In the context of offender risk assessment, risk communication is the process whereby the results of an assessment are conveyed in words, numbers, and/or pictures to decisions makers. The study of risk communication concerns the effect that the communication method, the characteristics of the individual receiving the communication, or the decision-making context has on people’s decisions. Forensic risk communication is a relatively new area of forensic psychology arising from the growth in offender risk assessment instruments since the 1990s. Its importance lies in its being the fulcrum of risk assessment’s influence over forensic decision-making and hence over the expression of the risk principle within the Risk-Need- Responsivity model of correctional intervention.
This article describes research into the communication formats and individual characteristics that influence risk perception and related decision-making, drawing on lessons learned from medical risk communication research, then focusing on violence risk communication. It reviews the effects of how numbers are presented, the role of individual numeracy, and the communication of risk in terms of probabilities or risk categories. This article considers what existing research suggests for best practice in offender risk communication.
The Importance of Forensic Risk Communication
In 1984, researchers Vladimir J. Konečni and Ebbe B. Ebbesen published a critical account of judges’ decision-making in criminal and mental health cases, describing them as “almost 100% rubberstamping” the recommendations by psychiatrists, prosecutors, and probation officers. This pattern, in which both criminal courts and mental health review boards rely heavily on the risk assessment results and recommendations communicated by assessors, was replicated several times over the next three decades. Studies showing that advice provided by clinicians or other assessors is a substantial influence on forensic decision-making raise the importance of not only assessing risk accurately but also understanding how to communicate risk effectively.
Lessons From Medical Risk Communication
The evidence base for forensic risk communication draws key lessons from medical risk communication. For example, information that is framed in terms of the possible risks of inaction (negative framing) is effective for persuading people to assess their health, but framing information in terms of the benefits of treatment (positive framing) is more effective for persuading people to choose treatment. Numeracy helps medical patients understand their risks better, and more numerate individuals are more likely to remember and use mathematical reasoning in their decision-making. However, even educated professionals struggle with numeric information.
Research into the communication of medical and other risks has found that the effects of risk illiteracy can be alleviated to some degree by communicating risks with the aid of graphs. Graphs can also help communicate the normalcy of a distribution and the unusualness of extreme outcomes. If used, graphs need to be purposely selected for their ability to communicate the gist of the relevant risk information and for their suitability to the purpose (e.g., identifying absolute risk of an unhealthy behavior vs. conveying the relative risks of treatment options). Any verbal description of the graph should be carefully written and tested to ensure that it is congruent with the graph and does not introduce the potential for misinterpretation.
The Framing of Risk
As found in medical research, how the risk is framed can influence violence risk perception. In forensic risk communication, risks are usually framed negatively, in terms of the likelihood of recidivism. Some assessments have been developed in which the items are phrased so that a positive score indicates lower likelihood of recidivism, drawing attention to the outcome of not offending, that is, positive framing. Research indicates that positive framing leads to more decisions to discharge offenders from custody than does negative framing. For instance, one study revealed that a risk described as a 26% chance of violence occurring was perceived by mock judges to be more dangerous than a 74% chance of no violence occurring, even though both statistics refer to the same probability of events. Graphs have been considered helpful in communicating such risks, partly because they can show simultaneously the probability of violence occurring and the probability of it not occurring, as in icon arrays.
Some research finds that risks also seem higher if they are communicated using frequencies (i.e., the number of recidivists per 100 or 1,000 offenders) rather than probabilities or percentages (e.g., .40 or 40% likelihood of recidivism). Again, graphs might be beneficial to the extent that they can be used to show information in frequencies and percentages at the same time. Different graphs might be employed to illustrate alternative framing or varying scales, which can also influence how risk is perceived, and each should be carefully designed and evaluated for its intended purpose. Also, the numeric information in graphs can still be difficult for less numerate individuals to understand.
Numeracy
There is evidence that numeracy affects how risks are perceived and used in forensic cases. In general, people tend to overestimate the risk of violent offending and ignore how common recidivism is in the whole forensic population (i.e., the base rate). Offender risk can be perceived as higher if case information that is unrelated to risk is communicated at the same time as the risk assessment, and there is evidence that decision makers in the corrections system often override the results of structured risk assessment using information that is unrelated to risk. More numerate individuals appear to be less swayed by such peripheral information. Less numerate individuals are more likely to be influenced by anecdotal and affective information rather than numeric information provided in forensic evidence.
