Abstract
Professional decision making in forensic clinical practice may have lifelong consequences for offenders. Although information on moral development is important for prediction of reoffending and referral to adequate treatment, conclusions regarding moral development are still largely based on unstructured clinical judgment instead of assessment instruments. For this study, the authors examined to what extent unstructured clinical judgment of both moral judgment and victim empathy concurred with test results in a group of young sex offenders. Moral judgment was measured with the Sociomoral Reflection Measure–Short Form (SRM-SF), whereas victim empathy was measured with an extended version of the Basic Empathy Scale (BES). No significant associations were found between clinical judgment of moral judgment and the mean scores on the SRM-SF. However, clinical judgment of victim empathy was significantly associated with victim empathy on the Victim Empathy Scale but not consistently in the expected direction. Juvenile sex offenders, who were judged by clinicians to show little victim empathy, displayed lower mean scores on the Victim Empathy Scale than juvenile sex offenders who were evaluated to lack victim empathy or to have intact victim empathy. This study showed unstructured clinical judgment of moral development not to concur with test results. To improve decision-making processes regarding moral development, clinicians are advised to rely on instruments that assess moral development to inform clinical judgment. Further research is needed to examine which predictions are more accurate and to establish the predictive validity of moral development evaluations.
Keywords
Clinical judgments are conclusions drawn by clinicians regarding the client’s health condition (e.g., psychopathology) and/or required treatment, and are based on actual observation of a client combined with clinical experience, theoretical knowledge, and, if available, test results obtained with appropriate assessment instruments. However, in the case of moral development, few valid and reliable assessment instruments are available, and to our knowledge, these instruments are mostly used for empirical research rather than for clinical examination. Therefore, clinicians mostly rely on subjective interpretations of information that they consider important for the examination of moral development.
In general, mechanical predictions, which are statistical, actuarial, or automatic assessments, further referred to as structured assessments, are assumed to be more accurate than clinical judgment (Grove, Zald, Lebow, Snitz, & Nelson, 2000). The predictive validity of mechanical assessments has been shown to be high, which may first of all be ascribed to the use of scientific concepts that have attained empirically established construct validity (e.g., Gibbs, Basinger, Grime, & Snary, 2007). For example, Grove et al. (2000) showed mechanical predictions to account for 10% more accurate predictions of educational and health outcomes than clinical judgments. The lower accuracy of clinical judgment might be explained by its proneness to a variety of possibly occurring biases (Lichtenberg, 2009).
First, there are biases that occur in the interaction with clients, such as mislabeling behavior of clients or patients, due to the fact that their behavior is being compared with clinical populations instead of normal populations (Garb, 1998, 2005; Langer & Abelson, 1974; Rosenhan, 1973). Second, biases that exist are related to (more) static characteristics of clients, such as race and gender biases, resulting in inaccurate assumptions or beliefs about a client’s race or gender affecting the eventual diagnosis (Garb, 1997). Third, clinical judgment can be affected by biases that occur during the collection of information, such as “confirmatory bias,” meaning that clinicians tend to gather information that coincides with already earlier formed beliefs and assumptions. Consequently, important information refuting earlier beliefs and assumptions is left out (Garb, 1998, 2005; Lichtenberg, 1997, 2009). Other cognitive biases that are known to exist are “hindsight bias” and “anchoring or adjustment bias,” which are the tendency to outweigh information that is consistent with a clinician’s final evaluation and the tendency to insufficiently adjust initial judgments when additional information is available, respectively (Lichtenberg, 1997, 2009).
However, assessment instruments are also limited in the sense that they may ignore case-specific information that is important for the development of individual treatment plans (Doyle & Dolan, 2002). This is in particular the case when dynamic personality factors are concerned, such as impulsivity and moral functioning, which change during life and often can only be understood in the context of an individual’s history or by gathering additional information (Groth-Marnat, 2009). In other words, assessment instruments may yield objective information but cannot take into account the uniqueness of each individual case, which limits the validity, reliability, and usefulness of assessment instruments in clinical practice (Hart, Michie, & Cooke, 2007).
In the field of moral development, treatment and referral decisions are generally based on unstructured clinical judgments rather than on structured clinical judgment that is informed by results obtained with assessment instruments (Hendriks, Rutten, Stams, & Brugman, 2006). Thus, it is important to further study possible differences and commonalities between clinical judgment and test results because referral decisions may have lifelong consequences for clients, in particular, in the case of forensic evaluations that concern sentencing decisions, such as length of detention and treatment type and duration.
Two widely researched constructs of moral development are moral judgment and empathy, that is, the reasons or justifications individuals give for decisions or values that pertain to just or benevolent social action (Gibbs, 2010) and the ability to share (affective empathy) and understand (cognitive empathy) emotional states of others (Cohen & Strayer, 1996; Jolliffe & Farrington, 2004, 2006), respectively.
