FORUM on Corrections Research

Correctional classification and the "responsivity principle"

It is no secret that the most effective correctional programs use correctional classification systems to conform to three principles of effective intervention: the risk principle, the needs principle and the responsivity principle.² The risk principle maintains that we should classify offenders according to risk of reoffending and then use our most intensive interventions and supervision models for the medium to high risk offender. The needs principle directs service providers to assess for individual traits that are related to future offending (such as criminogenic attitudes and values, criminogenic peers, etc.) and then focus our interventions on those characteristics rather than on traits such as self-esteem, which are not related to future offending. Finally, the responsivity principle holds that even when we meet the first two principles, we need to consider offenders' ability to participate in the programs that fit their level of risk and criminogenic need. The now popular cognitive skills programs, for example, are much less successful with offenders of below-average intellectual abilities than they are with those of average to above average intelligence.³ High anxiety inmates do not always respond to treatment strategies that involve confrontation.⁴ Similarly, offenders seem to benefit from being matched to supervising case managers on the basis of compatible personality traits.⁵

As these three principles take hold in correctional service delivery and research arenas, we see most efforts devoted to the first two principles -- risk and needs. The field of correctional classification has also witnessed increasing developments in the assessment of risk and of specific criminogenic needs. Evidence of this trend is apparent in current research journals as the overwhelming majority of classification articles address the assessment of risk and criminogenic need. The risk assessment technology continues to advance in terms of its predictive accuracy as well as in its ability to predict specific types of offences, including violence, sex offending and family abuse.⁶

But while these commendable advances promote the development of classification systems such as the Level of Supervision Inventory,⁷ the Hare Psychopathy Checklist⁸ and the Wisconsin Probation/Parole Classification System⁹ as well as specific assessments of criminogenic needs, the costs of ignoring responsivity should not be understated.

Actually, classification for purposes of responsivity or differential treatment enjoys a long tradition in corrections which predates the current generation of risk assessment models. At the core of this tradition are psychological and personality-based typologies that classify offenders according to cognitive complexity (for example, Conceptual Level,¹⁰ or Interpersonal Maturity Level¹¹) or criminal personality types (such as the Megargee MMPI-Based Typology,¹² Interpersonal Maturity Level,¹³ Client Management Classification¹⁴) or the Quay Behavioral Classification Systems.¹⁵ These systems were more widely used from the 1960s to the mid-1980s than they are today. This is probably because correctional priorities have shifted toward a more retributive focus and away from correctional treatment and intervention. The Client Management Classification (CMC) system, however, continues to be widely used.

The advantages of using a systematic process to classify according to responsivity

These earlier applications of the psychological classification systems worked successfully with the notion of differential treatment or matching.¹⁶ Matching clients to interventions, living units and case managers on the basis of their psychological classification harnessed the responsivity principle in a number of ways.

• Many case managers and client supervisors were not trained clinicians. The personality-based classification systems offered them a systematic, consistent approach for understanding important differences among clients.
• These differences identified personality traits such as low or high cognitive functioning, anxiety, accountability for criminal behaviour, criminal values and beliefs, and impulsivity. Many of these traits could inform decisions regarding intervention, supervision style and client-case manager interaction style. Specific treatment recommendations were incorporated into each system.¹⁷
• Evaluation findings could be informed by responsivity considerations. Indeed, in our current neglect of responsivity, we routinely "mask" the treatment effect.¹⁸ Over and over, we hear of programs that "failed" when in fact they probably succeeded with certain types of offenders and failed with others. For the group as a whole, our successes were cancelled by our failures.

Despite these advantages, psychological classification systems were considered to be geared too much toward rehabilitation to be of any use to later correctional practices which favored incapacitation and punishment. They were also accused of being too time consuming and cumbersome to use.

Psychological classification models: How efficient are they?

