Goals and Plans Card Sort Task: A Psychometric Assessment Tool to Measure and Support Life Goal Pursuits in People Who've Offended

Abstract Models of offender motivation to change exist, however there is a lack of theory-driven psychometric tools that measure motivational constructs to support offenders in positive life goal pursuits. This research extends the Personal Concerns Inventory (Offender Adaptation), presenting a Goals and Plans tool that supports users to: identify and prioritize goals in life domains of importance to them; detail how a goal can be attained; and consider obstacles to attainment. Literature informed the benchmark for the tool, which was evaluated through implementation in Prisons (n = 62) and Approved Premises (n = 105) across Wales, UK. Results indicate that goals in the life domains of Home & Future Living, Relationships, Physical & Mental Health and Learning & Working were prioritized. Goal attainment/restrictions were influenced by perceived control over a goal. Goals that relied upon external factors (i.e., services) reflected a more maladaptive motivational structure, whereas those that relied upon internal (self-focused) barriers to participation reflected an adaptive structure. This research offers practitioners a tool to assist users to ‘plan for the future’ and monitor progress by capturing motivations and assessing factors that might impact the likelihood of a goal being pursued.

Treatment programmes that aim to reduce recidivism rely upon an individual's motivation to engage with treatment and motivation to change (Mallion et al., 2020;Sellen et al., 2006). Indeed, motivation to change could be regarded as a criterion for participant selection into programmes, as resources are limited and because motivation is considered an essential component of treatment success (Hachtel et al., 2019;McMurran & Ward, 2010). However, there are a limited number of validated and theory-driven psychometric tools that measure motivational constructs in offenders (McMurran & Ward, 2004), with studies primarily relying on practitioner interviews and ratings (Mallion et al., 2020). Cox and Klinger (1988;1990;2002; propose motivational behavior to be goal orientated, founded on cognitive and emotional processes that drive goal choice and outcome responses. They consider motivational structures to be either 'adaptive' or 'maladaptive'; where a goal is representative of an adaptive motivational structure, an individual might recognize and commit to the goal, have pathways to achieving it and expect pleasurable engagement from goal achievement. Goals representing a maladaptive motivational structure are characterized by a perceived lack of control over achievement of goals, owing to an individual's lack of awareness of the strategies they will need to employ to reach their goals (Hosier & Cox, 2011). For example, emotional satisfaction may not be derived from goal achievement, or there is a perceived low level of expected success with goal attainment (Cox & Klinger, 2002). Core to this principle is the Theory of Current Concerns; a time-binding process which begins when a person becomes committed to their goal and ends when the goal is terminated . The Theory of Current Concerns was initially developed to understand the motivational structures associated with problem drinking  with the Personal Concerns Inventory created to examine motivational structures from the perspective of goal pursuits in this group (Cox & Klinger, 2000).
The Personal Concerns Inventory aimed to help users examine how alcohol misuse affected goal development and attainment, with the intention of using this to motivate change (Cox & Klinger, 2000). It requires users to identify and describe goals in 12 life areas (e.g., home and household; employment and finance) and to rate each goal on a scale of one (none at all) to 10 (the most that I can imagine) to best describe each goal's perceived importance, achievability and control (Cox & Klinger, 2000). Cox and Klinger (2002) report the predictive validity of the tool, identifying that adaptive motivation (a focus on the perceived importance, achievability and control over a goal), predicted motivation to change in individuals whose drinking was problematic. The selfreport rating scales described above where used to assess an individual's motivational profile in terms of importance, likelihood of attainment, and expected affective change if the goal were attained (Cox & Klinger, 2002). Sellen et al. (2006) propose Cox and Klinger (1988;1990; motivational model is also relevant in the context of offending behavior, suggesting that offending behaviors are maintained by prioritizing shortterm gains over long-term costs and failing to maximize potential in life. Further, Sellen et al. (2006) argue that the 12 life areas of the Personal Concerns Inventory resemble Ward's (2002) 'primary goods' in the Good Lives Model (GLM), which assumes that human's value certain states, personal characteristics and experiences. According to the GLM, offending behavior can stem from the desire to attain goals that are personally meaningful, but weaknesses within and surrounding the individual (i.e., their environment) prevents them from attaining goals in pro-social and sustainable ways (Ward, 2002). The relative weighting of goals is proposed to reflect an individual's values and priorities. Consequently, Sellen et al. (2006) proposed an adaptation of the tool for use with offenders; The Personal Concerns Inventory (Offender Adaptation) (PCI-OA). The PCI-OA