Sensing probation in Canada: Notes on affect and penal aesthetics in risk assessment

Abstract

Based on 6 years of probation practitioner experience in a metropolis of Canada, I provide an autoethnographic account reflecting on my fieldwork as I now commence doctoral studies. Contributing to discussions of experience in the penal atmosphere, I explore personal ethics and values, looking specifically to LSI-R software, where my experience with risk-based programming indicates a subjugation of both supervisees and supervisors. Studying penal aesthetics within the version of the software I used to assess criminogenic risk thus elucidates why evaluators tend to score their risk ratings upward rather than downward. Implications for a desistance paradigm are juxtaposed to the RNR model of offender management, where sensing visual and haptic stimuli pertains to an algorithmic governance mode limiting human connection. I conclude by reflecting on organisational values and behaviour to indicate where therapeutic alliances with criminalised people intersect criminalisation and desistance.

A long-standing thought I grapple with is how painfully un-user friendly our software is at work. Our main risk assessment tool is an application in software which seems like a switchboard of colourful checkmarks. Like little buttons waiting to be pressed. I recall feeling in the very early days of my career how awfully specialised it felt to have this training and knowledge of when to click—or not—the items that represented certain risk factors. … I wonder, should one of these tools be conducted on me by a close friend or colleague, how would the results be similar or different to my clientele? How colourful would the results be? I've always been one to struggle with double negatives, so I naturally hesitate when my selection of a check mark goes from yellow to green when I determine that the corresponding risk factor is present. In contrary determination, when I want to indicate the risk factor is not present, I must double click, and the item turns red. … Indeed, it seems more psychologically satisfying to me that I receive a green [checkmark] upon a single click making my job half as easy—as if green means go (Fieldnotes, 26 May 2020).

Introduction

Practising probation through the assessment of criminogenic risk is a form of job crafting that creates meaning as a form of coping with adversity (Mawby and Worrall, 2013). Self-modification, I argue, is a major limitation and form of liability for assessing criminogenic risk. Probation service users in Canadian urban settings, where I was situated when I wrote the above passage, often represent ‘multiply disadvantaged’ people who experience complex needs such as homelessness (Quirouette, 2016). The job of evaluating them evinces how risk logic intersects probation practice and criminal courts in such assessment (Quirouette, 2018). I observe this in terms of vulnerabilities for both organisational and personal limitations situating liability on criminogenic structures through individual action. For example, where criminal charges laid by probation officers for breaching high-risk probation supervisees in the administration of justice offences potentiate devastation in cumulative effect (Quirouette, 2018), such fundamental principles of the common law tradition—such as the presumption of innocence—appear to be evaporating into thin air (Myers, 2015, 2017, 2019; Pope and Bromwich, 2022; Quirouette, 2018). This has created a morally injurious climate that represents a high-stakes situation for practitioners (Griffin et al., 2019; Shay, 2014) as they experience cognitive dissonance: the state of tension between ideas, attitudes, and beliefs which become psychologically inconsistent with their work (Brown, 2021: 81). The experience of vulnerability is, therefore, a personal emotion that arrives from uncertainty, risk, and personal exposure (Brown, 2021: 13), in which this article warns of the social structures that exacerbate such alarm.

Showcasing value for practitioner experiences in Canada, my aim is to indicate an understanding of pervasive uncertainty based on more than half a decade in the field. As uncertainty is inherent to correctional work, it evinces how vulnerabilities exist in carceral workplaces. The experience of risk as ‘authority without control’ thus pertains to an alarming prevalence of mental disorders among probation practitioners in Canada (Ricciardelli et al., 2021: 16). Thus, I argue practitioners embody good people who become objectified by the dirty work of risk assessment when it is totally institutionalised. My intention is to ‘open a window’ into risk assessment programming vis-à-vis software and the specific design choices regarding its aesthetic appeal by elucidating a sense of probation culture in Canada (Worall et al., 2017). I, therefore, make sense of misrecognition and misrepresentation where folks on probation are seen badly (McNeill, 2018, 2019) as technocratic applications of risk represent a new form of governance which values algorithmic logic and constrains the rationale of practitioners (Pratt and Anderson, 2020).

I begin with literature about criminogenic risk assessment with implications for how penal aesthetics, such as colour, affect communicative meaning in software design. Next, I remark on the value of emotional labour as an analytic resource. Turning to vulnerability, I contextualise findings by Stanley Milgram and Philip Zimbardo about how obedience to authority (eg the Stanford Prison Experiment) implicates a totalising effect of risk as an institution. I argue aesthetics pertain to the social psychology of how probation practitioners sensorially experience vision and touch while assessing risk. I then synthesise the theory about mass probation and totalitarianism as it relates to key performance indicators (KPIs) and penal weight. I methodologically detail the implications and limitations of autoethnographic experience, reflecting on my approach to data collection and positionality. I present my findings in terms of journal excerpts maintained during my time in the field. Specifically, I share in the meanings for becoming acclimatised to a system where KPIs affect workload analysis related to the mass of penal weight. I conclude with a discussion of these novel findings, arguing for more autoethnographic accounts of probation in Canada.

Literature Review

Canada is a significant player in the development of risk assessment instruments globally (Luther and Mela, 2006; Ward and Maruna, 2007). However, the fettering of discretion for risk assessment decision-makers remains cautious, especially in how mechanisms of software and algorithmic logic impact accuracy and transparency (Pope and Bromwich, 2022). For example, understanding why evaluators tend to override risk ratings upward rather than downward (Orton et al., 2021) evinces concern about risk inflation and particularly how upward ratings exert a more positive effect on predictive validity (Guay and Parent, 2018). The proprietary nature of risk assessment tends to make details about its development opaque (Pope and Bromwich, 2022), and no known studies have explored the application of risk assessment software in terms of colour palettes, aesthetics and performance (Turner et al., 2022).