As Table 1 shows, psychological classification systems differ according to the number of psychological types they provide for and the method of administration required to arrive at a classification/assessment. As we compare these methods to the more common risk assessment instruments (such as the Statistical Index of Recidivism [SIR] Scale,¹⁹ the Salient Factor Score [SFS]²⁰ or the Wisconsin Probation/Parole Classification System²¹), it is easy to understand why psychological classification systems were sometimes viewed as too difficult to use. Recent research, however, offers some promise for simplification.

Table 1

An overview of psychological classification systems: Number of types and assessment procedures
System	Construct	No. of Types	Administration	Scoring
Interpersonal Maturity (1-Level) (Warren, 1983)	personality and cognitive development	13	interview	clinical
Jesness Inventory 1-Level (Jesness and Wedge, 1983)	personality and cognitive development	9	test	acturial
Conceptual Level (Harvey et al., 1961)	cognitive development	3	sentence	clinical
Quay Behavioural Categories (Quay and Parsons) 1972; Quay, 1983)	personality	5	observational checklist	acturial
Megaree MMPI-Based Typology (Megargee and Bohn, 1979)	personality	10	test	acturaial/ clinical
Client Management Classification (Lerner, et al; 1986)	personality	5	interview	clinical
Note: approximately 30 % of the MMPI results must be assigned clinically to personality types to break ties or resolve other discrepancies

Two sets of findings from a recent comparative assessment of several of the systems listed in Table 1 speak to this problem.²² The first found that the numerous types of psychological classification systems can be distilled down to four types central to adult male inmates.

1. Committed criminal: an offender who is comfortable with the criminal label, associates primarily with criminal peers and harbours beliefs, values and attitudes supportive of crime. Typically the criminal career is lengthy.
2. Situational: an offender whose criminal behaviour is the result of a recent life crisis, but who nevertheless shows a prosocial value system, positive self-concept and stable relationships. The criminal careers of these offenders are generally not extensive.
3. Neurotic anxious: an offender whose criminal behaviour represents the acting-out of a crisis or frustration. These offenders are highly anxious, have problems trusting others and possess dysfunctional strategies for managing anxiety. In prisons, they experience far more stress than other inmates.
4. Character disordered: an offender who does not openly admit to anxiety, but appears to be acting out anxiety in fairly impulsive, irresponsible manners. The offender's anxiety is highly defended, and some may try to charm and manipulate.

The second set of findings asserts that there is no need to avoid the systems which are more easily administered. The Jesness Inventory which involves administering 155 true/false items, for example, was just as valid and useful as the more complicated method for assessing I-level, which involves a much longer interview and assessment process.

Do responsivity-based personality systems predict correctional adjustment and recidivism?

Psychological classification systems are often described as classifying offenders according to responsivity rather than to risk of recidivism. Indeed, a recent critique of psychological classification research faults it for failing to conduct adequate recidivism studies. Even though it is not clear that the developers of the psychological systems intended them to predict recidivism, risk assessment models have developed strong track records for categorizing offenders according to their likelihood of reoffending -- both in validation studies and in more recent revalidation research.²³

In contrast, the few studies examining the relationship between the psychological typologies and recidivism have shown equivocal results. Megargee and Bohn's assessment of the relationship between the Megargee MMPI-Based Typology types and reoffending, for example, failed to identify any strong relationships.²⁴ A recent assessment of the CMC among probationers, on the other hand, noted that it was more accurate than a traditional risk assessment classification system in predicting future revocations, rearrests and absconding incidents.²⁵ In addition, with adult males, the psychological systems offer numerous predictions of prison adjustment,²⁶ and, in one study, the psychological types predicted prison adjustment measures just as well as more traditional risk-based predictors of adjustment.²⁷ Finally, it is well known that in the few instances where differential treatment and responsivity recommendations of the systems were used, there was an impact on offender recidivism.²⁸ Still, while the relevance of the psychological types to recidivism is not disputed, more research is clearly needed.

