Steffensmeier et al. 1998;Wooldredge et al. 2005). Judicial cognition and bias are theorised as producing the associations between the key variables in the aggregate sentencing decisions across a jurisdiction or over time.
The perspective developed in criminology known as focal concerns attributes certain considerations to decision makers (Ulmer 2012). This attribution is inferred from the statistical findings which link independent variables with the dependent outcome variables. Because recidivism is never fully predictable, and defendant character cannot be known entirely, court actors make assessments of dangerousness, blameworthiness or other relevant factors, partially based on attributions about the defendant according to their gender, employment status, family situation and race (Daly 1989;Light 2014;Steffensmeier and Demuth 2001;Steffensmeier et al. 1998). Ulmer (2012) and Kramer and Ulmer (1996) argue that legal factors, such as criminal history and crime severity, are used informally by judges as heuristics to assess blameworthiness and risks to community protection.
Sometimes, the attribution of statistically relevant factors to judicial subjectivity is direct. Based on observed statistical relationships between prior record, race/ethnicity, weapon use, pre-trial release outcome and sentence severity, Albonetti proposes a specific judicial subjectivity in sentencing decisions, reflecting a complex relationship between uncertainty avoidance, racial stereotypes and levels of punishment: '[w]hen judges attribute stable, enduring causes of crime to black offenders, the defendant's race affects the exercise of discretion' (Albonetti 1991: 261). However, she relies on no data directly from judges. In Australia, using state sentencing data and statistics Jeffries and Stenning (2014) find that, when Indigenous Australians appear before the court under circumstances similar to their non-Indigenous counterparts, as measured by key variables, a term of imprisonment was as likely or more likely in some states, but less likely in South Australia. Bond, Jeffries and Loban suggest that judges who give lesser sentences seem to make 'allowances for the circumstances of Indigenous offenders' and view the cohort as 'less blameworthy than their non-Indigenous counterparts, possibly due to Australia's legacy of colonisation, associated Indigenous social and economic marginalisation, and the potential of imprisonment to exacerbate this ' (2011: 66).
It is important to note the caution offered by Light, who concludes: 'I could not definitively determine the mechanisms driving the effects shown, a limitation of nearly all quantitative criminal justice research ' (2014: 473), suggesting that '[f]uture research would do well to explore this possibility through qualitative research with … judges'.
One research approach that aims to better understand any subjective or internal mechanisms that produce sentencing disparity-the fallibility in the conduit-is to investigate directly the psychology of judicial decision making. Social psychologists rely on experimental research designs to identify the role of cognitive shortcuts, or heuristics, in judicial decision making (Dhami  Guthrie et al. 2001;Ostrom et al. 2003;Rachlinski 1998). Such studies often use hypothetical scenarios and administer questionnaires to measure the influcence of cognitive illusions on the decisions judges make. Ultimately, this psychological approach focuses on the way decision makers organise information from various sources, as something that happens in their minds. This characterisation of judicial decisions as an internal cognitive process raises questions about how transferable these findings are to actual sentencing decisions in courts. In these primarily psychological research approaches, the judge remains an abstract self-contained entity, and the decision-making process is primarily an individual cognitive exercise. Travers (2007) points out problems with both quantitative sentencing analysis as well as some investigations of judicial subjectivity: a central problem with this … literature is that measuring inputs and outputs, in so far as this is possible, cannot explain how any particular decision gets made ... what happens in court, or what matters to the judge and other professionals involved in a particular legal case disappears or becomes irrelevant when the researcher tries to explain this using statistical methods, or even when interviews are employed to explore general sentencing principles' (Travers 2007: 24).
These research approaches remove the sentencing judge from everyday courtroom interrelations and cannot fully account for the collective and social dimensions of sentencing (Tata 2007).

Relocating the sentencing judge
What does sentencing research look like when directly considering the judge as a social actor and participant in an interactive sentencing process? There is growing scholarship on what judges actually do and how they experience their work. This research often draws on data obtained directly from judicial officers, including their accounts of sentencing and judicial decision making. This research complements and extends existing legal and quantitative research by investigating subjective experiences of sentencing, the activities undertaken by judges and the distinctive skills and techniques of 'judgecraft' (Kritzer 2007: 322; see also Fielding 2011; Moorhead and Cowan 2007;Roach Anleu and Mack 2015;Tata 2007).