is administered via a semi-structured interview where users are required to describe their current concerns in each life area and rate them in terms of value, attainability, controllability and imminence (Cox & Klinger, 2002;Sellen et al., 2006). The life concerns included are home and household matters; employment and finance; partner, family, and relatives; friends and acquaintances; love, intimacy, and sexual matters; selfchanges; education and training; health and medical matters; substance use; spiritual matters; hobbies, pastimes, and recreation; my offending behavior; current living arrangements; and other areas. Pilot data with 11 male prisoners showed that offenders' motivational structures could be identified within the PCI-OA by allowing individuals to identify specific concerns and help direct support needed for goal attainment (Sellen et al., 2006). While users in the pilot discussed the positive implications of breaking down concerns into smaller manageable problems, the authors do not report discussing solutions to these problems. Thus, it is unclear whether the tool is a measure of an individual's motivation to change, or a mechanism for enhancing motivation by encouraging users to consider how they might overcome concerns or both (Sellen et al., 2006).
The psychometric validation of the PCI-OA was further examined with 129 adult male prisoners, comparing outcomes from those motivated to change (in treatment programmes, n ¼ 64) to those not motivated to change (those not in treatment, n ¼ 65) (Sellen et al., 2009). Scale scores within and between the treatment and non-treatment groups were examined as a measure of predictive validity, with moderate changes over the course of treatment found for levels of control only. PCI-OA scores were also compared to scores from other measures of motivation to test concurrent validity: The University of Rhode Island Change Assessment (URICA; McConnaughy et al., 1983;1989); a measure of treatment motivation (Treatment Motivation Questionnaire) (TMQ; Ryan et al., 1995), and self and staff ratings of change. Outcomes showed some support for the tool's ability to index an adaptive and maladaptive motivation, with a positive correlation between the PCI-OA, The University of Rhode Island Change Assessment scores and the internal motivational profile (i.e., behavior that is self-determined) of the Treatment Motivation Questionnaire. No correlation was found between the PCI-OA and the external motivational profile of the Treatment Motivation Questionnaire, which relates to behavior motivated by external regulation (Sellen et al., 2009). However, it is worth noting that The University of Rhode Island Change Assessment has been criticized for not accurately representing the change process (Casey et al., 2005) and the Treatment Motivation Questionnaire lacks an empirical basis with offenders (Bonta & Andrews, 2003). Finally, structural validity was examined through a comparison of the PCI-OA factor structure with that of the original Personal Concerns Inventory, which showed some congruence between the adaptive and maladaptive index. However, there was a lack of validity for the third factor in the PCI-OAlack of directionwhich encompassed unhappiness in goal attainment, not knowing how to achieve the goal, and feelings of reward from offending behavior.
The strength of the PCI-OA (Sellen et al., 2006) is that it encourages users to describe their goals and provides a measure to quantify the users' perceived likelihood of achieving their goal, how much control they have over it, its importance, imminence and the expected affective change. Thus, the PCI-OA provides a basis for assessing an individual's motivational structure. However, the tool was not designed to allow users to prioritize and focus upon life domains of most importance to them. Nor does it consider how incentive values (i.e., the perceived value of a goal or action) might change overtime due to changes in circumstances or the environment.
The Model of Human Occupation (MOHO) (Kielhofner, 2008), which is used widely in occupational therapy in forensic mental health, explains the ways in which goals-based outcomes are driven by both personal and environmental factors that influence motivation and capacity to engage in occupation. The model acknowledges that the 'environment' provides opportunities, resources, conditions and restraints in relation to goals (Kielhofner, 2008). In its current form, the PCI-OA pays limited attention to identifying and overcoming the factors outside the individual which might affect motivation. For example, working toward a satisfying and positive life might include a focus on a few key areas such as positive support systems, pro-social behavior or participation in training or employment (Fortune, 2018;Mann et al., 2014). This is also central to forensic recovery approaches (Mann et al., 2014), which refers to the process of shifting goals, attitudes and values that promote a satisfying and positive life without being limited by offending behavior (Tomlin & Jordan, 2021). Thus, the assessment of goal prioritization and the explicit discussion of the steps needed to achieve positive change is essential when considering current concerns and motivation to succeed in goal attainment. Tools that aim to measure motivational structures would likely benefit from drawing upon forensic recovery approaches and the MOHO's holistic framework to consider how personal and environmental factors relate to motivation and performance (Connell, 2016).