At this point, we can only suggest some future research directions. For example, there appears to be a relationship between criminogenic attitudes and behaviours (noted risk factors), and the personality types identified above. These results emerged in the course of constructing the systems, but are validated in more recent research with two research populations -- adult male inmates incarcerated in a minimum security federal penitentiary, and adult male inmates incarcerated in a maximum security federal penitentiary, both in the United States.²⁹ The psychological classifications were derived from the Jesness Inventory, I-Level Classification System.³⁰ As shown in Table 2, characteristics such as impulsivity, criminal associates, history of criminal activity, lack of empathy, irresponsibility, comfort with the criminal label, lack of insight and failure to judge criminal behaviour as wrong, characterized one type -- the committed criminal (and sometimes character-disordered inmates) but not the other two -- neurotic anxious and situational inmates.

Table 2

Relationship (Gamma) between personality types and criminal orientations and behavioural patterns
Rating Item *	Committed	Criminal	Character	Disorder	Neurotic	Anxious	Situational
	Pen	camp	Pen	camp	Pen	camp	Pen	camp
Gang activityt	0.54***	0.49*	-0.39	0.49	-0.03*	-1.00	-1.00***	-0.35
Ever violentt	0.50***	0.24*	0.06	0.64***	-0.37**	-0.25*	-0.46**	-0.59***
Shows concern for otherst	0.43**	NV	-0.04	NV	-0.33	NV	-0.56**	NV
Shows empathy for otherst	0.39**	0.42**	0.15	0.41	-0.66***	-0.21	-0.24	-1.00***
Self-image (1 = favorable 0= not favorable)	0.50**	0.22	-0.28	-0.76**	-0.68***	-0.05	-0.15	0.09
Locus of control (1=external 0=internal)	0.30**	0.34**	-0.48**	-0.08	0.11	-0.21	-0.33*	-0.26
Responsiblet	0.32**	0.17	-0.22	0.04	0.16	-0.29	-0.28	0.00
Aware of the consequencest	-0.09	-0.41**	0.14	0.10	-0.19	-0.40**	0.23	-0.36
Aware of others' needst	0.29**	0.47***	0.14	-0.14	-0.38**	-0.20	-0.22	-0.46***
Aware of others' expectationst	0.26	0.59***	0.51**	0.26	-0.74***	-0.70***	-0.42*	-0.55**
Impulsivityt	0.24*	0.44**	0.16	0.01	-0.34*	-0.46**	-0.23	-0.22
Insight into own problemst	0.23*	0.25**	-0.25	0.03	-0.12	-0.17	-0.07	-0.19
Comfort with criminal labelt	0.08	0.26**	0.41**	0.15	-0.20	-0.22	-0.22	-0.23
Career criminalt	0.45***	0.61***	0.09	0.12	-0.54***	-0.57***	-0.35*	-0.64***
Crime acts out of crisist	-0.48***	-0.22**	0.19	0.16	0.00	0.12	0.64***	-0.10
Crime for conformity with criminal peerst	0.61***	0.61***	-0.24	-0.46*	-0.83***	-0.51	-0.23	-0.32
Everyone is out for #1t	0.20	0.28*	0.13	0.20	-0.21	-0.05	-0.24	-0.59***
Negativity judges behaviourt	0.33**	0.24*	-0.33*	0.10	-0.36**	0.00	0.04	0.45***
Introspectivet	0.27**	0.28**	-0.26	-0.40*	-0.11	0.12	-0.17	-0.35**
Notes: t 1 = No; 0 = Yes * p<0.10 ** p<0.05 ***p<0.01 NV: Limited variability *Items were obtained from raters' assessments of inmate interviews. Interrater reliability was greater than 70% for all items used in this table. Pen=penitentiary Camp=minimum secuity prison camp

One needs only to look toward a lengthy history of criminological research to understand the importance of this finding. Such traits are extremely important dynamic predictors of future offending.³¹ But because we lack the recidivism data and additional evaluations of programs that fully use the psychological classification systems, we can only suggest the following.

Traditional dynamic risk factors may interact with personality types in their relationship with recidivism.