This scholarship-driven partly by changes in judicial practice, such as increasing use of problemoriented courts and the rise of therapeutic jurisprudence (see King et al. 2014;Lens 2016; Mack and Roach Anleu 2011)-attaches priority to the social dimensions of sentencing. This research conceives the judge and judging not as an abstraction but, rather, locates the judge, and judicial perceptions and experiences of the sentencing process, as central to the research.

Sentencing as an interactional social process
The significance of the individual judge in sentencing was recognised during the 1970s, in Hogarth's study Sentencing as a Human Process, which did not infer judicial behaviour from sentencing outcomes. His research collected data directly from Canadian magistrates on their attitudes, beliefs and perceptions of the sentencing process, and then correlated them with social characteristics, such as age, ethnicity, religion and geographical location. The focus was on the social antecedents of attitudes which were then 'used to predict judicial behaviour ' (1971: 99). The paradigm remains an individual judge; sentencing in this context becomes an individual cognitive process, driven largely by the social backgrounds and attitudes of judges.
The conception of sentencing as an individual judicial exercise or intellectual struggle has, in the intervening years, increasingly featured in the literature (Hutton 2006;Mackenzie 2005;Tata 2007;Tombs and Jagger 2006 A key thread is the acknowledgement that judicial emotion plays an important role in the practice of sentencing, especially as the interactive moment in the courtroom between the judge and others comes to bear upon the sentence itself. Several socio-legal scholars have investigated such interactive moments, likening the courtroom to a theatre (Ball 1975;Friedman 2001;Grunwald 2012), and legal proceedings to 'a spectacle of legal performance art' (Abrams 1999: 908;see also Peters 2008;Rossmanith 2015;Tait 2002). Within the rubric of theatre, emotions are viewed as routinely contested (Rossner and Tait 2011) and central to judicial decision making.
Scholars have begun to map the ways and the extent to which judges use emotions as a means of managing other courtroom participants, as well as the courtroom processes associated with sentencing. Tata contends that emotions are in fact 'deployed' by judges in ways 'critical to the public performance of sentencing craft ' (2007: 431). Judges actively manage the display of emotions (their own and others) as a means of achieving certain responses in others (Maroney and Gross 2014). Mack and Roach Anleu's (2007: 341) study of Australian magistrates shows that emotions can be a mechanism for expediting court processing and 'getting through the list'. In Scottish courts, such practices are used to create delays or postponements and to manipulate workloads (Tata 2007). Other studies report that judges often manage the emotional tensions in the courtroom, especially those generated by others (for example, victims and families of victims). Judges can become 'active listeners' (Booth 2016: 82), who provide 'a space for the victim's voice in accordance with the legislation' and ensure that expressions of emotions are 'kept within socially approved limits' (Booth 2016: 112;also Schuster and Propen 2010).
Other important research acknowledges that judges operate as individuals and within a wider context-within a social world-which has been shown to influence sentencing decisions. In the courtroom, the judge is one, albeit perhaps the most important, participant in a workgroup, which includes lawyers, defendants, social workers, witnesses, victims and court staff. Eisenstein and colleagues argue that the 'most crucial decisions ' (1988: 37) in courtroom settings become a product of the inter-relationships between workgroup actors, and the shared informal/formal understandings of appropriate case outcomes by all involved (see also Flemming et al. 1992;Mather 1979;Roach Anleu and Mack 2010).
Some empirical research has sought to examine the effects of interrelated cultural, political, international, legal and professional conditions on sentencing across a range of jurisdictions. This includes investigations of professional/occupational subcultures (Eisenstein et al. 1988;Fielding 2011;Huck and Lee 2014;Hutton 2006;Ulmer 1997); the rules, policies, administrative and legal frameworks operating within courtrooms that constrain judges (Beyens and Scheirs 2010;Phoenix 2010); social interactions with other participants (Martyn and Levine 1998; Travers 2007); defendant-related factors (Bouhours and Daly 2007;Ulmer 1997); how other courtroom participants experience the judge in court (Jacobson et al. 2015); how the judge experiences other participants (Rossmanith 2015); and the possible impact of public opinion, however identified or perceived (Mackenzie et al. 2012;Roberts 2008).