Moral development has been shown to be important for delinquency per se, including first offending, as well as for criminal offense recidivism. Significant associations have been found between moral judgment and (first) offending (Stams et al., 2006) as well as between empathy and offending, with stronger associations for cognitive empathy than for affective empathy (Jolliffe & Farrington, 2004). Moreover, a recent meta-analysis found significant inverse relationships between moral judgment, empathy, shame and guilt, and general offense recidivism in both juvenile and adult offenders. Interestingly, much larger effect sizes were found for the relationship between moral development and general offense recidivism when assessment instruments were used (r = .57) than when evaluations were based on unstructured clinical judgment (r = .10; Van Vugt et al., 2011).
Adequate assessment of moral development is not only important to correctly refer offenders to the appropriate interventions according to their risk level (Andrews & Bonta, 2010) but also to assess improvements in moral functioning of offenders who are enrolled in interventions targeting moral development. This seems, in particular, important as more and more attention is being directed toward the theoretical and empirical foundation of judicial interventions.
The aim of this study is to examine whether unstructured clinical judgment and objective measurement of moral development are associated in a sample of young sex offenders, focusing on moral judgment and victim empathy. It seems important that clinicians can adequately judge moral development in young sex offenders because it has been shown that (juvenile) sex offenders show lower levels of moral judgment when questioned about their victim and lack victim empathy (Fernandez & Marshall, 2003; Knight & Prentky, 1993; Lakey, 1994; Marshall, Hamilton, & Fernandez, 2001; Marshall, Hudson, Jones, & Fernandez, 1995; Van Vugt et al., 2008; Varker & Devilly, 2007). Although we do not yet know whether lower levels of moral judgment and victim empathy are predictive of sexual offense recidivism in juvenile sexual offenders, a meta-analytic study, consisting of adult and juvenile sex offenders, did show lack of victim empathy to contribute to the continuation of violent nonsexual and general offense recidivism (Hanson & Morton-Bourgon, 2004). There is no reason to expect that this relationship would be different for juvenile sex offenders. As juvenile sex offenders largely recidivate to nonsexual offenses instead of sexual offenses (Hendriks, 2006; McCann & Lussier, 2008), it is questionable whether sex offender treatment should primarily target risk factors predicting sexual offense recidivism. From this perspective, moral judgment and empathy may be important targets in treatment that focuses on nonsexual offenses, especially because juvenile sex offenders do not differ from juvenile nonsex offenders where it concerns risk factors for general offense recidivism (Hanson & Bussière, 1998).
Although a significant association between unstructured clinical judgment and independent objective measurement of moral development would support the adequacy of moral judgment, lack of an association would cast doubt on the adequacy of unstructured clinical judgment of moral development. Such lack of association would call for the use of well-validated instruments to assess moral development to inform the clinical judgment of clinicians working with juvenile sex offenders.
Method
Sample
A total of 85 male sex offenders between 13 and 23 years of age, at the time of examination (M = 17.54, SD = 2.22), from three juvenile correctional facilities and six forensic outpatient treatment centers in the Netherlands participated in this study, with a response rate exceeding 90%. The majority of the participants attended special education schools (22.4%) or lower vocational education schools (61.2%), which prepares students for careers in (nonacademic) manual labor jobs. Most participants (77.6%) were native Dutch. We classified the sex offender group according to typologies that are used in clinical practice and scientific research. Most offenders were classified as solo sex offenders (87.1%), as they committed the sexual offense alone. Only 2.4% of the sample was identified as group sex offenders, whereas 7.1% of the sample committed solo and group sex offenses. For 3.4% of the total sample, this information was not available. Approximately 65.9% of the sex offenders were classified as child abusers and 24.7% as peer abusers, with the remaining group (9.4%) treated or sentenced for both offenses or for hands-off offenses (exhibitionism). The offender was familiar to the victim—being a family member, neighbor, or classmate—in 71.8% of the cases. Most offenders (55.3%) had female victims, 20% had male victims, and 24.7% had female and male victims.
A total of 22 clinicians participated in our study, who were working as clinicians either for the juvenile correctional facilities or for the outpatient treatment centers that participated in our study.
Procedure
All respondents declared to voluntarily participate in this study and gave the researcher permission to analyze psychological and criminal records by signing an informed consent. In case the participant had not yet reached the age of 16 years, a parent or guardian was asked to sign a consent. We explained to the respondents that they were allowed to withdraw from the research at any time and that withdrawal did not have any consequences for their treatment or detention situation. Each assessment started with a sociomoral interview (Sociomoral Reflection Measure–Short Form [SRM-SF]), which was recorded on audiotape, transcribed, and scored by the first and fourth author of this article. In the second part of the assessment, the respondent had to answer questions concerning victim empathy that were programmed on laptops. All respondents received a number to preserve anonymity and received a reward of 5 Euros for their cooperation after full participation.
