Positive reinforcement in probation practice: The practice and dilemmas of praise

Introduction

Positive reinforcements, such as praising and complimenting, are proposed to contribute to clients’ change processes in counselling and therapeutic literature by, for example, creating a positive atmosphere, acknowledging the positive actions clients have taken, enhancing clients’ self-efficacy and easing clients’ fears about being judged or the change process itself (Truax, 1966; Wall et al., 1989). Studies in correctional practices (Andrews et al., 1990; Raynor & Vanstone, 2016; Trotter, 1996, 2012) and child protection (Trotter, 2004) state positive reinforcement is central to effective practice with clients involved on an involuntary basis, under a prosocial modelling approach. This approach is based on learning theory, with the underlying assumption that people are influenced by the behaviour modelled by others and by positive and negative reinforcement of their own behaviours. Trotter (2015: 123) outlines four steps of prosocial modelling and reinforcement: “identifying client pro-social comments and actions; rewarding those comments and actions; presenting oneself as a pro-social model; and challenging antisocial or pro-criminal comments and actions.” In this paper, we examine the second of these steps: how probation officers reward or reinforce clients’ prosocial comments and behaviours through positive evaluations or assessments, and how clients respond.

Positive reinforcement in correctional practice

Trotter (2015) highlights the most powerful way a social worker can reinforce prosocial behaviours and comments is with praise. He advocates that praise should be used liberally and purposefully, and explicitly directed at the client's prosocial comment or behaviour. In their study of probation officers’ practice skills, Vanstone and Raynor (2012) included positive affirmation as a key element of prosocial modelling, and, under their canon of verbal/non-verbal communication skills as part of necessary relationship building skills, they highlighted probation officers should be placing emphasis on positives and expressing encouragement. They found prosocial modelling was correlated with client desistance (i.e., lower likelihood of reconviction) one year post order completion, and verbal and non-verbal communication skills were correlated with client desistance two years post order completion (Raynor et al., 2014; Raynor & Vanstone, 2016). Praise may help desistance by ‘de-labelling’ the individual, in promoting the prosocial elements of their behaviour and reflecting this back, effectively celebrating and reinforcing positive change (Kirkwood, 2023; Maruna & LeBel, 2010).

Trotter (2015) and Vanstone and Raynor (2012) give some examples of what positive reinforcement or praise might look like in a probation context:

PO: It is good to see you. I appreciate your coming on time. It does make things easier for me. Did you come on public transport?

These examples identify positive reinforcement as praising or complimenting any identified prosocial behaviour from punctual attendance at probation appointments to more trait like prosocial behaviours, e.g., honesty and insight. However, they do not address how clients respond to such positive reinforcement or consider the other functions it might serve in practice.

Praise as interaction

Research examining interaction has highlighted that praising and complimenting can be problematic for the recipient of the praise to manage. In her influential research, Pomerantz (1978) identified that, in English, people in ordinary conversations tend to downplay or deflect praise, as they balance the competing interactional pressures to both accept a positive assessment and avoid self-praise. For example:

R: You’re a good rower, Honey.

This is notable because, generally speaking, in conversation we typically agree with evaluative statements, with the common response being to upgrade the initial assessment (i.e., by providing a stronger evaluation; Pomerantz, 1978).

However, Shaw and Kitzinger's (2012: 231) research on home birth support phone lines highlighted these pressures may be different in institutional contexts, as they found a high proportion of compliments (just over half) were accepted by the praise recipients usually by a ‘thank you’ or similar, and often accompanied with a breathy laugh or smile voice, which worked to manage the issue of self-praise by displaying ‘surprise, taken-abackness, or never-thought-of-myself-that-way-beforeness’. Similarly, when Gathman et al. (2008) examined how participants in a research survey responded to interviewers complimenting their performance, they noted there were distinct differences between the patterns of responding in everyday ordinary conversation and this institutional context. Here, respondents were less likely to downplay or deflect the compliment, but tended to deal with the praise minimally (e.g., ‘okay’, ‘yeah’), or withhold any response at all. This is very rare in ordinary conversation but was common in this context, leading Gathman and colleagues to surmise the participants are orienting to the interview as a particular type of institutional activity. As the survey participants did not have the knowledge in this context of what constitutes a good or bad performance, they were unable to downplay or deflect the compliment as one might in ordinary conversation.

Giving praise is also not always positive as it can be used to reproach, criticise or interrupt; for example, praising someone's usual positive characteristic to highlight how they are presently being hypocritical (Golato, 2005). In institutional interactions, it can be used as a tool to achieve practitioners’ agendas, both those which are explicit (e.g., promoting positive change) and those which are not. For example, Weiste et al. (2021) demonstrated that practitioners in a mental health rehabilitation group used compliments to exert pressure on clients to respond (e.g., ‘Leo is an artistic person […] do you have some opinion?’ [228]), to close down discussion topics and to generate exclusion and prepare clients for a negative decision. It may also be used by clients to achieve social actions; for example, Hudak et al. (2010) noted that patients in medical interactions used compliments to pursue certain treatment options (i.e., by complimenting a surgeon's characteristic that aligns with a particular treatment approach). As such, both giving positive feedback, in the form of praise or compliments, and receiving it, is not as straightforward as it might initially seem. Understanding how this works in interaction in probation sessions will deepen our understanding of the function and role of positive reinforcement in this context.