As Wener (2012) observes, carceral spaces and prison design are traditionally coloured by shades of grey, and in terms of environmental psychology, natural light is imperative. Consider the illumination of a computer monitor where the projection of red elicits a reflexive response such as to danger (p. 226). Colour has the capacity to thus confuse, confound, and conflate observations about key findings. This is integral to affecting cognitive performance where red is specifically tied to the fear of failure and achievement attainment (Elliot et al., 2007) because the colour red generally signals danger and hazard, which affects human behaviour (Ha et al., 2012). Conversely, the colour green is understood to create balance and harmony by reducing stress (O’Connor, 2011). The aesthetic design of software elements, such as a colour palette (ie red, green, yellow), is impactful by colouring both user experience and performance. In other words, red ex-marks and green checkmarks are key to priming risk assessors as binary ones and zeros correspond to a form of symbolic punishment/reward as every visual stimulus signals and communicate inherent design choices intended to optimise user engagement (Lanier, 2018).

LSI-R risk assessment software

Canadian researchers conducted a meta-analysis over three decades of findings about the Level of Service suite of instruments detailing limitations of validity and identifying a need for continued research (Olver et al., 2014). The most widely used and researched risk/needs assessment tool in the world is the Level of Service Inventory-Revised (LSI-R) (Multi-Health Systems, 2022a). Established by the ‘Risk-Need-Responsivity’ (RNR) model of rehabilitation developed by Bonta and Andrews (2017), the LSI-R tool is an instrument for predicting recidivism based on a social learning theory of criminal conduct. A cacophony of subsequent scholarship about the value of such risk assessment and alternative models has ensued (eg Andrews et al., 1990, 2011; Bonta et al., 2011; Bonta and Andrews, 2007; Cullen et al., 2018; Focquaert et al., 2021; Herzog-Evans, 2018; McNeill, 2019; Polaschek, 2012; Ward and Maruna, 2007). As the tool I was accustomed to using in my fieldwork was a software-based iteration of the LSI-R instrument, this program formed the basis of my work conducting criminogenic risk analysis.

In terms of penalty, understanding the vulnerabilities of risk is vital to knowledge about the carceral estate of probation (Ricciardelli et al., 2015), especially as the LSI-R is a trademarked product employed by most Canadian jurisdictions and several countries throughout the world (Casey et al., 2014; Pope and Bromwich, 2022). As Carlen (2011) described risk-crazed governance, a new mode of programming now pertains to numerical values and KPIs. LSI-R assessments are therefore undertaken by scoring questions as either a yes (1) or a no (0) and the sum score represents a total risk result (Casey et al., 2014). LSI-R software users are therefore primed by variables in the assessment of risk that designers select while probationers become categorised (Barnes and Hyatt, 2012), for example, by using colour-coded profiles of risk: green/low, yellow/medium, and red/high (Pope and Bromwich, 2022: 159).

A leading publisher of the LSI-R software program advertises some of its design features explicitly, such as ‘fully digital administration [for] automatic scoring … with the click of a button’ (Multi-Health Systems, 2022b, emphasis added). It is imperative to note that pressing a computer mouse reinforces familiarity through haptic feedback by engaging the sense of touch (Miyashita et al., 2007), as humans and machines together assign meanings (Enriquez et al., 2004). Masco (2014: 22) explains that operationally, a ‘color-coded system attempts to fine-tune institutions, officials, and citizens through an escalating scale of … danger’. Where green refers to low risk and red represents a high degree of risk, it is understood as a grammatical matrix of threatening potential. This means that people become acclimatised, where the significance of green does not represent the absence of risk but reinforces its presence. As risk is inherently uncertain, practitioners tasked with assessing criminogenic risk are thus primed by colourful symbols which are tied to experience vis-à-vis their emotional responses (Brown, 2021: xxv).

While Robert Witkin (2005) contends that aesthetics are perceived optically by seeing the world through vision and haptically by feeling it through the sense of touch, these sensations affect the body (as cited in De la Fuente, 2007: 415). LSI-R software is therefore programmed with multi-sensory in/outputs as criminogenic risk assessment is haptically reinforced through the sense of touch (ie computer mouse clicks) and optically through the sight of green checkmarks and red ex-marks (Moran, 2016: 122). Haptic communication thus occurs through an interactive interpretation between humans and machines experienced by soft subtle cues (Takagi et al., 2018). Where human perception is highly vulnerable to the social psychology of software design (Harris, 2016), such computer programming pertains to a penal communication theory in how meanings about punishment are inherently conveyed and understood (Duff, 2002, 2003). Therefore acclimatising to these atmospheric pressures evinces a need for ethical design choice-making and discussion about what is valuable for carceral spaces (Turner et al., 2022).

Vulnerability and totalising risk

As Desmet (2022) observes in totalitarianism, the Grand Narrative of risk occurs when ‘mass formation’ atomises people by governing them through fear and coercive control. In probation, risk assessment programming through LSI-R in/outputs becomes institutionalised when it forms an ideology about roles, rules, and norms (Zimbardo et al., 2021). Recursive to how the Stanford Prison Experiment was once a ‘total situation’, where good people dehumanised others by succumbing to the circumstances of their dirty work (Zimbardo, 1974), probation practitioners focused primarily on risk necessitates the mechanisms for obedience to authority (Milgram, 2017; Perry et al., 2020). This social psychology legitimises big-picture theorising of ‘penal weight’ as totalising in how it represents a psychological burden (Crewe, 2011; McNeill, 2018, 2019; Piché and Walby, 2018). Where penology was once focused on managing aggregates (Feeley and Simon, 1992; Lynch, 1998), probation is now primarily concerned with binary algorithms (Bucher, 2018; Cinnamon, 2017; Hannah-Moffat, 2019; Kehl et al., 2017; Završnik, 2021).

