1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References


Offence paralleling behaviour (OPB) is a relatively new concept that emphasizes the assessment and modification of behavioural patterns that parallel criminal offending. Extant empirical research into OPB has typically focussed on the similarity between aggressive behaviours observed in custody and violent index offences perpetrated in the community; the results of these studies have been inconsistent, with the degree of similarity varying within subjects and across studies; (Daffern, Howells, Stacey, Hogue, & Mooney, 2008; Daffern, Howells, Mannion, & Tonkin, 2009). The aim of this study was to establish the level of similarity between OPB predictions and actual in-patient aggressive behaviours.


This prospective pilot study used a novel practice algorithm (Jones, 2010a) to develop a reference formulation from which OPBs and Pro-social alternate behaviours (PABs), the pro-social variants of OPBs, were predicted. Participants were five mentally disordered offenders resident in a UK medium secure psychiatric unit. Following generation of a reference formulation and OPB and PAB predictions the participants were monitored for 6 months. The degree of similarity between predicted and actual OPBs and PABs was calculated using Jaccard's coefficient.


Results indicated considerable similarity between matched (pairing predictions and their corresponding actual behaviours) OPBs, which were also more similar than random pairs (pairing randomly selected predictions and aggressive behaviours).


This study revealed the existence of OPB in mentally disordered offenders and has provided the first test of a novel practice algorithm, revealing its potential to guide OPB formulations.


  1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References

Violence is a ‘leading worldwide public health problem’ (World Health Organisation, 2002; p. 1); the total annual cost of homicide and wounding in England and Wales has been estimated at £13billion (National Audit Office, 2008). In Brazil, the financial burden associated with violence accounted for 1.9% of GDP in 1996 and 1997 (World Health Organisation, 2002). A significant proportion of the costs associated with violence are incurred as a result of repeat offending, with evidence suggesting ‘less than 5% of the male population commits between 50% and 70% of all violent crimes’ (Hodgins, 2007, p. 12). The financial, social and psychological costs of violence and aggression have motivated considerable research and a vast amount of psychological research has been directed towards understanding aggression and developing methods for assessing, managing and treating violent offenders (Howells, Daffern, & Day, 2008). A proliferation of structured risk assessment methods and treatment programmes has emerged, though more recently the value of case formulation driven assessment and treatment has been recognized (Daffern & Howells, 2009; Daffern, Jones, & Shine, 2010; Hart, Sturmey, Logan, & McMurran, 2011). The offence paralleling behaviour (OPB) framework is a case formulation driven approach to offender assessment that has emerged as an adjunct to structured assessment and treatment methods. Originally coined by Jones (2004), Daffern, Jones, et al. (2007) have recently defined OPB as: ‘a behavioural sequence incorporating overt behaviours (that may be muted by environmental factors), appraisals, expectations, beliefs, affects, goals and behavioural scripts, all of which may be influenced by the patient's mental disorder, that is functionally similar to behavioural sequences involved in previous criminal acts’. (p. 267)

Although it is a novel term, OPB has its foundation in existing theoretical and conceptual frameworks (for review see Jones, 2010b); the notion of monitoring and modifying manifestations of persistent maladaptive behaviours as they arise in therapy resembles strategies used in functional analytic psychotherapy (Kohlenberg & Tsai, 1994) and schema therapy (Young, Klosko, & Weishaar, 2003). Within the forensic field, empirical evidence from research on sexual crimes (Grubin, Kelly, & Brunsdon, 2001) and burglary (Bennell & Canter, 2002) has consistently supported the view that if similar psychological components are activated then consistency in behaviour may occur across different situations and crimes (McDougall, Clark, & Fisher, 1994; Shoda, 1999), an assumption that provides the theoretical foundation for case- linkage analysis (Woodhams, Hollin, & Bull, 2007).