Study aims
The current research was conducted in forensic settings across Wales, UK and aimed to develop and evaluate a Goals and Plans tool created to a) identify future goals and plans and b) assess (and troubleshoot) factors which could impact on the likelihood of the goal being pursued. Based on the PCI-OA, the tool aimed to address the limitations of the PCI-OA and improve usability by allowing users to prioritize goal choices in life domains that matter the most to them. Specifically, this study aimed to design a more user-friendly version of the PCI-OA by: (a) expanding the theoretical framework for identifying goals and plans to consider the personal and environmental factors that influence motivation and performance; (b) considering factors which could impact the likelihood of the goal being pursued; (c) evaluating the structural and implementation process of the tool; and d) responding to limitations and opportunities for improvement of the tool.

Ethical consideration
Data for this study were collected as part of a wider project examining the impact of the setting on staff and residents in prisons and approved premises (AP) (residential units that house, monitor and manage exprisoners in the community to primarily including those on license with a small number of individuals on bail or community sentences). Ethical approval was received from Wales REC 3 (Reference: 14/WA/ 0150) and the study was registered with the National Research Committee (reference: 2014-159).

Process of developing a goals and plans tool
The Goals and Plans Tool underwent two stages of development as described below.
Stage 1: Tool development, piloting and learning. The theoretical framework of the PCI-OA broadened that of the original Personal Concerns Inventory (i.e., the Theory of Current Concern) by reflecting core ideas from the GLM (Ward, 2002) to promote and facilitate goal attainment resulting in positive (prosocial) change. The current work aimed to extend this by allowing respondents to prioritize goals and consider how goal attainment is mediated by the interaction between personal and environmental factors. Thus, addressing the limitation of the PCI-OA by capturing a deeper understanding of a person's decisionmaking and motivation for change. In considering the principles of the GLM, recovery approach and MOHO, the theoretical framework of the Goals and Plans Tool enables practitioners to assess how a person is motivated, allowing for reflection of what is needed to achieve a goal and any factors that may restrict success (Connell, 2016). To tool therefore provides respondents an opportunity to consider how we perform occupation, set expectations and measure our competence in succeeding these expectations (Lee & Kielhofner, 2010). This was considered together with clinical input from forensic psychology, occupational therapy and probation research specialists, which included discussion around the tool's format and content based on experience implementing the PCI-OA.
The initial Goals and Plans Tool (see Table 1) took the form of a structured and sequential booklet and contained 10 life domains (adapted from the Personal Concerns Inventory; Cox & Klinger, 2000 and PCI-OA;Sellen et al., 2006) and an 'other' option to allow respondents to detail goals not presented.