Measures
Moral judgment was measured with the SRM-SF, which is a structured interview that contains 11 questions about a set of core values that are considered important in most societies: contract and truth, affiliation, life, property and law, and legal justice. The answers to these questions were scored for their stage of moral judgment (Gibbs, Basinger, & Fuller, 1992; Gibbs et al., 2007). “How important is it for judges to send people who break the law to jail?” is an example of one of the 11 questions of the SRM-SF. Internal consistency reliability analyses for the SRM-SF questions yielded α = .67.
For the purpose of this and other studies, an extended version of the Basic Empathy Scale (BES) was used to measure victim empathy. The BES is a validated questionnaire examining both cognitive and affective empathy, that is, the cognitive ability to recognize someone else’s emotional state and the affective ability to sympathize with and share the other person’s emotional state (Cohen & Strayer, 1996; Jolliffe & Farrington, 2004, 2006). A validation study of the BES for use in the Netherlands (Van Langen, Stams & Van Vugt, 2009) replicated the positive validation results of the original validation study by Jolliffe and Farrington (2006). The items of the Victim Empathy Scale are responded to on a 5-point Likert-type scale, ranging from 1 = strongly disagree to 5 = strongly agree. Examples of affective and cognitive victim empathy items are “I am concerned about the well-being of my victim” (affective) and “I understand my victim did not like what I did” (cognitive). The internal consistency reliabilities for the own abuse empathy scales were α = .67 (affective) and α = .84 (cognitive). As clinicians were asked to evaluate to what extent the referred juvenile sex offender(s) showed victim empathy, and were not specifically asked to distinguish between affective or cognitive empathy, we examined the strength of the association between the two subscales. We found a high and a significant correlation of r = .70 and therefore merged the two subscales into one subscale, measuring victim empathy.
To measure clinical judgment of moral development in young sex offenders, we used an additional questionnaire (see Appendix) that was completed by clinicians who treated the juvenile delinquents who participated in our study. The clinical judgment questions were derived from the standard psychological and psychiatric assessment used in juvenile court in the Netherlands (pro justitia examination; see also Le Sage, 2006). The following items were used: (a) How would you evaluate the level of moral judgment of the referred youngster? The following response options were given: 1 = little developed, 2 = well developed, and 3 = strongly developed and (b) To what extent does the referred youngster show victim empathy? The following response options were given: 1 = victim empathy is lacking, 2 = victim empathy is slightly present, and 3 = victim empathy is present.
Results
Relationship Between Clinical Judgment and Moral Judgment
We examined to what extent clinical judgment and test results of moral judgment and victim empathy in young sex offenders were related. To examine the relationship between moral judgment measured with the SRM-SF and clinical judgment regarding moral judgment (see Item 1, appendix), we performed a one-way ANOVA to examine differences in mean scores on the SRM for the three clinical judgment evaluations of “little”, “well,” and “strongly developed” moral judgment. No significant differences were found, even when controlled for type of treatment (residential vs. ambulatory), treatment duration, age of the offender, and type of sex offenders (child vs. peer abuser).
Relationship Between Empathy and Clinical Judgment
We also examined the association between victim empathy, measured with the Victim Empathy Scale, and clinical judgment of victim empathy (see Item 2, appendix). A one-way ANOVA was performed to identify differences in mean scores on the victim empathy measure for the three clinical judgment categories, namely, “victim empathy is present,” “victim empathy is slightly present,” and “victim empathy is lacking.” The mean scores on the Victim Empathy Scales significantly differed for the three clinical judgment evaluations, F(2) = 5.48, p < .01, whereas treatment type, treatment duration, age of the offender, and type of sex offenders (child vs. peer abuser) did not affect these results. Post hoc comparisons using the Student–Newman–Keuls test indicated that the mean score for the “victim empathy is slightly present” category (M = 3.96, SD = 1.01) was significantly lower than that of the “victim empathy is lacking” (M = 4.45, SD = 0.49) and “victim empathy is present” (M = 4.59, SD = 0.41) categories. The effect sizes for these differences were Cohen’s d = –.60 and Cohen’s d = –.75, respectively. The “victim empathy is lacking” category did not significantly differ from the “victim empathy is present” category. These results indicate that clinical judgment regarding victim empathy was not associated according to expectation with test results on the Victim Empathy Scale, as juvenile sex offenders who were evaluated to show “little” victim empathy had lower mean scores on the Victim Empathy Scale than juvenile sex offenders who were evaluated to lack victim empathy or who were thought to have intact victim empathy.