The OPB framework can be utilized by a range of clinical and custodial staff (Daffern et al., 2010) and has been seen as particularly promising within group-based in-custody treatment settings (e.g., custodial therapeutic communities), where behaviours emerging within custody but external to formal treatment sessions have historically been processed within psychotherapeutic groups. Engaging the multidisciplinary team to intervene as an offender enters a behavioural sequence that parallels their offending maximizes the potential for change. Additionally, OPBs can aid risk assessment as an adjunct to the nomothetic approach emphasized through many structured risk assessment instruments (Jones, 2010b). OPBs can also be used to guide intervention by identifying problem areas (i.e., those which maintain risk) and selecting the most appropriate type of intervention. This supports a person-centred approach to therapy and intervention. In addition, OPBs can be used in the monitoring and evaluation of treatment progress (Jones, 2010b). It is important to note here that despite considerable promise, the misapplication of the OPBs concept could be associated with detrimental consequences for clients. In the context of OPBs serving as an aid to the assessment of risk, misapplying the OPBs framework and interpreting non-relevant behaviours as OPBs could result in unduly restrictive risk management or intervention (Daffern et al., 2010).

OPB is a relatively new concept that remains in its developmental stage. There have however been several studies which have tested the OPBs concept in clinical settings (Daffern, Ferguson, Ogloff, Thomson, & Howells, 2007; Daffern, Howells, Stacey, Hogue, & Mooney, 2008; Daffern, Howells, Mannion, & Tonkin, 2009). This research has established that there is often some similarity between in-patient aggressive behaviours and patients' index offences (Daffern et al., 2009). However, many in-patient aggressive behaviours also lack similarity with index offences (Daffern et al., 2009) and some research has shown that sexually aggressive behaviour is not indicative of sexually abusive behaviour in custody (Daffern et al., 2008). Daffern, Ferguson, et al. (2007) did not find a relationship between in-patient aggressive behaviour and pre- and post-admission aggressive behaviour. However, they did find a relationship between pre- and post-admission aggression suggesting maintenance of a tendency towards aggression during hospitalization. It is possible that these tendencies were expressed in a topographically dissimilar manner, or that the prototypical precipitants to aggressive behaviour for this group were absent or that aggressive behaviour or its proxy (Gordon & Wong, 2010) were undetected during hospitalization (see Jones, 2010a). The difficulties in establishing reliable and valid OPBs formulations are consistent with the clinical case formulation literature (Hart et al., 2011), which shows limited reliability, validity and predictive power.

Concurrently, Jones (2010a) has developed a practice algorithm to assist staff with generating OPBs formulations (see Figure 1), which can then be used to guide treatment and assessment using the OPBs framework. According to Jones (2010a) the practice algorithm begins with the development of a reference formulation, which is based on a review of literature relevant to the offence and the individual case and ascertainment as to whether the information relevant to generic models and theory match the ‘case’ (see Jones, 2010a, pp. 71–88). The formulation can either be a ‘specific functional analysis/formulation of different individual offences’ or the ‘analysis of themes and common processes underpinning different offences’ (p. 72) peculiar to an offender. Once developed, this reference formulation is used to generate predictions of OPBs and Pro-social alternate behaviours (PABs; behavioural sequences manifesting in pro-social behaviour that otherwise resemble the OPB behavioural sequence) and to test whether these predictions come about during some follow-up period. Once the predictions are validated interventions are introduced to assist the client meet their needs in pro-social ways.


Figure 1. Practice algorithm for identifying offence paralleling behaviour. Reprinted from Jones, L. (2010a). Approaches to developing OPB formulations. In M. Daffern, L. Jones, & J. Shine (Eds.), Offence paralleling behaviour: A case formulation approach to offender assessment and intervention (pp. 71–88). UK: Wiley-Blackwell, with permission from John Wiley and Sons.

Download figure to PowerPoint

Aims and Hypothesis

This study aimed to prospectively examine a novel method for developing OPBs formulations, based on Jones (2010a) practice algorithm, by examining the level of similarity between OPBs predictions derived from formulations based on previous behavioural repertoires, and actual in-patient aggressive behaviours in a Medium Secure Psychiatric Unit. PABs were conceptualized based upon anticipated pro-social behaviours that might be expected to satisfy the functional needs of the predicted OPBs. It was hypothesized that there would be similarity between predicted OPBs, derived from formulations based on previous offending and other aggressive acts, and actual observed aggressive behaviours. This would suggest that similar psychological components were active across previous and in-patient aggressive behaviours and that the formulations encouraged by Jones (2010a) algorithm are capable to identifying these OPBs and PABs.


  1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References


The study was conducted across four male wards (58 beds) of a Medium Secure Unit (MSU) in Bethlem Royal Hospital, UK. The Unit specialises in the management and treatment of offenders whose offending is related to a mental disorder.