Three questions pertaining to goal choices, attainment and obstacles were included for each life domain: 'what I want'; 'how will I do this'; and 'what could get in the way of this?' Textboxes were provided for respondents to write comments and thoughts for each question. A rating scale for goals in each domain were included on a scale of 0 (not at all) to 10 (extremely) for: a) goal importance, b) control (relating to goal achievement), c) happiness (if the goal was achieved), d) knowing what to do (to achieve the goal) and e) the likelihood of achieving the goal. Highly adaptive motivation was considered present when scores on all factors were high (individual score of 7 and above). Areas of maladaptive motivation were indicated by low scores (individual score of 3 and under) on any factor (indicating a possible need for specific intervention) with highly maladaptive motivation/goal seeking indicated when two or more factors received low scores. A timescale in which the goal might be achieved was also included. As with Sellen et al. (2009) study, users were given four options: within 1-month, within 6-months, within the next year and sometime in the future. Three rating scales originally included in the PCI-OA were not included: unhappiness, offending help and offending interference. This followed pilot use of the PCI-OA by a probation research specialist (author) which identified confusion amongst respondents as to what these categories meant. This echoed Sellen et al. (2009) validation of the tool.
Men from four Approved Premises in Wales, UK (N ¼ 18) and a category C prison in Wales (N ¼ 20) undertook the Goals and Plans booklet between December 2014-June 2015. Category C prisons house those who cannot be trusted in open conditions but who are unlikely to try to escape. Residents were recruited opportunistically, either through open access on the wing/AP for researchers to approach potential participants, or via access to specific areas within the setting (i.e., education wing, lounge). Respondents were asked to examine the extent to which the measure covered areas important to them and the extent to which the tool enabled them to provide full answers which reflected their goals, views about reaching these goals and concerns about obstacles. All residents were recruited opportunistically, were male over the age of 18 and convicted of a range of offenses.
No specific demographic or offense data is available due to the requirements for anonymised data stipulated as part of the ethical approval. APs were included as previous research has primarily focused on prison populations (Sellen et al., 2006;Sellen et al., 2009). The domains from the Goals and Plans Tool were presented sequentially in an A4 booklet, the order and format of which was consistent across all participants. If able, participants completed the task independently (following a brief set of instructions and supported by a short completion guide); where needed participants were provided Homefuture living Hopes and plans in relation to living accommodation, location, and the level of domestic support needed for maximum independence.

Relationshipsfriends and family
Hopes and plans in relation to future relationships with family and friends. Physical and mental health Hopes and plans in relation to physical and mental wellbeing.

Hobbies and interests
Hopes and plans in relation to personal interests and hobbies which may supplement other activity (e.g. work) or be a focus for 'occupation' in the future. Learning and working Hopes and plans in relation to future work (paid or voluntary) and education / learning desires and opportunities.

Money
Hopes and plans in relation to finances including how the individual will manage their money. Self-changes / Personal development Hopes and plans in relation to personal development and changes the individual would like to make to themselves. Drug and alcohol management Hopes and plans in relation to dealing with substances.