In practice, this has meant shifting focus from offender management through tough responses to ‘smarter’ governance (Goodman et al., 2017), where an aesthetic turn in management has generally created a culture that values measurable performance (Minahan and Cox, 2018). As algorithms operate through if/then logic, the qualitative nuances of practitioner discretion become diminished (Grant and McNeill, 2015; Robinson et al., 2014; Shapland et al., 2012). For example, deskilled practitioners then become decoupled from resources (Fitzgibbon, 2007) as managerialism and compliance myopia represent a new norm in the information age of probation (Phillips, 2016, 2017). This has negatively influenced practitioner morale (Farrow, 2004), inviting questions about the value of probation and its legitimacy (Canton and Dominey, 2018; RA Duff, 2003; Lovins et al., 2018; McNeill, 2011; Tidmarsh, 2022).

Risk as a metric of performance has therefore created an alternative criminogenic policy around data analytics (Hannah-Moffat, 2019), where the total focus has become KPIs (Hamai et al., 2005). Thus practitioners experience monotony in generic assessments of risk leading to ethical implications about mass supervision (Cracknell, 2022). Where ethics refer to doing things for the right reasons and potentiate coherence and integrity, what it means to ‘care’ in probation may be difficult to operationalise but vital to understanding the value of practitioners (Dominey and Canton, 2022). In Canada and transnationally, probation en masse forms a large part of the ‘penal dragnet’ that reshapes the sociosymbolic landscape (Benekos and Merlo, 2014; Wacquant, 2008, 2010, 2016). As technocracy and statecraft mix anew, this environment has created a ‘transcarceral habitus’ marked by its strict rules, regulations and policies, which devalue the human experiences of trust and evoke scepticism and vigilance (Quinn, 2021).

Therefore, ‘mass probation’ quintessentially demonstrates this variation in state punishment (Phelps, 2017) as it promotes division across moral lines of race, class, and gender (Phelps, 2018). That discrimination occurs in a paradoxical way which represents an alternative to imprisonment for some, while for others, incarceration becomes more likely (Phelps, 2013). Probation is thus fertile ground for potentiating an orientation to rehabilitation (Canton and Dominey, 2018; RA Duff, 2003; Durnescu, 2011; McNeill, 2011; Mcwilliams and Pease, 1990; Ward and Maruna, 2007); as well as the possibility for architectural reform in the context of penal aesthetics, choice-making, and design (Jewkes, 2017, 2018; McNeill, 2006; Phillips, 2014; Shah, 2020; Tidmarsh, 2021). While the value of a therapeutic alliance in probation practice means working with criminalised people (Ricciardelli, 2018), a paradigm governed by risk is convergent with the values of criminal desistance (McNeill, 2006; Ward and Maruna, 2007).

Emotional labour as an analytic resource

As Turner and colleagues (2022) discuss, attention to the aesthetic and architectural design of carceral spaces represent atmospheric pressures—in which competing emotions call for a ‘turn to affect’ (Pile, 2010; Thrift, 2004). Probation is likewise acclimatised to ‘sensory penalties’ that relate to human experiences of punishment and spaces of social control (Herrity et al., 2021). Thus, valuing the qualities of affect organisational outcomes aimed at efficacy (rather than efficiency) potentiates a focus on well-being through personal narratives (Durnescu and McNeill, 2014; Robinson et al., 2014; Shapland et al., 2012; Ward and Maruna, 2007). This is crucial for administrations to thus contend with ways in which probation practitioners manage their feelings and balance inward experiences with outward displays of emotion to manage workplace demands (Hochschild, 1983; Phillips et al., 2020). This concept of emotional labour pertains to how probation practitioners portray organisational values, Phillips and colleagues (2021) explain, which is a powerful analytic lens for shedding light on the embedded penal philosophies to create meaning about why and how institutions function. Moreover, by appreciating the affectual qualities of practitioners, organisations become enabled to prevent sickness, absenteeism, and turnover and learn about how to effectively support their labour forces through retention.

As Brown (2021) notes, pain is caused by the misuse of power as it relates to perfectionism. Where predicting the uncertainty of risk and attempting to manage it through KPIs is an impossible strategy it inevitably leads to a diminished sense of personal control for probation practitioners. In other words, the fear of penalty leads to a form of desperation in realising there are limits to pain both in how it is intended and the ways in which it is experienced (Christie, 1981). Thus, Appreciative Inquiry (AI) is a way of seeing and being in the world, Watkins and colleagues (2011) explain, where evaluation invokes negative feelings about the judgment that focuses on faults and evokes discord. The alternative approach to which I turn now is one of valuation that instead requires looking back to see what was most significant in an organisation (ie the project of risk assessment) with a view to increasing its merits in future.

Methodology

Following Worall and colleagues (2017), I sought to explore and describe normative dimensions and empirical realities about the meaning of probation by ‘opening a window’ and providing insights about LSI-R software programming. Inspired by ‘counter-visual ethnography’ (Schept, 2014), I aimed to perceive ideological facets of the cultural landscape in probation work by leveraging my own experience with probation as a carceral estate. In practice, this meant using autoethnography as qualitative inquiry and an ‘intellectual resource’—acknowledging myself as an active participant whose identity and biography influence and inform understandings of this research (Jewkes, 2012: 69). My data consists of subjective interpretations about my own experiences as a ‘complete participant’ in the field, employed as a fully qualified probation practitioner over a 6-year period (Hesse-Biber, 2016). As probation is inherently secretive (Mawby and Worrall, 2013; Shah, 2020), this required ‘studying up’ the enigmatic topic of community corrections by recognising the inverted balance of power of the organisation that employed me and my own personal experience as the subject of research (Beizsley, 2019).