Study population

All male patients between the ages of 18 and 64 years who could communicate effectively in English, who had committed a violent index offence, including rape of an adult, and had at least one aggressive incident during their hospital stay were considered eligible for this study. An aggressive incident was defined as any behaviour which can be defined as such according to the Overt Aggression Scale (Silver & Yudofsky, 1987) except for aggression towards the self. Several exclusion criteria were also applied. Patients with learning disabilities and neurological problems impairing capacity for consent were excluded. Patients in preparation for discharged were also excluded, as follow-up lasting 6 months would have not been possible.

Using the aforementioned criteria, 16 (27.6%) patients were eligible to participate, of which five consented. Information on age, diagnosis and index offence was gathered, in order to better understand the composition of the participants who decided to take part. This information was also helpful in assessing how representative of the MSU population, the recruited sample is, thus informing the generalizability of the study. Participants' age ranged from 21 to 49 years (= 34, SD = 13.47). The primary DSM-IV diagnosis, derived from review of patients' records, was: Antisocial Personality Disorder (N = 1), Borderline Personality Disorder (N = 1), Schizoaffective Disorder (N = 1) and Schizophrenia (N = 2). Regarding participant's index offence, two had been convicted of Rape and three of Grievous Bodily Harm (GBH).


All stages of this protocol were executed by an Honorary Researcher, with Bachelors degree in Psychology and a Masters degree in Forensic Mental Health Science. Prior to data collection, this researcher had also worked for 2 years as an Assistant Psychologist conducting Cognitive Behavioural Therapy therapy. This was an advantage for several reasons- the researcher had an experience of working in in-patient settings and applying theoretical frameworks to develop understanding of difficulties. Furthermore, through clinical work, the researcher was experienced in conducting formulations using similar components to those used in this study (e.g., behaviour, thoughts, and emotions) to form a better understanding of behaviours and their functions. During the course of the study regular supervision was provided by a Consultant Forensic Psychiatrist and a Clinical Psychologist.


All patients who satisfied the inclusion criteria were approached in person by the researcher. They were given an information sheet, the details of which were also read aloud, in order to ensure understanding and to give an opportunity to ask questions. They were offered a week to consider participating. The information sheet contained information about the study (i.e., procedures and expected duration); it highlighted that participation was voluntary and that declining to take part would not impact on any aspect of care; participants would not receive any payments, benefits or incentives; full contact details of the researcher and advice on how to withdraw from the study were also provided. Additionally, patients were given details of the hospital's Patient Advice and Liaison Service, in case they wanted to discuss the study with an independent person. Following a decision, patients who were agreeable to participate were given a consent form. The content of this consent form was read aloud and patients were given the opportunity to ask questions. Following an agreement to participate both the researcher and the patient signed the consent form. A copy was retained by both the participant and the researcher. Participants were advised that following the completion of the study, they would receive a newsletter with the findings. They were encouraged to contact the main researcher with any questions regarding the newsletter.

Ethical considerations and approval

The study raised several ethical considerations and these were considered carefully prior to commencing the study. In the first instance, recruiting participants with mental health problems required careful consideration of capacity to consent. This was judged by clinical staff on the wards. It was decided that should a participant no longer have the capacity to participate, they would be withdrawn from the study; however any information obtained to that date would still be retained. This was also made explicit on the information sheet and consent form at the beginning of the study.

A further ethical issue was related to predicting potential aggressive behaviours on the wards and whether these predictions should be shared with clinical staff. A decision not to share predictions with clinical staff was made on the basis of several factors. Disclosing predictions to staff might have impacted on staff observations by making them more vigilant to participants' behaviours and perhaps have led to the introduction of restrictive interventions that might have escalated the risk of aggression. This in turn could have had a negative impact on patient-staff relationship. In addition, all predictions were based on clinical notes; therefore the likelihood of the researchers detecting significant risk-related information that had been previously undetected was small. Finally, the accuracy of predictions made on the basis of OPB formulations is still unclear. Nevertheless, the researcher discussed the assessments with the main supervisor to determine whether any risk-related disclosures were required.

Research Ethics Committee and Research and Development approvals were obtained (Ref: 10/H0807/85) from The South East London Research Ethics Committee prior to commencing the study.