Traditional dynamic criminogenic risk factors may characterize only one or possibly two types of offenders as identified by the psychological systems. At the same time, traits such as positive self-image, external locus of control, dysfunctional responses to crises, while not associated with committed criminals, may be associated with the recidivism of inmates classified as other types (for example, neurotic anxious and situational).

There may be important treatment implications. Just as criminogenic needs vary across risk categories (for example, low risk inmates do not possess the most important dynamic risk factors), they may also vary across the psychological types, to the extent that the responsivity systems may also direct us to a differential identification of criminogenic needs.

Conclusion

Even when we have classified offenders according to risk and criminogenic need and targeted our interventions to key criminogenic needs, important considerations remain -- human traits that affect an offender's capability to respond to the approaches of our program.³² Clearly, factors such as intelligence, anxiety, cognitive maturity, attention deficit disorder and learning style will translate into treatment amenability or an offender's likelihood of achieving success in our program. Many of these traits are embedded in the types identified by psychological classification systems.

Although previous research demonstrates the value of psychological systems, these systems are not used to their full potential in current correctional practice. Indeed, the entire notion of responsivity is often ignored in offender programming endeavours. In citing some of the most recent research findings pertinent to the psychological systems, this article attempts to lend new support to the value of these classification systems.

---

1. Division of Criminal Justice, P.O. Box 210389, University of Cincinnati, Cincinnati, Ohio, 45221-0389.

2. D. Andrews, J. Bonta and R. Hoge, "Classification for Effective Rehabilitation: Rediscovering Psychology," Criminal Justice and Behavior, 17, 1 (1990): 19-52. See also D. Andrews and J. Bonta, The Psychology of Criminal Conduct (Cincinnati: Anderson Press, 1994).

3. E. Fabiano, F. Porporino and D. Robinson, Corrections Today, August (1991): 102-108.

4. M. Warren, "Application of Interpersonal Maturity Theory to Offender Populations," in Personality Theory, Moral Development, and Criminal Behavior, W. Laufer and J. Day (eds.) (Lexington, Massachusetts: Lexington Books, 1983).

5. Andrews, Bonta and Hoge, "Classification for Effective Rehabilitation: Rediscovering Psychology." See also T. Palmer, Individualized Intervention with Young Multiple Offenders: The California Community Treatment Project (Hampton, Connecticut: Garland Press, forthcoming).

6. P. Van Voorhis and K. Brown, Risk Classification in the 1990s (Washington: National Institute of Corrections, 1996).

7. D. Andrews and J. Bonta, The Level of Service Inventory - Revised (LSI-R) (Toronto: Multi-Health Systems, 1995).

8. R. Hare, The Psychopathy Checklist - Revised (Toronto: Multi-Health Systems, 1991).

9. C. Baird, R. Heinz and B. Bemus, The Wisconsin Case Classification/Staff Deployment Project, Project Report No. 14 (Madison, Wisconsin: Department of Health and Social Services, Division of Corrections, 1979). See also C. Baird, R. Prestine and B. Klockziem, Revalidation of the Wisconsin Probation/Parole Classification System (Madison, Wisconsin: National Institute of Crime & Delinquency, 1989).

10. O. Harvey, D. Hunt and H. Schroder, Conceptual Systems and Personality Organization (New York: John Wiley, 1961). See also M. Reitsma-Street and A. Leschied, "The Conceptual Level Matching Model in Corrections," Criminal Justice and Behavior, 15, 1 (1988): 92-108.

11. C. Jesness and R. Wedge, Classifying Offenders: The Jesness Inventory Classification System (Sacramento, California: Youth Authority, 1983). See also Warren, "Application of Interpersonal Maturity Theory to Offender Populations."

12. E. Megargee and M. Bohn, Classifying Criminal Offenders: A New System Based on the MMPI (Beverly Hills, California: Sage Publications, 1979).

13. Jesness and Wedge, Classifying Offenders. See also Warren, "Application of Interpersonal Maturity Theory to Offender Populations."

14. K. Lerner, G. Arling and C. Baird, "Client Management Classification Strategies for Case Supervision," Crime & Delinquency, 32 (1986): 254-271.