Reflecting on personal thoughts, feelings, and behaviours, I followed an interpretative method of analysis to gain a ‘thick description’ (Geertz, 1973) by unearthing facets of detail about programming and criminogenic risk. These experiences about probation thus attend to perceived harm not only to others but also to myself (Hesse-Biber, 2016). As Jewkes (2012) notes about positionality, revealing experiences of justice and anxiety based on personal perspectives may enrich research findings. As the sole participant of this study, such phenomenology thus generates knowledge about what it means to ‘be in the crime’ (Massé, 2018) as I pursued criminological verstehen to denote a qualified position for sharing in the meanings of fellow probation practitioners (Ferrell, 1997: 10). My ‘professional curiosity’ is thus derived from my probation practice (Phillips et al., 2022), and in particular, the paradox that probation represents the most common form of supervision program in Canada (Malakieh, 2019, 2020), while experiences of probation practitioners, among all other criminal justice actors, remain the least visible, heard, or understood (Worall et al., 2017).

Field notes

The handwritten field notes I present in this article were made between 24 May 2020 and 22 June 2020 while I was situated within a Canadian metropolis. Data consisted of sensory observations and experiences about workplace events based on field notes containing job analysis, interpretation, and theories about personal matters where I reflexively memorised daily events (Hesse-Biber, 2016). My field notes amount to approximately 8500 words, were written when I had approximately 3.5 years of probation practitioner experience, and were often done during reflection at home and thus not ‘jotted down on the fly’ (Hesse-Biber, 2016: 184). This occurred in the evenings while decompressing from workday events, although sometimes I did write on weekends. Admittedly, there were days I did not make entries resulting from ‘secondary trauma’ when I experienced too much emotional, cognitive, and physical exhaustion to write (Severson and Pettus-Davis, 2013). Although, I found writing overall to be cathartic in purifying toxic experiences of danger and fear (Doerries, 2015: 36). With an emic perspective, I aimed to give voice while empathising with concerns experienced on the inside (Markee, 2012). All field notes were transcribed verbatim into digital text, helping to enhance reflexivity and notice themes, which turned into an integral part of my analytic process (Noaks and Wincup, 2004: 129). Excerpts of my field notes that form the basis of this article were formatted for grammar and minor syntax but otherwise read as originally intended.

Positionality

As Ricciardelli (2022) notes, sensations of vulnerability, confusion, and self-doubt were experienced as I reflected on my autoethnography. Albeit, I worked full-time in the field, allowing me to conceptualise data in pursuit of a PhD rather than vice versa (Brajuha and Hallowell, 1986). As a novice doctoral student, I can relate to the considerable anxiety of exiting the probation field as a practitioner while re-entering academia (Jewkes, 2012). As others have observed, resonating here is the politics and dichotomy of insider/outsider status (Jewkes, 2012), where my membership has been fluid in traversing roles (Ricciardelli, 2022). For example, I first entered the field of probation in an ‘active’ training role which transitioned to ‘complete membership’ as I gained tenure and experience (Adler and Adler, 1987). For a period, I went anthropologically ‘native’ by becoming totally involved, losing objectivity while immersed in field participation (Adler et al., 1986). Reflecting on the benefits of being there, I have earned a level of trust that comes with privileged access (Tope et al., 2005). My experience in practice has provided insights into the ethics and values of practice with which I hope to develop partnerships for penal institutions and academia throughout my ensuring career (Ricciardelli, 2022).

Moreover, I recognise that my perspective is inherently biased as it is imbued by liminality through my identity and social positions that span race, class, age, gender, and marital status (Jewkes, 2012). For example, I recall ways in which my dress at work changed from business formal to downright casual—not necessarily intentionally—but inevitably read in terms of status, power, and significance (Jewkes, 2012: 67). I was more apt to groom myself when entering performance reviews and to interact with official executives whereas in other occasions I recollect colleagues coaxing me to join them in sporting a beard—which eventually became rather straggly at times during my fieldwork while interacting with clientele, many of whom were homeless and themselves unkempt. Elucidating such distinctions in appearance, I subsequently learned this to be a tactic in presenting myself in carceral places while undertaking autoethnography (Jacobs, 2015; Jewkes, 2012). As probation values the inherent moral significance of race, class, and gender (Phelps, 2018), I can relate to my own identification as a White male with socioeconomic privilege (Jewkes, 2012; Phillips and Earle, 2010). Thus pertaining to the performances of masculinities and risk as a vulnerability (Ricciardelli et al., 2015), these qualities informed my analysis and were paramount to negotiating access while navigating the relation between public and private life in and outside of work.

Exiting the field

As Jewkes (2012) references, my professional work and personal life pertain to social performance and management of self (Ellis, 2020). Exiting the field was abrupt as I realised staying longer made it harder to pull back and maintain objectivity (Hesse-Biber, 2016: 198). Being admitted to a doctoral program thus facilitated my level of analysis in gaining scholarly support while creating space to critically interrogate my roles and responsibilities—to those I mentored, worked with, and supervised in the field. Yet most importantly, I realised that to exit the field was to prioritise my own well-being. I semi-consciously strategised a public departure by announcing my entry into a PhD program while maintaining reciprocity and rapport among friendships that may last a lifetime (p. 201). Having passed a point of saturation, where remaining in the field would no longer reveal any new insights (Saunders et al., 2018), I began navigating the stigma of leaving an identity behind and realised a toll that impression management had assumed (Goffman, 1963). From this arrived feelings of a ‘residual category’ (Garot, 2004; Jewkes, 2012), where I noticed, with help from my intimate partner, the need to prioritise my own well-being. This pertained to the stories I told to and about myself and the extent to which I immersed myself in probation work (Ellis, 2020; Jewkes, 2012; Steier, 1991). I noticed how this began to devolve into consequences that were ‘dire’ to my mental and physical health (Jewkes, 2012). Thus, engaging in therapy has required being vulnerable, that is, going to a place where things become overwhelmingly uncertain (Brown, 2021: 2).