Methods and data

Using conversation analysis, we studied 21 video recorded probation interviews which took place between 2007 and 2010 as part of the Jersey Supervision Skills Study undertaken in the British Channel Islands. Full reports of the study are available in Raynor, Ugwudike and Vanstone (2014) and Raynor and Vanstone (2016). Briefly, the study collected 95 video recorded interviews from probation staff, which were used to analyse the skills used by officers in the interviews. The skills analysis was guided by a checklist and manual (Vanstone and Raynor 2012) drawn from international literature on effective correctional skills, and produced a numerical score for each interview based on the observed use of skills. Subsequent analysis of reconvictions showed that clients from higher scoring interviews were significantly less likely to be reconvicted than clients from lower scoring interviews, although both groups were initially assessed as showing similar risks of reconviction. Information was also provided about which skills were strongly associated with desistance over one year and two years. The Jersey study did not cover in detail the response of clients to the skills used or the process of interaction: the focus was more on ‘what works’ than on ‘how it works’. The current study is intended as a contribution to filling this gap.

The authors received ethical approval for secondary analysis from the University of Edinburgh and the Jersey Probation Service, and twenty recorded interviews were selected from the higher and lower ends of the skill score distribution for analysis in order to maximise the opportunity to observe differences, together with a further ten interviews selected to show three sequences in which officers recorded repeated interviews with the same clients. The study of praise and positive reinforcement which is reported in this article is based on 21 of these interviews, involving 8 probation officers and 17 clients. Twelve interviews are at the higher end of the skill distribution and nine at the lower end. These 21 sessions were selected due to practical reasons, i.e., quality of the audio and video, and content of the session, i.e., report interviews were excluded. To preserve anonymity and avoid the use of official case identifiers, the interviews are labelled here with a letter for the probation officer, number for the client and H or L to identify high and low, e.g., B6H. Where the session is from a sequence the video is noted in the brackets (e.g., (#2)). All names and identifying information have been anonymised.

Conversation analysis is an approach that involves the detailed examination of talk as interaction (Schegloff, 2007). It attends to the unfolding nature of conversation, such as patterns in the turn-by-turn sequential ordering of interactions, typically through analysing detailed transcripts of audio or video recordings of ‘naturally occurring’ exchanges, including fine-grain detail such as hesitations, pauses and overlapping speech. This type of analysis reveals how intersubjectivity is created and maintained through interaction, such as when people check mutual understanding, correct themselves or repair misunderstandings. As well as being applied to ‘everyday conversations’, conversation analysis has been used to study a wide range of institutional settings, and provides insight into how the business of such institutions is done, with the potential to enrich our understanding, reveal surprising findings, and guide practice. Given the central role of talk and interaction for probation practice, conversation analysis is a valuable approach for understanding how it operates (Kirkwood, 2016; Mullins et al., 2022), such as in this case of exploring how praise features and functions within probation sessions.

The video recorded probation sessions were transcribed by a professional transcription service. The first author viewed all the videos with the transcripts and identified instances of positive reinforcement, in the form of praise or complimenting. The second author viewed examples of the videos and transcripts and assisted with coding ambiguous instances. In line with previous research in probation (e.g., Trotter, 2015; Vanstone & Raynor, 2012), positive reinforcement here was considered as any positive evaluation or assessment of clients’ prosocial behaviours, thoughts, comments, circumstances or other aspect. In conversation analysis research, positive assessment in interaction has been commonly used to identify the actions of ‘doing praise’ or ‘doing complimenting’ (Gathman et al., 2008; Golato, 2005; Hudak et al., 2010; Pomerantz, 1978). However, as discussed above, positive assessments are not always doing these social actions, and praise or complimenting are not always in the form of positive assessments, and can instead appear in the form of, for example, questions or corrections (Golato, 2005; Shaw & Kitzinger, 2012). As Trotter (2015; Trotter & Ward, 2013) highlight, what is considered ‘prosocial’ is value laden, and runs the risk of the practitioner (or researcher) moralising or imposing their own values. For the purpose of this analysis, what was identified as ‘prosocial’ was informed by previous research on probation practices and encapsulated behaviours, expressions, comments and attitudes consistent with pursuing a non-offending lifestyle, such as gaining employment, addressing drug use, ‘staying out of trouble’, or pursuing positive interests (i.e., Andrews & Kiessling, 1979; Raynor et al., 2014; Trotter, 1993, 1996, 2015; Trotter & Evans, 2012).

Across the 21 sessions, 188 instances of positive reinforcement, as praise, were identified. These instances could be as short as just one utterance (e.g., ‘congratulations’) or expanded over several utterances. These were analysed to identify 1) what the action or function of the positive reinforcement was and 2) how clients responded to positive reinforcement. The first author also identified 28 instances where there was an interactional opportunity for praise or positive reinforcement, but the probation officer did not pursue this, or it was delayed. These instances were identified as times when the client made a prosocial comment or action (e.g., reported reduced drinking or engaging in a positive interest), but this was not clearly positively assessed by the probation officers. These instances were identified and analysed to consider how clients respond when praise is a relevant action in the interaction but not present. Given the difficulties in pinpointing something that is absent, identifying these missed opportunities for praise is unlikely to be exhaustive but gives us some indication of the potential role of praise.