Risk management
This domain in concerned with hopes and plans in relation to managing issues of future risk. Pro-social behavior Hopes and plans in relation to positive interactions and role in society. Other This can be used to capture any other goals or plans the individual might have that don't fit into the previous areas.
with support from a researcher. Most respondents took between 15-30 minutes to complete the tool, with a small number of individuals taking longer as they considered each question. This extended testing phase revealed that respondents had difficulty in providing information on all presented life domains with a noticeable waning of interest when providing answers to later occurring domains. It was also noted that important life goals were sometimes listed against early (less appropriate) domains when a more appropriate domain existed later in the booklet. Respondents also reported difficulty estimating the timescale in which some domains could be achieved, especially where achieving the goal required input from others/services. The average number of domains identified by each user at this stage was 4.3, with a range of 1-7. Almost all AP residents (94%: n ¼ 17/18) identified goals in the domain of Home and Future Living, which primarily related to finding accommodation. Goals in the domain of Learning and Working were provided by 61% of residents (n ¼ 11/18) and almost half of the participants identified goals in the domains of Relationships, Physical & Mental Health and Hobbies & Interests (44%: n ¼ 8/18). Less than a third of participants indicated having a goal of managing their risk (n ¼ 5/18), which included avoiding certain areas or behaviors and/ or self-improving and maintaining current standards.
Stage 2: Tool modification and implementation in practice. The learning provided by the extended testing phase enabled a number of revisions to improve usability. These changes related to the approach to completion and the content of the tool. The rating of each domain was replaced by a card sort activity (Blanchard & Banerji, 2016) during which the life domains could be ranked in order of importance to the individual (1-10). This removed the assumption that all domains were relevant to all respondents and enabled respondents to visualize and physically move goals to compare and reflect on domain importance to reach the final ranking. It was anticipated that this approach to prioritizing domains could also facilitate the assessment of changes to individual priorities over time, and enable common goal differences between prison and AP populations to be identified (possibly influenced by environment or wider circumstances). Further, the modified version only requires respondents to answer questions relating to the highest ranking four domains. This number reflected the mean number of domains completed by participants in the extended testing phase.
The content of the tool was revised in three ways. The 'other' domain option was removed as goals identified under this heading could be readily categorized into one of the existing 10 domains, and the 'timescale' option was removed as respondents found this difficult to estimate. Finally, the 'knowing what to do' question was removed as outcomes from this question (yes or no) were not considered useful and did not allow for assessment as to whether respondents fully understood the steps required for goal attainment, or if these steps include maladaptive motivations.
The Goals and Plans Tool was therefore reworked into a card sort activity (e.g., Blanchard & Banerji, 2016) where respondents were asked to sort and rank concepts (i.e., life domains) in order of importance, where 1 represented the most important life domain and 10 the least, and identify specific goal(s) for their four top-ranked life domains. For each goal, respondents were asked to write comments in text boxes relating to 'how will I do this'; 'what could get in the way of this' and 'how I can overcome things that get in the way' and rate these goal(s) on a scale of 0 (not at all) to 10 (extremely) on levels of importance, likelihood of achievement, level of control and commitment. Once the four domains were worked through, respondents were permitted to add further domains if they wished to.
In the revised task, each of 10 life domains from stage 1 were presented on individual A6 cards which also included prompts to help respondents consider potential goals in relation to each domain. An A4 Goals and Plans Recording Form was developed for respondents to record their responses as they worked through the procedure (i.e., identify goals, write the steps needed for goal attainment, describe potential obstacles and rate goals to index importance, control, happiness and likelihood of achieving the goal for their top four goals). The adapted card sort task was evaluated on a larger sample of prison and AP residents in Wales to assess its utility across the services.

Participants: card-sort task
Opportunistic sampling was used to recruit residents from three public sector and one private sector category B/C adult male prisons (n ¼ 62) and 4 AP (n ¼ 87) sites in Wales, UK between April 2016-October 2019. Private sector prisons are managed by private companies but are inspected by Her Majesty's Inspectorate of Prisons in the same way as government run (public sector) establishments. Those placed in Category B prisons do not require maximum security, but for whom escape still needs to be made very difficult. Security categorization C, include those who cannot be trusted in open conditions but who are unlikely to try to escape. Five people were also recruited having moved from an AP into their own accommodation. Some sites provided open access for the researchers to approach potential participants within the wing or specific setting (i.e., workshop or lounge area). At other sites staff brought residents to the researchers. With the latter format it is not known if all residents were invited to participate or if staff unofficially 'selected' possible participants.
Participants were male, over the age of 18 and convicted of a range of offenses. No specific demographic or offense data is available due to the requirements for anonymised data stipulated as part of the ethical approval. This is further addressed in the limitations sections of the discussion.