Limitations

As Brown (2021: 190) states, ‘to love with any level of intensity and honesty is to become vulnerable’, and yet I realise that fear, stress, and anxiety are symptoms of my experience—that has led to my avoidance of publishing this research and also the excitement of a ‘sneaky thrill’ (Katz, 1988). Therefore, my affect is both a possible source of strength and limitation in which I am inherently biased and where generalisability is caution. Further, as my data is limited to a 4-week period, it may have been inadvertently impacted by affectual temporalities. For example, I was apt to write as a result of experiencing injustice where negative emotions tend to lead more quickly to engagement (Lanier, 2018: 18), thus elucidating a need for positive accounts with transcarceral habitus in future research. Also, for reasons of confidentiality, I limited some fieldnote details (Lancaster, 2017). Finally, despite my attempt at epistemological and theoretical rigour, it remains unclear if feelings of emotional investment produce more honest knowledge (Jewkes, 2012: 72).

Findings

Getting with the program

On 7 June 2020, I observed concern about ‘the wider project of exercising compliance’ as I felt my training and acclimation to the field had served as ‘indoctrination’ illustrated by the scoring of green checkmarks. I felt the LSI-R software application was problematic and recognised the rationality of assessments worked in tandem with legal enforcement. Demonstrating favourable performance was thus achieved in two main ways, by authorising LSI-R assessments and laying criminal charges where the probation practitioner:

has to prove themselves, and this is often accomplished by the evaluation of risk (i.e., being able to detect all factors) or through enforcement—in which, by an officer laying a heavy number of breaches, they indirectly communicate their value as an employee because of the work they produce (7 June 2020).

I felt the authorisation of LSI-R assessments was gratuitous because, in practice, ‘the risk assessment tool accounts for both the disposition as well as the details of the police occurrence report—the worst of each’ (21 June 2020). I recognised prejudice in valuing higher risk assessment scores. On 18 June 2020, I noted, ‘I spoke with my colleague about how to manipulate the RNR tool using professional discretion. … all checks applicable are inferred, and the worst-case justifiable is assumed’. By this, I meant value was demonstrated to my employer by showing higher scores of risk, which inadvertently led to a greater likelihood of legal enforcement. On 15 June 2020, I reflected on observations about what favourable performance meant and what was being seen in my work.

What I mean by optics is more than just public perception. It is a driver in shaping the culture of probation work because it refers to the practice of building a justifiable and rational defence for taking … enforcement action.

The implicit message was that compliance is non-negotiable, which I interpreted to mean even if adherence to the policy was improbable, the expectation was to conform. ‘If an assessment, for example, remains unauthorised past the 8-week allotted time frame because the nature of our inner urban city office is a site of gross poverty, inequity and mass desperation’ (11 June 2020), the noticeable expectation was to find a way to comply regardless and authorise the assessment. On 29 May 2020, I reflected on the rigid mechanics and technocratic application of risk that my administration favoured over nuance, especially as it related to the assessor’s point of view:

Life is grey—or colourful—and not black ink on white paper. This extensive “lack of discretion” is a hindrance to my professional capacity—especially when the expectations of the policy cannot be met, and my job becomes rationalising my actions … I am not a robot.

I noticed text-based data by way of case file information was privileged over the quality of human interactions. Standardised text as a unit of data entry conformed with the format of the LSI-R software application. That is, I conducted LSI-R assessments on the computer where data input consisted of mouse clicks and text word processing. These clinical judgements were supposed to be based on an amalgamation of documented case file details such as collateral contacts with community partners and interviews with the criminalised person under assessment. Often, however, the client themselves were unavailable to partake in their own assessment despite the 8-week deadline assigned to the assessor.

Individuals have been pressured by management to get the assessments done even if based on file information with a rationale and justification for that violation of policy—should the client fail to report over several weeks—leading to my need to violate policy and complete an assessment on an offender who I never spoke to, “get the assessment done anyway” we are told. That is the beginning of the slippery slope (18 June 2020).

I recognised that ‘once it becomes apparent that what the policy requires is an [impossibility], it loses its legitimacy’ (18 June 2020), creating a sense of discomfort that caused me to feel both manipulated and manipulative. I reflected on an informal discussion I had with my peer about deception. ‘My colleague and I after work shot the breeze about how we had mastered the program’, specifically:

We spoke about “scoring up”—which becomes the ad hoc and necessary evil in demonstrating my hard work [producing] numbers, which by all accounts are flawed as a result of what they’re aimed to capture, and which have been inflated either consciously or not in favour of the administration (18 June 2020).

Reflecting on this while journaling led to feelings of uncertainty regarding my own personal liability for the outcomes of this work. A sense of vulnerability and exposure led me to question if I had, in fact, mastered the program or had been mastered by it—effectively duped. This awareness of cognitive dissonance was a troubling state of tension between my beliefs and attitudes that were psychologically inconsistent with my behaviour and workplace expectations (Brown, 2021: 81).

