Core Correctional Practices

Background

In 1980, Andrews and Kiessling introduced five Core Correctional Practices (CCP), i.e. core activities for professionals working with clients to reduce delinquent behaviour. Dowden and Andrews (2004) describe them as follows:

- Effective use of authority: firm but fair, clarity about the rules but also supporting clients to comply with these rules in a friendly, non-dominant way;

- Prosocial modelling and reinforcing: demonstrating desirable behaviour and rewarding clients if they display such behaviour, disapproving of undesirable behaviour and helping clients to convert this into desired behaviour;

- Problem-solving: teaching clients to solve their problems in a prosocial way;

- Use of sources in society: engaging necessary assistance from agencies (e.g. debt assistance ) with or on behalf of the client;

- Quality of interpersonal relationships between the probation officer and client: an open, warm, stimulating and solution-oriented style of communication.

Cognitive techniques and motivational interviewing were added to this in subsequent studies (Bonta et al., 2011; Pearson et al., 2011). The CCPs are mainly based on the social learning theory.

The empirical substantiation for the combined use of the CCP is described below. These are discussed per practice in Chapter 5.

Empirical substantiation

A study from Denmark concluded that the probation officer partly determines the effectiveness of the supervision. This study analysed the data of 19,534 clients and 371 probation officers. Clients were randomly assigned to probation officers in one part of the study. Small but significant differences were found in recidivism (one year after completion of supervision) among clients of different probation officers, on the basis of which the researchers concluded that it matters which probation officer is responsible for the supervision (Andersen & Wildeman, 2015). This study does not provide insight into the question of what determines that difference. Studies on the use of RNR and CCP provide clues for this.

A meta-analysis of 273 studies examined whether CCP is mentioned in different programme descriptions and, if so, whether these programmes lead to a reduction in recidivism. Four out of the five practices appear to show strong associations with reduced recidivism. The mere use of sources in society was in itself not associated with less recidivism. The use of CCP appears to be particularly effective in combination with the RNR principles (Dowden & Andrews, 2004).

25

Various studies describe initiatives in the United Kingdom, the US and Canada where probation supervision is implemented through a combination of the RNR model with CCP. Slightly different definitions of the CCP are used in the various studies but they generally correspond. In many of these projects, probation officers complete a training course of 3 or 3.5 days focusing on the application of CCP, complemented by follow-up sessions. It is also encouraged to organise peer supervision meetings about the new practice, with probation officers receiving individual feedback from researchers on the basis of audio recordings of conversations with clients. Chadwick et al. (2015) conducted a meta-analysis that involved 10 studies on such an approach.²¹ They found a small but significant effect of this practice on recidivism by clients (effect size expressed in Cohen’s d average of 0.22).²² Expressed in recidivism percentages: an average of 36% of clients of trained probation officers reoffend, rising to 50% in the event of untrained probation officers.

In this research, we found 10 studies, 6 of which were included in the meta-analysis of Chadwick et al. (2010). Since a number of more recent studies provide relevant additional results, we summarise these briefly below. The table below provides an overview of these projects and studies.

Table 1: overview of studies into the effectiveness of using RNR and CCP Project Country Study Scope Core correctional

practices

21 Six studies from the meta-analysis by Chadwick et al. are also described in this section. They are: Bonta et al., 2011; Pearson et al., 2011; Latessa et al., 2013; Raynor et al., 2014; Taxman, 2008; Lowenkamp et al., 2014.

22 Cohen’s d can take any value but usually fluctuates between -2.0 and +2.0; an effect size of 0.22 represents a small effect.

23 More probation officers had been trained, but 28 did not submit a recording and were not part of the study.

24 Initially, 88 probation officers were recruited for the study, 21 of whom withdrew and 8 probation officers dropped out later on.

26

Several studies found that, after the training, the probation officers pay more attention to dynamic criminogenic needs, use CCP more often and also use a broader range of CCP in the supervision (Bonta et al., 2011; Raynor, Ugwudike, & Vanstone, 2014; Robinson et al., 2012). The findings from these studies show that the use of a combination of RNR and CCP in probation supervision has a modest but positive effect on reducing recidivism, both in the US and Canada, as well as in the United Kingdom. Although the effects are not significant, they all point in the same direction. It is however striking that the first studies achieved better results. When STICS was repeated in another province, no significant differences were found between the experimental group and control group for the approach as a whole (except for a part thereof) (Bonta, Rugge, Bourgon, & Wanamaker, 2019).

Contrary to previous results, repeated effect measurements in the STARR (US) and Citizenship (United Kingdom) projects no longer yielded any significant differences either (Lowenkamp et al., 2014; Pearson et al., 2016). As a possible explanation for these less positive findings, the authors of all three studies indicate that there were implementation issues: no use of additional counselling or post-training feedback to ensure a thorough embedding of the practice (Bonta et al., 2019), or no additional programming after the training, as a result of which it is unclear to what extent probation officers actually put the newly learned skills into practice (Lowenkamp et al., 2014). In addition, the following explanation is given: low numbers, making it more difficult to find statistically significant results (Bonta et al., 2019) and a stronger research design in the repeated study (RCT) which is known to often produce less positive results (Pearson et al., 2016).

A different approach was applied in a small study by Raynor et al. (2014). The CCP is subdivided into relational skills (shaping conversation, verbal and non-verbal communication, effective use of authority) and structuring skills (motivational interviewing, prosocial modelling, stimulating problem-solving ability, cognitive restructuring). Based on an analysis of interview recordings of ten different

27

probation officers with 75 different clients, the officers are subdivided into two groups: high versus low scoring in the use of skills. Raynor et al. found that high-scoring probation officers use both relational skills and structuring skills, whereas low-scoring probation officers mainly use relational skills. High-scoring probation officers also use both skills more consistently across the various interviews. Clients who are guided by high-scoring probation officers reoffend less than clients who are guided by low-scoring probation officers (26% versus 58%). The authors conclude that the combination of relational and structuring skills is particularly important: relational skills probably form a basis for being able to apply structuring skills (Raynor et al., 2014).