Overview of available assessment instrumentsThis review of available risk assessment instruments has evaluated them against five different criteria. These five different criteria can be identified by the question(s) they ask regarding a particular instrument:
Applied
to six different assessment instruments, these five questions result in 30
different outcomes. Table 1 summarizes whether the instruments reviewed comply with the ethical or testing standard corresponding to each question. As seen in Table 1, there are only three instances of compliance, or partial compliance.
Table
1.
|
APA 2.02 refers to questions of sensitivity, specificity, false positives, and false negatives. | |
EPT 1.1 refers to questions of validity. | |
EPT 2.8 refers to questions of inter-rater reliability. | |
EPT 5.1 refers to manual availability. | |
APA 2.04 refers to questions of peer review. | |
UCJ refers to unaided clinical judgment. | |
GCRA refers to guided clinical risk assessments. | |
SVR-20 refers to the Sexual Violence Risk-20. | |
RRASOR refers to the Rapid Risk Assessment for Sexual Offender Recidivism. | |
MnSOST refers to the Minnesota Sexual Offender Screening Tool. | |
SORAG refers to the Sexual Offender Risk Appraisal Guide. | |
C corresponds to compliant. | |
NC corresponds to noncompliant. | |
PC corresponds to partially compliant. |
|
UCJ |
GCRA |
SVR-20 |
RRASOR |
MnSOST |
SORAG |
APA 2.02 |
NC |
NC |
NC |
NC |
PC |
PC |
EPT 1.1 |
NC |
NC |
NC |
NC |
NC |
NC |
EPT 2.8 |
NC |
NC |
NC |
NC |
NC |
NC |
EPT 5.1 |
NC |
NC |
C |
NC |
NC |
NC |
APA 2.04 |
NC |
NC |
NC |
NC |
NC |
NC |
[Page
128]
Beyond
those three outcomes, the six instruments reviewed fail to comply with the
remaining ethical or testing standards.
Given
the extent to which these instruments fall short of various practice standards,
their use creates difficult ethical problems for the psychologists who rely on
them.
Ethical Standard 2.04(b) states:
"Psychologists recognize limits to the certainty with which diagnoses, judgments, or predictions can be made about individuals"
(American Psychological Association, 1992, p. 1603).
This
standard obligates psychologists to recognize the profound limits
limiting the accuracy with which they can identify recidivism risk in cases of
previously convicted sexual offenders.
Relatedly, Ethical Standard 7 ,04(b) states:
"Whenever necessary to avoid misleading, psychologists acknowledge the limits of their data or conclusions"
(American Psychological Association, 1992, p. 1610).
This
standard obligates psychologists to acknowledge that, given the current level of
knowledge regarding risk assessments, the instruments typically used in sexual
predator evaluations cannot claim general acceptance by the relevant scientific
or professional community. General acceptance of these instruments necessitates
that they comply with appropriate ethical and practice standards. Quite clearly,
however, those instruments fail to comply with the relevant standards by a margin
of nine to one.
In
1996, Grisso and Tomkins expressed guarded optimism regarding the progress being
made in predicting future dangerousness.
When properly translated, the results of the new generation of violence risk studies might soon provide mental health professionals with a more reliable scientific foundation for describing a person's violence risk, thereby assisting society in deciding when those risks are sufficient to take action to protect the person and others.
We have not yet achieved this capacity (p. 929) (emphasis added).
Unfortunately, it appears that this verdict currently remains as appropriate as it was in 1996. For assessing the recidivism risk of previously convicted sexual offenders, UCJ and GCRAs do not possess sufficient evidentiary reliability to support expert testimony.
Ongoing research related to developing actuarial instruments for risk assessment appears promising. Nonetheless, there still remains considerable work to be done before psychologists can use these instruments to support their testimony in a legal proceeding.
Without meeting relevant ethical and practice standards, these procedures cannot claim general acceptance per Frye. In addition to falling short of general acceptance, the rate of error associated with these procedures typically remains unknown; and, consequently, they also fail to satisfy Daubert criteria.