Approach to analysis
Descriptive statistics were used to characterize goal choices as an index of goal importance, control, happiness, knowing what to do and likelihood of achieving the goal. Pearson's correlation analysis explored the potential relationship amongst the ratings made. Correlation coefficient of .10 represents a weak correlation; a correlation coefficient of .30 is moderate; and a correlation coefficient of .50 and above represents a large correlation (Cohen, 1977). The Mann-Whitney U test was used to determine if there were differences between AP and prison respondents for the top ranked domains. Analysis was carried out using SPSS version 23 (IBM Corporation, Armonk, NY, USA). Prior to analysis the data were first reviewed to check quality and accuracy and to establish if they met assumptions for the parametric test.
Qualitative findings will be reported in the form of text box responses relating to identified goals, steps needed for goal attainment and potential obstacles. Text boxes were assessed descriptively to consider the main themes that emerged from the data (Lambert & Lambert, 2012). These descriptive texts provide additional context to participant's responses. All participants provided explanatory text for the four top rated life domains.

Card sort task: identification of life domains across prison and AP residents
As shown in Table 2  Notes: Import.
Goals that were perceived to rely upon support from or facilitation by others reflected a more maladaptive profile. For example, Home and Future Living was ranked as a priority goal for 30% of respondents (n ¼ 47) with a high level of importance and commitment. However, this domain also had the lowest perceived level of control and likelihood in achieving the goal: Goal: "I will be homeless when I get out of jail. The goal of gaining safe accommodation is of the highest priority so I do not re-offend." To achieve: "all I can do on this matter is to do my very best to follow the instructions of the housing team and hope I get placed in supported housing" Obstacles: "This goal is effectively out of my control" (Participant (P) 1) Text box responses for identified goals, steps needed for goal attainment and potential obstacles found that goals where achievement and obstacles were self-focused represented a more adaptive structure. For example, goal attainment in the domains of Relationship (ranked as a top priority by 28% (n ¼ 43) of respondents), and Physical and Mental Health (ranked as a top priority by 9% (n ¼ 14)) were considered to rely upon respondents making changes to their behavior or lifestyle. Levels of control and likelihood of goal achievement were rated higher than for Home and Future Living in both the domains of Relationships and Physical and Mental Health.
Goal: "I want to gain a healthy and close relationship with my family as I have not been able to spend time with them. That is needed due to being in and out of jail for 10 years." To achieve: First I need to let go of the anger inside of me from the past and realise how valuable relationships are. I bottle up and hide away." Obstacles: "Myself, due to being anxious and having hate towards everyone" (P2) Goals in the domain of Drugs and Alcohol, which were identified as a priority domain by only 4% (n ¼ 6) of respondents, principally reflected an adaptive profile. This was the domain with the highest level of control, commitment and happiness, although achievability of this goal was perceived to be lower. Goal achievement and obstacles to attainment were primarily self-focused and relied upon the individual's motivation to commit to treatment and seek support.
Pro-Social Behavior was identified as a priority domain by only two participants (1%) across both the AP and prison residents. Yet, goals in this domain were scored highly on levels of importance, commitment and happiness. This may in part be due to goals in this area overlapping with other domains. For example, goals in this area were sometimes seen as a means for reaching goals in the domain of Relationships (developing and maintaining connections with family and friends).

Difference between AP and prison respondents
As shown in Table 3, AP respondents prioritized goals in the domain of Home & Future Living, with prison respondents prioritizing the domain of Relationships. Both groups ranked Hobbies & Interests and Pro-Social Behavior as being the least important, with only one person in each group prioritizing goals in these domains. However, a Mann-Whitney U test found the top ranked domains were not statistically significantly different between AP and prison respondents (U ¼ 3,056, p ¼ .119, r ¼ .13).