Discussion
Altogether, this study showed that unstructured clinical judgment of moral development did not concur with the scores on measures of moral judgment and victim empathy. Although several studies reported structured assessment to outperform unguided clinical judgment in the prediction of recidivism of offenders, even in the case of moral development (Bengtson & Långström, 2007; Craig, Beech, & Browne, 2006; De Vogel, De Ruiter, Hildebrand, Bos, & Van de Ven, 2004; Van Vugt et al., 2011), this study did not examine the predictive validity of either clinical judgment or structured assessment instruments of moral development. Therefore, it cannot be automatically deduced that unguided or unstructured judgment of moral development yields invalid information.
Because clinical judgment can easily be distorted by irrelevant client characteristics and cognitive therapeutic biases (Garb, 1997, 1998, 2005; Langer & Abelson, 1974; Lichtenberg, 1997, 2009; Rosenhan, 1973), it is recommended that clinicians also rely on multiple sources of information and objective assessment instruments when examining dynamic risk factors for delinquency as targets for effective treatment (see also Andrews & Bonta, 2010; Ward, Melser, & Yates, 2007). The use of objective assessment instruments makes clinicians less vulnerable to therapeutic biases and transforms their unstructured clinical judgment (Hendriks et al., 2006), which is informed by nonscientific concepts of morality, into a structured clinical judgment of moral development that is based on well-validated scientific concepts that have empirically been tested (Gibbs et al., 2007). Another advantage of the use of assessment instruments is that these are replicable and controllable.
As moral judgment and (victim) empathy have been shown to play an important role in the continuation of offending (Hanson & Morton-Bourgon, 2004; Van Vugt et al., 2011) and may be important treatment targets for sex offenders, moral judgment and victim empathy should be assessed correctly. According to the “what works” principles of effective interventions (Andrews & Bonta, 2010), interventions should target dynamic risk factors, such as moral development (Van Vugt et al., 2011), as these factors can be modified. In the case of dynamic risk factors, clinical judgment remains important, as clinicians are expected to weigh the seriousness of these factors (e.g., in the case of risk assessment evaluations) and their possible interplay. Therefore, the emphasis should be on the development of valid and reliable instruments that promote “systemization and consistency yet are flexible enough to account for case-specific influences and the contexts in which assessments are conducted” (Doyle & Dolan, 2002, p. 652).
To summarize, clinical judgment remains relevant as it contextualizes structured information based on assessment instruments. Clinicians should attend to the individual circumstances of clients and interpret test results in the context of additional information, their particular observations, and clinical experience (see Ward et al., 2007). For example, if test results are not in line with the clinician’s view regarding the problem of the referred client, he or she should ascertain whether the assessment took place correctly, whether he or she owns complete information concerning the client, and whether additional information is needed to understand the conflicting information.
Some limitations of this study should be mentioned. First, it is possible that our instrument that was used to assess clinical judgment does not adequately capture clinical judgment. However, clinicians were handed questions that they were familiar with from forensic psychological and psychiatric examinations that were conducted in the Netherlands to inform the judge or prosecutor about the offender’s health and personality (see Le Sage, 2006). Second, the assessment of clinical judgment did not include the kind of sources the clinicians used to evaluate the juveniles’ level of moral development. Clinical judgment may have been based on observations and conversations during treatment or on other available information regarding moral development of the client. Possibly, clinicians may have included important information in their evaluations of the young sex offenders’ moral development that is not captured by the instruments. Third, the SRM-SF showed relatively low internal consistency reliability, which could have affected the outcomes. Finally, our findings were based on cross-sectional data, and therefore, it was not possible to establish the predictive validity of either clinical judgment or assessment instruments regarding moral development. Future research should focus on the decision-making processes of clinicians and examine which information regarding moral development is focused on and which is not.
To conclude, there are few instruments available in the Netherlands to assess moral development with adequate assessment instruments, and thus, forensic clinical practice, due to unfamiliarity with the available instruments, is still largely relying on unstructured clinical judgment. Although this study was unable to provide insight into neither the correctness of clinical judgment nor the correctness of structured assessment of moral development, we showed that clinical judgment of moral development was not associated with scores on measures of moral development. This study indicates that clinical judgment and structured assessment are important topics in clinical practice, in particular, in the case of moral development. Especially given the possible long-term consequences of decisions made by clinicians and other professionals in the domain of moral development, it is important to validly and reliably assess moral development of juvenile offenders. More research is needed to understand why clinical judgment and structured assessment of moral development are not associated, which information clinician’s include in their judgments, and how assessment instruments can best be used by clinicians to produce a valid and reliable judgment of moral development.
Footnotes
Appendix
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.