Scrutinising efficiency measures

I recognised a flaw in the design about how performance was measured, noting ‘the basis of key performance indicators, that by all accounts from the frontline fail to include much of the meaningful work done by professionals’ was totally objectifying as ‘no effective mechanism appears to exist to allow for useful criticism and feedback from frontline staff’ (11 June 2020). I later noted ‘there appears little oversight about the quality of my work—since it’s all about the numbers—I figure [if] I keep within the “green” of my assessment currencies, and maintain up-to-date authorisation periods for these assessments, then there is little need to look [further]’ (22 June 2020). By ‘currencies’, I was referencing a metric rolled out with remote work to monitor monthly LSI-R authorisation reports during the pandemic. An LSI-R assessment was digitally authorised once the assessor completed the assessment in software. The currency of authorised assessments represented a monthly rate of productivity coloured in red if below the expected threshold. This concern for scrutiny was apparent elsewhere in my field notes:

I know that management is reviewing my numbers, and the work that I do is quantified, so I have some objective targets to meet. I am cognizant, in other words, of my proclivity to score up and do so quickly (26 May 2020).

Recognising that metrics were designed to forecast and predict organisational needs such as human resources, I took issue with the application of how the RNR model was operationalised in practice. This objectification framed a picture of my work in quantifiable terms. ‘These images are flawed because they are largely premised on artificial data—after all, how is the need assessed without input from the offender?’ (18 June 2020). Here, I noticed efficiency was privileged as a mechanism of computation where a zero-sum equation disfavoured the human client-centred relationship practitioners most valued. I described this calculable objectification where:

the work I am putting in has little quality to it—just a hollow shape of little substance so the veteran and senior officer most familiar with how the metrics work will self-modify to adjust and meet or exceed the standard. Thus, the sites we set in terms of how work is measured must be tested for precision. Why is no weighting of “productivity” informed by the client? By the public? By the probation officer? Are we all untrustworthy? Where is the meaning in these numbers? (18 June 2020).

Primarily, I was concerned with the efficacy of my time as a resource. On 2 June 2020, I elaborated further on this by describing a monthly committee meeting I was a part of in administering cases at my office. There, we discussed ‘the workload analysis tool—devised to distribute equal caseload weighting and help with the administration of case assignments in our office’:

My manager today announced that the region had taken a position from our struggling inner urban office because a fiscal decision was made higher up according to the numbers. “It's a game”, my manager said during our committee meeting—and emphasised the need for everyone to have assessments completed for each of their cases—the authorisation of which is how workload analysis points are assigned. … [the] ‘tool’ purports to determine a quantifiable allotment of points that roughly translates to personnel hours or FTE (2 June 2020).

Here, my concern about human resources is evident as I reference the unit of Full-Time Equivalent (FTE) as a quantitative metric my administration valued. My awareness of this operational facet may be indicative of the degree to which I aspired for a management position. I elaborated in detail on this concern regarding the mix of criminogenic assessment and the vulnerability of human behaviour:

Our policy, for example, requires at least one in-person interview [with the client] before an assessment may be authorised and uploaded to the mainframe database. Yet in our urban inner-city office where the clients are often transient and in crises they frequently do not report … what else is one to do? … with key performance indicators aimed at the quantity and severity of risk [it] becomes a systemically unworkable problem because the monthly reports reflect an average official caseload weighting premised upon an authorised assessment for each case within so many weeks from the time a case is assigned…. [When an assessment is not completed in time] the ‘analysis’ points awarded to the assessment period are lost and cease to exist. … So zero points for multiple cases are captured for each caseload … throughout the entire office thereby erroneously seeming like little work is being done based on the metrics the way they are designed. We are told that the higher-ups at the regional corporate office oversee the analysis of these points which translate to real world things like awarding or removing full-time compliment (FTE). We were just told the other week that a position for our … sinking office was … lost to another office due to workload pressures … it had double the average analysis points (18 June 2020).

The administration of cases impacted operations and vice versa, where on aggregate, the organisation valued assessors who worked efficiently by locating risk quickly and failing to do so meant more work. The penalty of not authorising assessments on time meant workload analysis points went uncaptured for the assessor’s caseload. The reward for adhering to the program meant the assessor benefited from negative reinforcement, where less work was assigned to that caseload. Therefore failing to prioritise LSI-R assessments meant caseload weighting points would not be awarded, which was compounded by each office at the regional level. In other words, this produced a scenario of falling behind where ‘catching-up becomes systematically impossible as more cases arrive as a result of the appearance that you’re not getting your work done because the assessments you don’t have time to complete [as they] remain unauthorised—and therefore no workload analysis points are assigned’ (2 June 2020).

Social psychology and ‘total’ control

As my observations occurred in the midst of a pandemic, I felt conflicted between the flexibility allowed by remote work and the potential for accelerated manipulation. On the one hand, I was pleased with the decision my administration took to allow for rotational days working from home. While on the other, I recognised concern for advancing the application of risk as a metric of performance. Reflecting on my productivity in this technological shift toward digital workflow, I considered efficiency in terms of efficacy.

Welcome, all of a sudden, to the digital age much thanks to the need for continued social distancing. I keep telling myself it is for the better and believe so, from an operational side, things will become much more ‘efficient’ yet with this level of modernisation will come a preoccupation with key performance indicators which, in my experience … are poorly orientated and weak. … [I] spoke candidly within our virtual staff meeting today—trying to explain to the junior staff less enthralled than I with how the governance of our operation actually works—that “it’s all about the numbers” I said, and gave a few examples to explain why the [workload analysis] tool is flawed; it doesn’t capture data outside of rather narrowly focussed protocols (11 June 2020).

I recognised an insidious concern for conflating insecurity with safety, writing ‘my specialisation [assessing criminogenic behaviour] requires work and if all the risk assessments that I conducted were evaluated as “low risk” I would be reducing the likelihood of my need for continued employment’ (26 May 2020). I realised my administration valued higher risk scores, but I felt uncritically adopting risk as a performance metric was myopic. Here, the totalising effect of caseload weighting points was apparent as I recognised the social psychology of applying risk technology. I described this ‘behaviour as a sort of currency’ in which experienced POs were motivated by the reward of workload analysis points.