Rating scales
As shown in the correlation matrix (Table 4), levels of importance showed a significant and positive correlation with commitment (r ¼ .923, p < .001) and happiness (r ¼ .918, p < .001). Commitment and happiness also positively correlated (r ¼ .933, p < .001). This pattern can be seen in Table 3; goals rated high on levels of importance were also rated high on levels of commitment and happiness, representing an adaptive structure (for example, Drugs & Alcohol).
Control over and likelihood of achieving a goal positively correlated (r ¼ .652, p ¼ .040). Where control represented an adaptive profile (i.e., rated highly), so did likelihood of achievement (for example, Risk  Management). Conversely, goals with maladaptive patterns (low ratings) of control also had lower perceptions of goal achievement (for example, Home and Future Living). Overall, whilst individuals rated goals as being important, there was a perceived low level of control over goal attainment. This might in part be due to respondents considering some goals to be reliant upon external services providing support (i.e., probation or housing agencies finding appropriate accommodation). So, whilst respondents might express commitment to goal attainment, expectations for goals' achievements are likely impacted by license conditions and/or and reliance on services.

Discussion
This research aimed to present a Goals and Plans Tool for use within forensic settings to help individuals explore goals in life domains that are compatible with a future of non-offending. The final card sortbased tool extends the work of Sellen et al. (2006;PCI-OA) and is theoretically grounded in Cox and Klinger's motivational model, the GLM, recovery approach and MOHO (i.e., Cox & Klinger, 2000;Mann et al., 2014;Ward, 2002). Asking users to prioritize goals, consider how a goal might be achieved and to reflect upon potential external (i.e., the environment) and internal (self-focused) barriers to participation addresses the limitations of existing tools. Capturing barriers to participation can support practitioners in their capacity to intervene and support prosocial participation (Connell, 2016;Connell et al., 2019).
The current work offers an assessment tool that can be used to measure these factors across multiple settings in the criminal justice system (i.e., prison to an Approved Premises in the community), centering on the individual's motivation to lead a positive life and presenting a strength-based approach to self-managed behavior. No significant differences in the prioritization of life domains was found between the two participant groups, with both groups prioritizing goals in the domains of 'Relationships' and 'Housing & Future Living.' This may be because social needs are critical to the achievement of basic resources such as employment and money (Bowman & Ely, 2020;Denney et al., 2014). Access to a support system with pro-social individuals is vital to overcome barriers to participation, and much work has reported on the link between social ties and support and reduced likelihood of recidivism for those living in both prison and the community (Connell et al., 2019, Maruna, 2001, Ward, 2002.
The data showed that individuals were more motivated to succeed for goals of higher importance (and those which could promote a pleasurable response) when control and 'likelihood of achievement' were high. This reflects the foundation of Cox and Klinger's motivational model which proposes goal attainment and outcomes are connected to emotional experience, and further emphasizes the value of interventions that facilitate goal attainment in areas of personal interest to the individual, thus promoting positive change (Connell, 2016;Ward, 2002). However, goals ranked as being low priority, such as Pro-Social Behavior, also scored highly on levels of importance, commitment and happiness; it might be that there are hierarchies of the domains or that some domains are nested within others. For example, where they were reported, goals relating to Pro-Social Behavior were commonly relevant to other domain areas, such as Relationships. Thus, success or failure of these lower-rank needs might act to motivate behavior for higher ranking domains. In this way, some goals might be classed as secondary goods which function to mediate the attainment of primary goods as described in the GLM (Ward, 2002;. The GLM proposes that secondary goods can be maladaptive and interventions should focus on the interchange between maladaptive to adaptive goods (i.e., socially acceptable goals) (Ward, 2002;. Primary and secondary goods can act as both essential human needs and motivators for offending behavior depending on factors such as access to social support, economic or housing stability (Chu et al., 2015;Bowman & Ely, 2020;Willis & Ward, 2011). Prosocial behavior, for example, is reported to increase when individuals are supported to become financially independent and secure stable housing (Bowman & Ely, 2020). In this case, independence and stability is the strategy (i.e., secondary goods) used to increased optimism in attaining long-term change and purposeful and fulfilling goals (Bowman & Ely, 2020). Therefore, goal attainment in domains not prioritized in the present study but ranked highly on levels of importance (i.e., Pro-Social Behavior) might provide individuals with the resources and skills required to succeed in higher-ranked domains.
The strength of the current research is that it offers practitioners a way to capture maladaptive motivations and dysfunctional occupation to assess factors (both personal and environmental) that might impact on the likelihood of a goal being pursued. This is vital to ensure a holistic approach to the assessment of motivations in offenders that consider how the setting can provide a significant context for performance which can subsequently shape behavior and promote or decrease motivations (Parkinson et al., 2006). Additionally, detailing obstacles and focusing on how these might be overcome provides a clear way for individuals and supporting staff to intervene to maximize the chances of goal achievement.