Operational procedures

Information regarding age, mental health diagnosis, index offence, previous and follow-up OPB and PAB was obtained from the patients electronic medical records (ePJS). An electronic incidents database (Datix) was also used to collect information regarding in-patient aggressive behaviours. Incidents and reports on Datix which were incomplete, unclear or reported on behalf of a person who was hurt rather than the one who displayed the aggressive behaviour were not taken into consideration.

Following patients' consent to take part in the study, Jones (2010a) algorithm for formulating OPBs was followed by the first author to develop OPB predictions:

  1. Literature and observation based case formulation of index offence and other relevant behaviours. Individual formulations were created for each participant using current relevant causal models and theoretical frameworks. For example, for sexual offences the Integrated Theory of Sexual Offending (ITSO; Ward & Beech, 2006) was used where clinical data was organized into the pre-defined categories such as clinical symptoms/state factors; distal and proximal factors.
  2. Thematic analysis. Common themes across each individual's offending and other relevant behaviours were identified and analysed through multiple sequential functional analyses (Gresswell & Hollin, 1992). The ABC model (identifying Antecedents; Behaviours (overt/covert); Consequences; Hollin, 1990; Sturmey, 2007) was used as a guide to pull apart relevant behaviours in order to identify common themes. For example, a recurrent theme of overt behaviour, derived from such analysis could be discovering that the individual withdraws from all social interaction prior to aggressive behaviours. During the above process, behaviours were parsed using Clark and Crossland's (1985) principal: ‘judging where the greatest discontinuities of form or function lie within the behavioural stream’ (p. 103). A reference formulation was then developed, which was tested on behaviours to date, ensuring that it accounted adequately for these. If it did, the researcher would proceed to making OPBs hypothesis. If it did not, the formulation was revised. The above process was designed to increase the validity of the formulation.
  3. Hypothesis about OPBs and PABs predictions. Two hypothesis were generated, one regarding OPBs and one of PABs. Prior to generating predictions about future behaviours, nine causal features and characteristics for each OPB and PAB and actual behaviours were identified to enable the researcher to examine the similarity between the predicted and actual behaviours. These features were: stressors, context, cognition, emotion, mental state, physiological state, behaviour, drug/alcohol and function. The above characteristics were compiled through identifying relevant features during the previous two steps of the practice algorithm and through existing research on OPBs. For example Daffern et al. (2009) classified patients' aggression using similar categories to those outlined above, to compare similarity between patient's index offence and in-patient aggressive behaviours. However, this study used different methods to categorize information (e.g., the Behaviour Cycle Logs, a structured document designed to encourage patients to reflect on their behaviours on the wards and to gather more information about the behaviour). In addition, guidelines proposed by Jones (2010a) were also followed, such as identifying: stressors and/or triggers; social context variables; self-regulation changes in the lead up to the offence and Detection Evasion Skills. Following this, two predictions were generated, one about aggressive behaviours (the predicted OPBs) and another about positive behaviours (the predicted PABs).
  4. Outcome of predictions. Participants were followed up for 6 months, from January until June 2011, in order to establish whether predicted behaviours (OPB and PAB) manifest. This was assessed through regularly reviewing clinical notes and the untoward incidents database (Datix). In cases where relevant behaviours were identified, the first researcher met with the participant in order to complete a Behaviour Cycle Log for the relevant behaviour. Following the above process, follow-up behaviours were then coded on the nine causal features and characteristic (see Step 3). The features were coded either as present or absent.
  5. Refutation and Revision. Jones (2010a) algorithm encourages subsequent refutation or revision of the OPB formulation following a period of observation. However, for the purpose of this study the test of similarity, as described in the statistical analysis plan section, was used to determine whether the identified behaviours were similar to the predicted OPBs. The similarity found by this study was also compared to similarity which could be expect by chance, by randomly pairing actual behaviours and OPB predictions from different participants and comparing those to the matched pairs.

Statistical analysis plan

The degree of similarity between predicted and actual behaviours was calculated using Jaccard's coefficient. The Jaccard's coefficient allows for the ‘measurement of similarity between two events or pairs; its value ranges from 0 to 1, with 0 indicating no similarity and 1 indicating a perfect match’ (Woodhams et al., 2007, p. 242). Previous research on OPBs (Daffern et al., 2009) has used this method to measure similarity.

In order to determine whether the degree of similarity calculated by the Jaccard's coefficient is different from that which could be found by chance, the predicted and actual behaviours were also paired at random, in order to establish a baseline value. The similarity coefficients of the randomly paired behaviours were then compared to the similarity coefficients of the matched pairs.