It reminds me of Stanley Milgram’s research on obedience to authority while at the same time incorporating a digital version of the Stanford Prison Experiment … by Philip Zimbardo. Milgram because the sense of ‘weight’ afforded to caseload analysis points which are allotted in a manner that presumes to indicate how productive I am as a probation officer. ... On the other hand, Zimbardo has demonstrated how typically good people in wrong [carceral] environments can be easily coaxed to do extraordinarily bad things. I suggest a combination of these two occurrences are at play with digital software and its interaction with and by its users in probation work (26 May 2020).

I questioned the function of my role and the efficacy of my work while administering criminal justice, especially with my dependence on technology:

at work today our network was down. There was no Internet. Operations had nearly ceased until 1630h when the contracted communication company personnel arrived to flick a switch … We sat around in each other’s offices doing very little all day long (29 May 2020).

My concern was that my time as a human resource was effectively diminished by technocracy. I felt frustrated that I was prohibited from entering a particular room where I had building access to effectively troubleshoot a problem that involved pressing an electric breaker switch. This curious irony sat unwell with me as the task was apparently outside of my specialism despite feeling my role was effectively clicking buttons in risk assessment software. I later elaborated on this point regarding the nuance and challenging nature of my work given its situation in an inner urban setting:

my time as a probation officer is spent putting out fires dealing with crises involving homelessness, addiction, and polysubstance abuse, and concurrent mental health disorders—unlike in other cities … which had nearly double the average caseload weighting [points] … according to the numbers … all analysis points are being captured by diligent probation officers who know how the operation works [and] authorise all assessments and maintain them up-to-date (11 June 2020).

Here, I reflected on a conflation of risk in casework and the technocratic allocation of resources at the regional level. As I had experience working in different geographical settings, I realised the quality of social issues differed between rural and urban. Thus, the quick-paced environment of the metropolis where I was located was voluminous and measurement units of risk were inequitably distributed—meaning, an interdependence to play ‘catch up’ equated to the need to ‘score up’.

[T]he rate at which a probation officer would be required to complete assessments is too quick for much meaningful interaction or prosocial modelling of behaviour … Rather it [the administration] preoccupies itself with imprecise calculations as a forecast of how and where resources should be allocated in the future (2 June 2020).

This led me to realise my original interest in this study to understand how practitioners ‘explain and describe their engagement with electronic technology’ and how it impacts ‘probation habitus’ in the age of managerialism and calculability (24 May 2020). That is, I posed in writing to myself, ‘how is technology used to interact with individuals in the justice field?’ (19 June 2020) and recognised that management meant focusing on administering a cycle of enforceability through the assessment of risk.

So now the use of digital and electronic technology to speed up this … process runs the risk of even more quickly failing to meet the needs of humans in utterly helpless circumstances—that liminal space of being made the clientele of our products and services (11 June 2020).

Being an admirer of Nils Christie (1977, 2017), I recognised that an essential function of text word processing for data entry ‘exemplifies a technocratic approach to crime control as an industry based on conflicts as property’ in which my ‘specialisation’ and ‘expertise’ were insignificant (3 June 2020). The confluence of this led me to note the ‘service we provide therefore has less to do with rehabilitation and more to do with public confidence [where] the product is a sense of protection’ (15 June 2020).

Discussion and Conclusion

Probation software programming is a facet of contemporary punishment and social control imbued by the senses (Herrity et al., 2021). I have explored optics as visual stimuli that signal interpretative meanings, whereas haptics mediates symbolism through subtle cues between humans and machines in the design of LSI-R software applications. Probation is powerful both emotionally and sensorially in how it embodies non-verbal and algorithmic experiences through atmospheric pressures (Turner et al., 2022). I am cautious about this interplay, sensing it relates to an algorithmic logic of the RNR model where efficacy for desistance and rehabilitation becomes subsumed by a risk paradigm (Ward and Maruna, 2007). In short, the experience of LSI-R data in/output, where software has mediated reality in favour of risk through designed aesthetic objectives, was like playing a video game (Frome, 2007: 835). As penal governance preemptively organises itself around risk rationality to mitigate liability (Mythen, 2020), this paradoxically puts both people and organisations in a vulnerable state. That is, vulnerability is the first thing we look for while assessing others, yet the last thing we wish to show about ourselves (Brown, 2021: 83–84). As my autoethnographic experience and self-reflection have evinced, I am uneasy about ways in which the mediation of risk-based software applications coloured my personal experience as a probation practitioner for more than half a decade.

Where the RNR model of offender rehabilitation is ubiquitous in modern penal philosophy, it is the factors of risk which is foremost in principle—where criminogenic need and, to a much greater extent, responsivity become disparate to technocratic politics (Hannah-Moffat, 2005: 44, 2013; Werth, 2019). This is evinced to the extent that KPIs function to control and govern behavioural values and organisational performance (Hannah-Moffat, 2013; Maurutto and Hannah-Moffat, 2006). And demonstrable to the extent that I observed a norm about authorising LSI-R assessments based on mere static file information, which led to scoring up in aggregation. This morally injurious terrain requires an appreciation for the ethics and values of probation—and its ‘moral repair’, in which addressing such feelings of guilt, shame, and betrayal are necessary for restoring the fragmentation of self (Griffin et al., 2019: 358). My ethic for data entry based on case files inherently valued risk by privileging policies and procedures over people. Negated, then, are the dynamic factors which fail to become integrated into practice (Serin et al., 2019), as subjective experience and human agency are major considerations not only for ethics but the efficacy of risk assessment itself. If third and fourth-generation assessment tools are legitimised by the evaluation of dynamic risk (Hannah-Moffat, 2005, 2013; Orton et al., 2021; Werth, 2019), how can assessments—where input by criminalised people themselves remain absent—be considered valid or reliable? The desistance paradigm, by contrast, is inherently dynamic as it focuses on responsivity for individual needs, thus maintaining risk in check by balancing focus toward ethical design, valuation, and intentionality.