As well as numeracy, experience with risk assessment information and attitudes toward social science and evidence-based practice has been associated with how risk information is understood and used in decisions. Graphs can increase people’s attention to the probability of harm and for that reason have the potential to be especially helpful for communicating the risk of violent behavior, but research evaluating the use of graphs in forensic risk communication remains very limited.
Numerical and Categorical Communication of Risk
Offender risk assessment is generally conceived of as prognostic rather than diagnostic, and the results of risk assessment make a statement about the likelihood of offending in the future. As a result, risk may be expressed as a probability. Actuarial risk assessment instruments typically yield an empirically derived probability of recidivism associated with categories of risk, although these categories may be determined through a rational process. For risk communication, advantages of using actuarial data include the precision and reliability of assessment results. Effective communication of actuarial risk can aid the allocation of risk management resources to individual offenders according to their risk of offending, which supports the efficiency and effectiveness of the correctional system. However, some discussion remains in the literature as to which statistics are the most important to communicate (e.g., relative risk vs. absolute probability of recidivism). Furthermore, assessors and decision makers are not always comfortable with the use of statistics.
Risk assessments that take a structured professional judgment approach typically result in a nonnumerical risk category such as low risk or high risk. A practical advantage to the use of categorical risk communication is that clinicians are more accustomed to a diagnostic than a probabilistic approach and prefer using categories. However, as with medical risk communication, people do not interpret these risk categories consistently. Variation in the interpretation of risk is a problem for the practice of risk communication because it is difficult to agree on a disposition or the allocation of correctional resources if different assessors and decision makers each have a different idea of what an offender’s risk category means. Reformulations of structured professional judgment tools in the mid-2010s emphasized case formulation rather than categorization, aiming to integrate the analysis of risk factors with suggestions for risk management. Initial research examining the communication of risk management recommendations to law enforcement, social service, and correctional agencies suggests that most recipients carried out at least some of the recommendations, especially those affirming court-ordered conditions.
Recommendations for Practice
In their foundational article, Vladimir J. Konecni and Ebbe B. Ebbesen conducted some of the first research into risk communication in real-life settings, and they looked forward to changes in forensic decision-making that allowed for a greater role of analytic data and feedback provided to the decision makers. By 30 years later, risk assessment data were widely available, and there was growing evidence (experimental and naturalistic) that forensic decisions are made in accordance with the risk principle. However, studies suggest that clinicians, jurors, and other decision makers still do not necessarily view offender assessment, or key contributors to risk such as psychopathy, as a particularly important piece of information. To be effective, risk communication must help people overcome any existing misunderstandings (including a tendency to overestimate risk) and redirect their attention to relevant and useful information.
Effective risk communication helps people make a meaningful connection between risk information and decision-making. A standard for evaluating the effectiveness of forensic risk communication has been proposed that measures the distinction made in risk management decisions between offenders with measurably different risk of recidivism. Experimental research employing this standard has reported that providing numerical information can improve risk communication. Recommendations emerging from this line of research include explicitly tying any categorical risk communication to nonarbitrary numerical risk estimates.
The use of graphs has been recommended to facilitate the communication of offender risk assessment. Although there is limited empirical evidence pertaining to graphs in forensic decision-making, initial research suggests that a simple histogram of the proportion of offenders who recidivate as a function of assessed risk category can improve risk communication. Graphs used for this purpose should be selected with attention to negative and positive framing of information, and the benefits of presenting information in multiple formats should be evaluated.
Evidence that forensic decision makers rely heavily on information provided by professionals who assess risk points to the need for these experts to develop expertise in effective risk communication. Training in risk assessment techniques has already been successful; in the future, training in how to evaluate and explain offender risk assessments and their results might also be valuable. Assessors serving as expert witnesses must learn to help a nonscientific audience to overcome misconceptions about risk and understand how statistical interpretations can be applied to decisions about individual offenders.
An emerging area of offender risk assessment is analytic procedures that demonstrate the ability to measure individual offenders’ risk and its change after treatment. In the future, the practice of risk communication might also need to develop methods to communicate the extent and limits of change in risk.
References:
- Hilton, N. Z., Scurich, N., & Helmus, L. M. (2015). Communicating the risk of violent and offending behavior: Review and introduction to this special issue. Behavioral Sciences & the Law, 33, 1–18. doi:10.1002/bsl.2160
- Konečni, V., & Ebbesen, E. G. (1984). The mythology of legal decision making. International Journal of Law and Psychiatry, 7, 5–18.
- Scurich, N., Monahan, J., & John, R. S. (2012). Innumeracy and unpacking: Bridging the nomothetic/ idiographic divide in violence risk assessment. Law and Human Behavior. Advance online publication. doi:10.1037/h0093994