15. H. Quay, Technical Manual for the Behavioral Classification System for Adult Offenders (Washington: Department of Justice, 1983). See also H. Quay and L. Parsons, The Differential Behavioral Classification of the Juvenile Offender (Washington: Department of Justice, 1972).

16. T. Palmer, "The Youth Authority's Community Treatment Project," Federal Probation, 38, 1 (974): 3-14. See also Palmer, Individualized Intervention with Young Multiple Offenders. See also Reitsma-Street and Leschied, "The Conceptual Level Matching Model in Corrections." See also Warren, "Application of Interpersonal Maturity Theory to Offender Populations."

17. A summary of recommendations culled from previous research is available from the author.

18. P. Van Voorhis, "Correctional Effectiveness: The High Cost of Ignoring Success," Federal Probation, 51, 1 (1987): 56-62.

19. J. Nuffield, Parole Decision-making in Canada: Research towards Decision Guidelines (Ottawa: Solicitor General of Canada, 1982).

20. D. Gottfredson, C. Cosgrove, L. Wilkins, J. Wallerstein and L. Rauh, Classification for Parole Decision Policy (Washington: National Institute of Law Enforcement and Criminal Justice, 1978). See also P. Hoffman, "Screening for Risk: A Revised Salient Factor Score (SFS-81)," Journal of Criminal Justice, 11 (1983): 539-547.

21. Baird, Prestine and Klockziem, Revalidation of the Wisconsin Probation/Parole Classification System.

22. P. Van Voorhis, Psychological Classification of the Adult Male, Prison Inmate (Albany, New York: SUNY Press, 1994).

23. Andrews and Bonta, The Psychology of Criminal Conduct. See also Baird, Prestine and Klockziem, Revalidation of the Wisconsin Probation/Parole Classification System. And see J. Bonta, W. Harmon, R. Hann and R. Cormier, "The Prediction of Recidivism among Federally Sentenced Offenders: A Re-validation of the SIR Scale," Canadian Journal of Criminology, 79 (1996): 61-79. And see P. Hoffman, "Twenty Years of Operational Use of a Risk Prediction Instrument: The United States Parole Commission's Salient Factor Score," Journal of Criminal Justice, 22, 6 (1994): 477-494.

24. Megargee and Bohn, Classifying Criminal Offenders.

25. P. Harris, "Client Management Classification and Prediction of Probation Outcome," Crime and Delinquency, 40 (1994): 154-174. It is noteworthy that the rearrest rate for this sample was low (10.9% for a 17-month follow-up period), suggesting a rather low risk sample.

26. Megargee and Bohn, Classifying Criminal Offenders. See also Van Voorhis, Psychological Classification of the Adult Male, Prison Inmate.

27. P. Van Voorhis, "Psychological Determinants of the Prison Experience," The Prison Journal, 73, 1 (1993): 72-102.

28. Palmer, "The Youth Authority's Community Treatment Project." See also Palmer, Individualized Intervention with Young Multiple Offenders. And see Warren, "Application of Interpersonal Maturity Theory to Offender Population."

29. Van Voorhis, Psychological Classification of the Adult Male, Prison Inmate.

30. Jesness and Wedge, Classifying Offenders: The Jesness Inventory Classification System.

31. Andrews and Bonta, The Psychology of Criminal Conduct. See also A. Caspi, T. Moffitt, P. Silva, M. Stouthamer-Loeber, R. Krueger and P. Schmutte, "Are Some People Crime Prone? Replications of the Personality-Crime Relationship Across Countries, Genders, Races and Methods," Criminology, 32 (1994): 163-195. And see P. Gendreau, T. Little and C. Goggin, "A Meta-analysis of the Predictors of Adult Offender Recidivism: What Works!" Criminology (in press). And see S. Glueck and E. Glueck, Unraveling Juvenile Delinquency (Cambridge: Harvard University Press, 1950).

32. Andrews, Bonta and Hoge, "Classification for Effective Rehabilitation."