Findings

Discussion

Although praise is seen as an important part of effective practice, ironically it can create problems in the interaction. That is, for practitioners the challenge is how to praise clients without creating pressure for them to reject that praise; for clients, a normative response of deflecting or down-playing the praise could be seen as not complying with the institutional agenda. So examining how praise is enacted, how it functions, and how clients respond to this, is important for understanding a key part of probation practice.

As shown, practitioners could be explicit about what they were praising and why it was good. This would help identify and reinforce pro-social thoughts and behaviours, with the potential to build and sustain motivation for engaging with interventions and the process of desistance. In the moment-to-moment unfolding of interactions, praise could also help to clarify that their engagement with the tasks was going well and to encourage them to stick with it. In this way, it helps build social solidarity between clients and practitioners, as well as affirm that positive change is possible, key elements of desistance where change is acknowledged in the relationship and reflected back to the clients (Maruna & LeBel, 2010). However, where praise required a response from the client, explicit and continuous praising may be difficult for clients to directly accept and could create problems in the flow of the interaction.

Although clients tended not to respond explicitly to praise, this was often related to the way the praise was given, such as in a position that did not require a response from the client, in a conversational sense. This helped avoid the interactional trouble that may otherwise result, and allowed the task at hand to continue. Minimal responses (e.g., ‘yeah’) were a simple way for clients to accept the praise, giving due respect for the practitioner's authority to provide positive assessments, while avoiding potentially problematic responses of either rejecting the praise or seeming to self-praise. Orienting to a different action – such as commenting on the benefits of the probation intervention – was another strategy that dealt with the interactional constraints, where the client could implicitly accept the praise and align themselves with the institutional agenda.

Our coding and analysis of the data also showed that there were some instances where POs’ provision of praise was relevant, but was not given. Indeed, as shown in extract 10, clients could persist in describing their actions in ways that functioned to elicit praise, indicating that they treated probation sessions as an appropriate forum for discussing pro-social behaviour, and for POs to confirm and reinforce that their behaviour was correct and valuable. There may be many reasons POs do not praise clients’ reported behaviours or beliefs, for example concerns about risk or to direct discussions a different way. However, in these instances rather than an absence of praise there is a presence of another action, e.g., challenge or redirection. Although we can only speculate here on the implications of POs not ‘biting’ when their clients are fishing for compliments on their pro-social behaviour, we suggest that interactionally these minimal responses may jeopardise social solidarity and therefore engagement, as well as potentially convey that such behaviours are not desirable or worthwhile. This may point to why positive reinforcement, as praise, is an important aspect of probation.

Although we noted that missed opportunities for praise were more frequent in the sessions coded as demonstrating low skill, these are necessarily tentative comparisons as there are several factors which impact making such comparisons. Firstly, our data set was limited with only four POs in each category. Secondly, the length and content of the sessions was quite different between the low and high categories; the low skill category sessions were shorter and less likely to involve structured programme work and more likely to be report interviews or briefer check-ins. This may have limited the opportunities for clients to present prosocial comments and actions, and for POs to offer positive assessments. Thirdly, it is important to consider the clients’ stance and contribution in these sessions as influential to how the interactions have unfolded, that is whether they have provided accounts or examples of prosocial behaviours and attitudes to be assessed, and the capacity of the PO to provide positive assessment for prosocial comments and actions. Relatedly, whether the ethos of the organisation itself is prosocial or the individual practitioners are prosocial in their wider actions and values (e.g., being punctual, maintaining confidentiality etc.) will impact the interactions in situ that we have examined here.

As we have mentioned, defining praise can be difficult; we have treated it as one form of positive reinforcement, but it overlaps with related concepts such as compliments, not all of which might constitute praise. Moreover, as noted in previous research (e.g., Golato, 2005; Hudak et al., 2010; Weiste et al., 2021), not all apparent praise may function as forms of positive reinforcement, but could perform other social actions. The definition of ‘pro-social’ is also debatable and context-specific, which is important for considering how these particular interactions unfold. The data relate to a specific time, place and cultural context, and further research would be needed to explore how praise and other related forms of positive reinforcement manifest and function in criminal justice contexts.

Overall, our research demonstrates that praise is part of the business of probation practice, and its absence is noticed. However, praise raises interactional challenges that practitioners can manage skilfully to positively reinforce prosocial comments and behaviours and promote engagement. It highlights the ongoing relevance of positive reinforcement in practice, and the important role this has in maintaining ongoing engagement with clients. We hope this study helps practitioners understand the dilemmas clients have in responding to praise whilst also understanding that praise is a necessary part of practice. This study enhances our theoretical understanding of that nature and function of positive reinforcement, and provides concrete examples that can guide practice by giving insight into how and where practitioners can use praise effectively in their interactions with clients.