The research designs used to explore sentencing's complex experiential, emotional and social dimensions diverge considerably from those used within the legal and multivariate social science traditions. Data are typically sourced directly from courtroom participants (including the judiciary), rather than relying upon outputs such as decisions or aggregate statistics. Research designs include surveys, interviews and focus groups, simulated cases and observational studies of sentencing proceedings, as well as archival analysis of transcripts and other courtroom data. For example, Beyens and Schiers' multi-method study of sentencing judges in Belgium deploys IJCJ&SD 55 Online version via www.crimejusticejournal.com © 2017 6(2) self-administered questionnaires, simulated exercises, semi-structured interviews and focus groups as a means of extracting the multiplicity of judicial and non-judicial 'voices' (2010: 309), emanating from 'historical, legal, local and institutional contexts' (2010: 324) that bear upon decision making. Elsewhere, Bouhours and Daly (2007) analyse transcripts from South Australian sentencing remarks to show how the context and seriousness of the offences, as well as the characteristics of offenders and victims, directly bear upon sentencing decisions. More recently, Huck and Lee's (2014) innovative self-report study of US judicial officers augments conventional multivariate social research by detailing the predictive capacity of situational and contextual social pressures on courtroom participants (especially the desire to be viewed favourably by others the judge cares about) and the effects on sentencing outcomes.
These studies contribute to a narrative about the complex social world of sentencing and its constituent parts. They specify the social role of the sentencing judge as a core (and human) actor, and map active engagement through social interactions that negotiate these conditions (Hutton 2006; see also Bourdieu 1987;Hawkins 2003). Importantly, this research shows that the external conditions present within the social world of sentencing both enable decisions to be made and constrain these decisions. Judges, in exercising their judicial role, must think and act strategically as they negotiate and respond to various conditions and legal requirements. They operate not in abstraction, wholly captured by their legal and social contexts or inputs; nor are they entirely free agents behaving independently of these contexts (Sewell 1992).

Conclusion
This article first describes how different sentencing research traditions construct the judge and judicial decision making and, second, identifies a research strand that locates the judge as a more fully human actor in a complex social and legal process. In one sense, all sentencing research locates the judge as central to sentencing outcomes, exercising judicial authority through which the law operates, and exercising discretion through which social forces operate. However, in much legal and quantitative research, the judge is cast primarily as a conduit, a means through which these other imperatives operate. Understanding the judge's place in sentencing must go beyond the construction of the judge as a conduit for law envisaged by legal analysis or as a mechanism for translating extra-legal factors into the unwarranted sentencing disparities identified in quantitative social science research.
In contrast to these approaches, there is socio-legal literature which confronts these abstractions and undertakes a more direct, grounded investigation into the affective, social and experiential dimensions of judging. This approach increasingly characterises decision making as more than an individual, cognitive and rational calculation. Sentencing is cast as interactive, where the judge attaches emotional significance to pertinent information as a means of negotiating the sentencing processes (Bennett and Broe 2007). This is not to say that the sentencing process is purely an emotional undertaking by an individual judge (Roberts and Bradford 2015). Rather, sentencing necessarily involves striking an intricate balance fusing emotion and legal-rational requirements within the confines of the complex social world of sentencing.
A fuller understanding of sentencing requires research that locates the judge within this complex socio-legal process. It also asks researchers to acknowledge their implicit or explicit conceptualisations of the judge and their role in sentencing. This entails research designs that seek information directly from judicial officers so as to investigate the practical and interdependent dimensions of the sentencing process, which is necessarily absent in research primarily investigating sentencing law and practice, or using aggregate statistical patterns.
Inevitably, legal and quantitative research are limited in their capacity to provide rich explanations for judicial sentencing behaviour or to examine judicial subjectivity. The insights generated by emerging research directions complement, expand and explain findings of legal and quantitative research. Emphasising sentencing as a contingent, independent process, with IJCJ&SD 56 Online version via www.crimejusticejournal.com © 2017 6(2) emotional dimensions, discloses the nature of judging in sentencing as a deeply bounded and simultaneously dynamic, interactional process.