Limitations and future direction
The strength of the current research is that it offers practitioners a way to capture motivations and assess factors that might impact the likelihood of a goal being pursued. The Goals and Plans Tool may also be useful as an outcome/monitoring tool through repeated use over time. The current work did not aim to explore changes over time, however the potential to review and revisit goals and motivation across time and in different settings/points within the criminal justice system could: a) help identify goal achievement; b) identify barriers to goals; and c) better explain the relationships between goals and offending. However, longitudinal research is needed to understand how goals are impacted by other variables such as the prison or AP environment (Liebling et al., 2019). Research is also needed to assess how successful individuals might be at following plans to achieve goals and overcoming the potential barriers they have identified and developed a mitigating strategy for, such as the need to access additional support to facilitate goal attainment in the long-term.
For this study, it was not possible to consider the impact of offense type or sentence length on goal attainment due to restrictions within our ethical approval. Nor was it possible to consider heterogeneity across different demographics groups, limiting the generalization of findings. Future work should consider the utility of the tool across other populations, such as women and young offenders, and whether demographic factors, such as age and ethnicity, might impact goal selection and attainment. Future studies should also consult with practitioners to consider how useful they consider the tool to be and address any limitations or challenges they might experience in implementing this in prisons and the community. Additionally, it would be useful to consult with relevant staff, such as occupational therapists and offender managers, to examine whether they consider the adapted Goals and Plans Tool to be more user friendly that the PCI-OA and whether the format of the tool is appropriate. Whilst card sorts tasks have been used in a variety of contexts (Blanchard & Banerji, 2016), and have been found to be less taxing and tedious than alternative psychological approaches, and more pleasurable, further assessment of the tool's usability is warranted. Further, there is a need to formally consider the validity and reliability of the tool across a larger and more diverse sample of prison and community participants.
Finally, future research should consider how the tool might be used to support the resettlement of prisoners, especially as motivational structures may be modified following changes in environment or lifestyle (Cox & Klinger, 2000;Sellen et al., 2006). Several researchers have identified links between individual struggles (i.e., adapting to life outside of prison), maladaptive motivations and offending behavior (Liebling et al., 2019;Williams, 2016). Whilst the current research explored potential differences in the motivational profile of prison and AP residents, the tool might be used to prepare prisoners for release and support them in goal attainment whilst on probation and beyond.

Implications for practice
The tool was designed to be completed by or under the guidance of an Occupational Therapist, Psychologist or appropriately trained individual with the information collected being incorporated into intervention, support and/or care packages. Whilst the tool can be used as a 'one off' assessment, it is considered most useful when used at intervals to support the respondent to monitor the progress toward goals and to reevaluate their goal priorities over time. In this way, the tool can help users and practitioners to examine the impact of environmental changes (e.g., a move from a prison to an AP or the community), to inform pathway/sentence priorities and planning, and to assess the impact of goal attainment on the individual's engagement in a pro-social and healthy lifestyle.

Conclusion
Approaches to reducing recidivism have, in the past, primarily focused on criminogenic needs related to offending behaviors and risk (Maruna & LeBel, 2003;Ward et al., 2006). However, to reduce reoffending forensic services must consider contributing factors and assess the impact interventions have on risk (Connell, 2016). The Goals and Plans Tool presented here offers a psychological assessment tool to help respondents 'plan for the future' and assist practitioners in monitoring progress. The Goals and Plans Card Sort Task and Guidance for Use document is available from the authors upon request.