  1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References

In total, 20 aggressive behaviours (the purported OPBs) and three positive behaviours (the purported PABs) were identified. All participants were aggressive on at least one occasion. The number of aggressive behaviours per participant ranged from 1 (n = 2) to 5 (n = 2). Behaviours for which there was insufficient information to score the different categories used to compare predictions and actual behaviours were excluded. Five (25%) aggressive behaviours were excluded for this reason. Therefore, the final analysis included 15 actual aggressive behaviours and three positive behaviours.

A Jaccard's coefficient was derived for each comparison between a prediction and a follow-up behaviour. Table 1 displays the minimum and maximum Jaccard's coefficients derived from the analysis per participant for aggressive behaviours. In addition, similarity coefficients were derived for actual PABs and their corresponding predictions (see Table 1).

Table 1. Minimum and maximum Jaccard's coefficients per participant
ParticipantMinimum–maximum aggressive behavioursMinimum–maximum positive behaviours
  1. a

    A coefficient of .000 indicates that only one Jaccard's coefficient could be calculated for this participant.

  2. b

    A coefficient of .000 indicates that no positive behaviours were detected for this participant.

2.857–1.00 .000b–.000b
3.556–.625 .000b–.000b
4.000a–.833 .000b–.000b
5.667–1.00 1.001.00

In order to establish a baseline level of similarity, predictions were also randomly paired with actual aggressive behaviours. Table 2 displays the minimum and maximum similarity coefficients for 15 randomly paired predictions and actual aggressive behaviours.

Table 2. Minimum and Maximum Jaccard's coefficients for randomly paired aggressive behaviours and OPB predictions
ParticipantMinimum and Maximum Jaccard's coefficients
  1. a

    A coefficient of .000 means that only one Jaccard's coefficient could be calculated for this participant.


The Jaccard's coefficients of matched pairs were compared to the Jaccard's coefficients of random pairs. The Jaccard's scores for matched pairs were not normally distributed (matched D(15) = .25, p < .05 and random D(15) = .15, p > .05). Therefore, Wilcoxon signed-rank test was used to determine whether there was a statistically significant difference between the matched and paired Jaccard's scores. A statistically significant difference was found between the Jaccard's scores of the matched and random pairs; matched pairs (Median = .857) had a significantly higher similarity coefficient compared to random pairs (Median = .286), z = −3.108, p < .01).


  1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References

This study prospectively tested the OPB framework using predictions of OPBs and PABs derived from a novel practice algorithm (Jones, 2010a). Results revealed greater similarity between OPBs predictions derived from structured analysis of previous aggressive acts and actual incidents of inpatient aggression than randomly paired predictions and inpatient aggressive behaviours. The similarity of predicted and actual behaviour was significantly higher than what would be expected by chance. This result provides some validation for the OPBs framework by showing that predictions of OPBs based on careful analysis of an individual's history of violent behaviour can be used to predict inpatient aggressive behaviours.

The current findings differ somewhat from previous research conducted by Daffern et al. (2009) who did not find a significant difference between matched and random pairs of index offences and inpatient aggressive behaviour. Other studies (Daffern, Ferguson, et al., 2007, 2008) have also reported no relationship between pre-admission and in-patient aggression and between sexually abusive behaviour during hospitalization and sexual offence history.

There are several significant differences between the present study and previous research, which provide useful insights into the development and use of the OPBs framework. Firstly, Daffern et al. (2009) examined similarity between an index offence and in-patient aggressive behaviours whereas the current study considered all previous behaviours that might be relevant to the individual's violent offending cycle. By examining various offences and relevant behaviours the dominant themes within an individual's offending can be elucidated; analysing a discrete offence may not allow for persistent themes to be identified, resulting in invalid OPBs predictions. The current study's approach also allowed developmental sequences (Jones, 2004) to be identified. The lack of similarity between matched and random pairs in Daffern et al. (2009) study and the results of the present study suggest that a careful analysis of an individual's full offending and personal history must be taken into account when developing OPBs formulations and making OPBs predictions; an OPBs analysis should not be completed on an index offence alone.