Ethical design and valuation

Returning to this article’s epigraph, I reflect on the mediation of clicking my computer mouse once to illuminate a green checkmark and twice to discover a red ex-mark. This system of respective signalling—a job that was well or poorly undertaken—has created a conflict of interests where assessing risk relates to a broader compliance project inherent to algorithms. That is, ‘engaging’ with the program mediated by software has altered my interactions via stimuli, tugging at emotions through the provision of symbolic penalties and rewards (Lanier, 2018). I experience guilt about this now, recognising that to score up has presupposed the zero-sum equation that rewarded me in terms of currency (eg LSI-R caseload weighting points) while simultaneously penalising the clientele I supervised. By valuing clients as higher risk, I contributed to probation’s mass and made the legal entanglement of my clients more probable. To the extent this activity has been compromising, it transgresses into a ‘transformative risk subject’ (Hannah-Moffat, 2005) as my self-modification evinces the mechanics of performance indicators leading to an ethic of productivity which valued expediency. Metrics that form the basis of risk-based governance thus elucidate why evaluators tend to override risk ratings upward rather than downward (Orton et al., 2021).

By reducing criminalised people into fragments and effectively atomising them into mouse clicks about case file text, I have engaged and experienced coercive control through the psychology of totalitarianism (Desmet, 2022). Evidenced by reflexive concern about my own obedience to authority (Milgram, 2017; Zimbardo, 1974), this has created anxiety that feels like an escalating scene of fear and loss of control (Brown, 2021: 8). And with uncertainty about the technocratic application of risk and its aesthetic software design, I have realised the need to prioritise well-being. Where LSI-R software attributes exemplify authority without control, the risk is inherently unpredictable and thus related to organisational well-being—especially given an alarming 59.2% of Canadian probation practitioners screen positive for a mental disorder (Ricciardelli et al., 2021: 16). My acclimatisation into this culture provides for an appreciative inquiry regarding the values and ethics of practice. Yet, given the closed and confidential nature of training and operational secrecy, the degree to which staff strive for a shared vision or experience confusion and role conflict is wanting further inquiry (Ricciardelli, 2022). As KPIs and metrics have coloured my experience about a data-driven concern for recidivism (Brown, 2020), I remain uncertain about how my own ethics and values, as a penal professional, have contributed to—or inhibited, the desistance paradigm that involves working with criminalised people (McNeill, 2006).

Affectual weight and penal aesthetics

Evoking my sense of penal weight is, therefore, literal, as in the burden of caseload points, and figurative, as it represents a mechanism of oppression (Crewe, 2011). This experience of dissonance elucidates what McNeill (2018, 2019) describes as ‘maloptic’ in the misrecognition of people who become seen in their poorest light through the misrepresentation of punishment itself. I argue this is also malhaptic in how technocracy and risk-based governance are applied to policies and practices that remain out of touch with human agency. Where risk-crazed governance promotes a program of uncertainty by attempting to control what is unknowable (Carlen, 2011), algorithms require cold, calculated ‘inhuman data’ (Bosworth et al., 2005: 259). In practice, this means subjectivities become truncated into worst-case scenarios based on pre-and post-criminal convictions. Hence, as Pope and Bromwich (2022) observe, the fragmentation of personal narratives into mechanically rigid modes of data threatens the nuanced discretion required of the common law tradition in Canada. Moreover, designers of actuarial tools and software applications must not only show how the weight of factors are determined from assessment to final risk result—but also the proprietary nature of software development and its aesthetic design. For example, why program green checkmarks in a single rather than double click? What decision-making occurs about symbolism and selection of colour palette?

Penal weight in terms of workload analysis points further elucidates how service providers experience human geography in terms of their performative value. Rethinking carceral space virtually, in terms of aesthetic software design—and how implicit choices may colour experience, calls attention to penal structures that foster human agency, nature, and well-being (Jewkes et al., 2020; Moran and Jewkes, 2014; Turner et al., 2020, 2022). A just design, therefore, focuses on hope and health in its governance and architectural orientation to criminality (Jewkes, 2018: 6). Yet my experience with KPIs represents a pain in which risk has been counterproductive to the goal of desistance—hindering probation’s rehabilitative potential (Durnescu, 2011). How, then, as McNeill (2006) posits, do we rehabilitate programming and social structures that enable conduits for human and social capital? Penal design requires both sense and sensitivity (Porporino, 2010) to understand how mediatised assessments serve the practitioner-client relationship. While a therapeutic alliance requires agency, the social structures behind homelessness, addiction, and mental illness represent building blocks from which first noncriminogenic and then criminogenic needs can be successfully addressed (Ricciardelli, 2018: 16).

Care is both an ethic for and value of probation, as described by practitioners in deciding to undertake the job in the first place by ameliorating criminalised peoples’ vulnerability (Dominey and Canton, 2022). Morale is thus about working with and for human beings, not machines, and thus representing a currency in itself for probation staff (Farrow, 2004). My experience framed by uncertainty for risk assessment has led to the emotional exposure of vulnerability (Brown, 2021: 13). This reflexivity and autoethnographic exposure has further revealed a sense of danger about technocracy where the complexity of human relationships becomes neutralised (Jewkes, 2012). Yet these subjectivities paradoxically have enabled a creative route for survival through this research by paying homage to my experience within the transcarceral habitus (Quinn and Tomczak, 2021). This leads me to wonder, if automation and digital scoring of criminogenic risk become the norm, what is my role—and whose interests will I serve in the administration of Canadian criminal justice?