In addition, this study utilized a wider variety of situational and dispositional factors to measure similarity than has been used previously. This makes the OPBs formulation process more robust and less prone to error or bias. It is also important to highlight that the present study used similar behavioural facets as the Daffern et al. (2009) study, although there were substantial differences in the protocol with regard to how information was extracted and recorded from case notes to enhance the coding of these nine characteristics. Specifically, this study used a variety of bottom-up (information derived from clinical notes and past Behaviour Cycle Logs) and top-down (theory driven) processes to formulate OPBs. A further strength of the present study was its prospective approach. This allowed for data to be collected systematically as it was entered on the clinical records, thus minimizing the chance of recall bias.


There are several issues that need to be considered when interpreting the results of this study. Firstly, this study evaluated OPBs in a small number of patients so although significant positive findings were evident further comparable research using a larger and more diverse participant population is required to confirm the reliability, validity and utility of Jones' practice algorithm. Although the number of participants in this study limits generalizability, the study aimed to pilot a new algorithm and to investigate whether it has sufficient validity to support further application on larger scale. This is important in ensuring optimal use of resources and clinicians' time.

In addition, there are several factors, which may have affected participation rate. For example, compared with routine clinical practice, where staff might have a good rapport with patients facilitating engagement in treatment, greater opportunity for observation and open discussion about patient's psychological issues, this study was conducted by an Honorary Researcher. Participants were unfamiliar with the researcher and participation was voluntary; available data was limited to that contained in patients' files. This is quite different to standard clinical practice where staff are generally trying to engage patients in treatment and management and therefore have increased access to a greater amount of clinical data to assess treatment needs and outcomes. In view of the above limitations and the positive findings of this study, it is possible that if this study is conducted by clinical staff in in-patient settings, the levels of OPBs and PABs could be higher. However, this needs further testing.

Although patients were involved in data collection and formulation, through the completion of the Behaviour Cycle Log, most of the data were derived from electronic records entered by clinical staff. This posed several issues: there was variation in the level of detail and amount of information entered by staff, which hampers comprehensiveness. This issue was addressed by identifying multiple sources of information for the specific behaviour and reviewing clinical notes 2 days before and after the behaviour. Also, entries into clinical files were not specifically related to the aims and purposes of the study; as such, some OPBs and PABs may have been observed but not reported in the clinical files.

A further limitation was posed by the use of multiple types of data triangulation, such as data and observer triangulation (collating data from multiple sources and observers) to increase validity and reliability (Flick, 1992). This procedure may have increased ‘leakage’ during the development of the reference formulations. This term refers to the use of multiple sources of information to generate hypotheses and formulations and to detect characteristics of follow-up behaviours, which in turn can lead to increased chance of detecting information that is being searched (C. Evans, personal communication). The researchers attempted to address this by following a structured protocol and testing the formulations against existing behaviours, which increased the chance of refuting them.

In addition to testing OPBs, this study set out to establish similarity between predicted and actual PABs. However, very few PABs were recorded so analysis of similarity was not possible. The low number of positive behaviours recorded raises concerns about bias when staff makes file entries. Although one reason for the lack of actual PABs may be due to the fact that the predicted PABs were invalid, or that prosocial behaviour did not emerge in these patients during the course of follow up, the present results may also suggest that when making clinical entries, staff are more likely to record problematic behaviour. This may be due to heightened focus on risk related behaviour. Finally, it may have been that patients who improved were granted increased unsupervised community leave; if these patients engaged in pro-social behaviour in the community these behaviours would have not been observed and recorded by staff. Future research on PABs could confirm the importance of establishing processes to help with the recognition and re-enforcement of PABs. It is also critical that staff extend their attention to PABs as the emergence of these behaviours could signal the development of pro-social change rendering the individual less likely to recidivate; this is critical information for risk assessors.

Lastly, although this study detected 15 aggressive behaviours, two participants were aggressive only once. This may provoke discussion on whether the OPB framework is a cost and time efficient tool for patients who do not display frequent aggressive behaviours on the wards. During the course of this study, several key areas emerged, which highlighted why exploring functional similarity across behaviours for all patients is crucial. Aggressive OPBs were the focus of this study because they are hypothesized to be more similar to sequences of behaviour evident in previous aggressive behaviours than other OPBs; if aggressive OPBs are topographically dissimilar to previous aggressive behaviours then it is conceivable that there would be very little similarity between aggressive behaviours in the community and in-patient behaviours. Since databases were reviewed to identify only aggressive OPBs, a number of topographically dissimilar OPBs may have occurred, which were not taken into considerations. More specifically, it is possible that non-aggressive OPBs exist, which have the same sequence and function as aggressive OPBs, but they differ in topography in that they are not represented by aggressive behaviour (e.g., disengaging from therapeutic activities, self-harm or defiant actions). In clinical practice non-aggressive OPBs may be more likely to go unnoticed and unchallenged than aggressive OPBs and thus may maintain offending behaviour. This may have significant impact on the assessment, intervention and management of offending behaviour in this population. An alternative explanation for the low number of aggressive behaviours observed could be that the patients have acquired Detection Evasion Skills (DES; Daffern, Jones, et al., 2007). Previous research (Daffern, Jones, et al., 2007) has noted that the absence of OPBs and PABs could signify that the individual has acquired skills to evade detection. Though all patients engaged in OPBs, which suggests DES were not fully establishes, the presence of DES must also be considered.

Conclusion and recommendations

The results of this study validate the OPB framework and suggest there is a similarity between predictions, based on OPB formulations derived from accumulated information on previous aggressive and other behaviour and actual in-patient aggressive behaviours. This finding has several important implications. Firstly, it suggests that in-patient OPBs can be predicted using a novel structured practice algorithm to establish a reference formulation and to generate OPB and PAB predictions. Prospective analysis of an individual's previous offending and personal history allows for important themes in an individual's offending to be identified, for the individual's full behavioural repertoire to be established and for its development and trajectory to be determined. This study has shown that predicted aggressive OPBs based on Jones (2010a) practice algorithm are similar to manifested aggressive behaviours within custody. This result differs from previous research on OPBs and emphasizes the need for clinicians to establish formulations based on an individual's entire history of aggressive behaviour, rather than their index offence alone, which has been shown to frequently lack similarity with in-custody aggressive behaviours (Daffern et al., 2009).

In relations to PABs, this study has highlighted an important clinical issue, namely how positive behaviours are recorded in clinical notes, and the extent to which they are considered with regard to patients' treatment and progress. It is possible that bias exists in the way clinical notes are recorded. Future research on PABs could re-affirm the need for establishing, in clinical settings, processes to help with the recognition and re-enforcement of PABs. It is also possible that by helping patients to recognize PABs this would positively impact on their belief in behavioural change and self-efficacy. However, this hypothesis needs testing.

The validity and reliability of the OPB practice algorithm explored here would be strengthened by replication of this study using a larger sample of in-patients or prisoners and including review and monitoring of a wider range of OPBs, in particular expanding the study of in-custody behaviours to include non-aggressive OPBs.


  1. Top of page
  2. Abstract
  3. Background
  4. Method
  5. Results
  6. Discussion
  7. References
  • Bennell, C., & Canter, D. (2002). Linking commercial burglaries by modus operandi: Tests using regression and ROC analysis. Science and Justice, 42, 153164.
  • Clark, D. D., & Crossland, J. (1985). Action systems: An introduction to the analysis of complex behaviour. London and New York: Methuen.
  • Daffern, M., Ferguson, M., Ogloff, J., Thomson, L., & Howells, K. (2007). Appropriate treatment targets or products of a demanding environment? The relationship between aggression in a forensic psychiatric hospital with aggressive behaviour preceding admission and violent recidivism. Psychology, Crime and Law, 13, 431441. doi:10.1080/10683160601060861
  • Daffern, M., & Howells, K. (2009). The function of aggression in personality disordered patients. Journal of Interpersonal Violence, 24, 586600. doi:10.1177/0886260508317178
  • Daffern, M., Howells, K., Mannion, A., & Tonkin, M. (2009). A test of methodology intended to assist detection of aggressive offence paralleling behaviour within secure settings. Legal and Criminological Psychology, 14, 213226. doi:10.1348/135532508X342919
  • Daffern, D., Howells, K., Stacey, J., Hogue, T., & Mooney, P. (2008). Is sexually abusive behaviour in personality disordered inpatients analogous to sexual offences committed prior to hospitalisation? Journal of Sexual Aggression, 14, 123133. doi:10.1080/13552600802267080
  • Daffern, M., Jones, L., Howells, K., Shine, J., Mikton, C., & Tunbridge, V. C. (2007). Refining the definition of offence paralleling behaviour. Criminal Behaviour and Mental Health, 17, 265273. doi:10.1002/cbm.671
  • Daffern, M., Jones, L., & Shine, J. (2010). Offence paralleling behaviours. UK: Wiley-Blackwell.
  • Flick, U. (1992). Triangulation revisited: Strategy of or alternative to validation of qualitative data. Journal of the Theory of Social Behaviour, 22, 175197. doi:10.1111/j.1468-5914.1992.tb00215.x
  • Gordon, A., & Wong, S. C. P. (2010). Offence analogue behaviours as indicators of criminogenic need and treatment progress in custodial settings. In M. Daffern, L. Jones, & J. Shine (Eds.), Offence paralleling behaviour: A case formulation approach to offender assessment and treatment (pp. 105120). Chichester, UK: John Wiley & Sons.
  • Gresswell, D., & Hollin, C. R. (1992). Towards a new methodology for making sense of case material: An illustrative case involving multiple murder. Criminal Behaviour and Mental Health, 2, 329341.
  • Grubin, D., Kelly, P., & Brunsdon, C. (2001). Linking serious sexual assaults through behaviour: HORS 215. London: Home Office Research and Development and Statistical Directorate.
  • Hart, S., Sturmey, P., Logan, C., & McMurran, M. (2011). Forensic case formulation. International Journal of Forensic Mental Health, 10, 118126. doi:10.1080/14999013.2011.577137
  • Hodgins, S. (2007). Persistent violent offending: What do we know? British Journal of Psychiatry, 190, s12s14. doi:10.1192/bjp.190.5.s12
  • Hollin, C. R. (1990). Cognitive-behavioural interventions with young offenders. New York: Pergamon Press.
  • Howells, K., Daffern, M., & Day, A. (2008). Aggression and violence. In K. Soothill, M. Dolan & P. Rogers (Eds.), Handbook of Forensic Mental Health (pp. 351374). Cullompton: Willan Publishing.
  • Jones, L. (2004). Offence paralleling behaviour (OPB) as a framework for assessment and intervention with offenders. In A. Needs & G. Towl (Eds.), Applying psychology to forensic practice (pp. 3463). Oxford: Blackwell Publishing.
  • Jones, L. (2010a). Approaches to developing OPB formulations. In M. Daffern, L. Jones, & J. Shine (Eds.), Offence paralleling behaviour: A case formulation approach to offender assessment and intervention (pp. 7188). UK: Wiley-Blackwell.
  • Jones, L. (2010b). History of the offence paralleling behaviour construct and related concepts. In M. Daffern, L. Jones, & J. Shine (Eds.), Offence paralleling behaviour: A case formulation approach to offender assessment and treatment (pp. 324). Chichester, UK: John Wiley & Sons.
  • Kohlenberg, R. J., & Tsai, M. (1994). Functional analytic psychotherapy: A radical behavioural approach to treatment and integration. Journal of Psychotherapy Integration, 4, 175201.
  • McDougall, C., Clark, D. A., & Fisher, M. (1994). Assessment of violent offenders. In M. McMurran & J. Hodge (Eds.), The assessment of criminal behaviours in secure settings. London: Jessica Kingsley.
  • National Audit Office. (2008). Reducing the risk of violent crime. (HC 241, Session 2007-2008).
  • Shoda, Y. (1999). A unified framework for the study of behavioural consistency: Bridging person x situation interaction and the consistency paradox. European Journal of Personality, 13, 361387. doi:10.1002/(SICI)1099-0984(199909/10)13:5<361
  • Silver, J. M., & Yudofsky, S. C. (1987). Documentation of aggression in the assessment of the violent patient. Psychiatric Annals, 17, 375384.
  • Sturmey, P. (2007). Functional analysis in clinical treatment. Burlington: Elsevier.
  • Ward, T., & Beech, T. (2006). An integrated theory of sexual offending. Aggression and Violent Offending, 11, 4463. doi:10.1016/j.avb.2005.05.002
  • Woodhams, J., Hollin, C. R., & Bull, R. (2007). The psychology of linking crimes: A review of the evidence. Legal and Criminological Psychology, 12, 233249. doi:10.1348/135532506X118631
  • World Health Organisation (2002). World report on violence and health- summary. Geneva: World Health Organisation.
  • Young, J., Klosko, J. S., & Weishaar, M. E. (2003). Schema therapy: A practitioner's guide. New York: Guilford.