This paper investigates whether the prediction of dangerousness in mentally ill and criminal populations can be considered a scientific enterprise. It begins by distinguishing dangerousness from risk assessment and tracing the historical evolution of how societies have classified dangerous individuals. The paper then critically evaluates the empirical foundations of modern risk assessment instruments, arguing that confounding variables, problems of reproducibility, and the impossibility of isolating independent and dependent variables undermine their scientific validity. Landmark court cases — including Daubert v. Merrell Dow Pharmaceuticals and Kumho Tire Co. v. Carmichael — are examined as legal illustrations of these same empirical shortcomings. The paper concludes that risk assessment should inform, but never solely determine, judicial and clinical decisions about dangerousness.
Dangerousness refers to the likelihood that a mentally ill person or criminal will participate in an act that harms themselves or others. The prediction of dangerousness in mentally ill patients is one of the key challenges facing psychologists today. Psychologists have a responsibility to keep the public safe, yet this must be balanced against the need to avoid drastic measures when they are not warranted. This is often a difficult decision, and psychologists struggle to find the right balance and to devise better methods of assessing risk among their mentally ill patients.
There are many factors that can affect the level of risk in any given situation. These factors represent internal and external variables that may or may not be easily controlled. The existence of these variables makes it difficult to make an accurate assessment of dangerousness. This paper supports the thesis that the prediction of dangerousness is not a science.
The first task is to clarify what dangerousness means and how it differs from risk assessment. Dangerousness is typically used to describe those who have committed violent crimes or who have mental illnesses that create the potential for violence. Risk assessment is the process of determining the likelihood that a person will commit a violent act. No reliable roadmap exists that can lead to an accurate prediction of whether a person will commit a violent crime in the future. Many factors influence an individual's decisions, and circumstances can change in ways that affect outcomes. Nevertheless, psychologists, social workers, criminologists, and other professionals are routinely asked to assess the dangerousness of a person and evaluate their risk to society, so that proper actions can be taken that protect both the rights of the individual and the safety of the public.
One of the more controversial outgrowths of this issue is legislation in Germany and the EU calling for the preventative detention of offenders convicted of certain crimes (Dunkel & van Zyl Smit, 2004). This law primarily involves sex offenders. The premise of the law is not the main source of contention; rather, it is the execution of the law that draws the most criticism. The legislation provides for mandatory detention of certain offenders even when no provision for detention existed in the original sentence (Dunkel & van Zyl Smit, 2004). It also allows for the extension of sentences if an offender's development in prison suggests that they may commit further serious offenses (Dunkel & van Zyl Smit, 2004).
The issues that surround this legislation mirror those that arise in the determination of dangerousness more broadly. The legislation makes sweeping assumptions and presumes guilt rather than evaluating individual risk levels. The key criticism is that it is based on generalizations unsupported by significant empirical evidence. The law is reactionary in nature, and individual rights dissolve into those generalizations — one of its most fundamental flaws.
One of the key requirements of a civilized society is devising a means to maintain order and enforce its rules. In a utopian society, there would be no need for a criminal justice system, but no such society exists. Humanity is therefore continually confronted with the problem of how to prevent those who commit crimes from repeating their actions and disrupting social order (Brown & Pratt, 2000). Our current conception of dangerousness emerged from changes in the social arrangements of modern society. In the past, class, gender, and social structure played a significant role in the assessment of dangerousness (Brown & Pratt, 2000). The power structure of earlier societies was weighted toward the upper classes, who held ultimate control over crime, punishment, and the classification of dangerous individuals. Risk assessment was often based on a person's station in life rather than on individual circumstances or characteristics, and the poorer classes were almost always presumed to be the more dangerous elements of society.
Modern risk assessment has moved away from this kind of categorical, class-based approach (Brown & Pratt, 2000). The methods used today are the product of the interplay between knowledge, values, and evolving class and gender relations. Social status no longer plays a formal role in classifying an individual's dangerousness.
Nevertheless, modern risk assessment still takes a categorical approach (Brown & Pratt, 2000). The difference is that today's categories are grounded in a long history of academic research on repeat offending and violent or criminal behavior rather than in social hierarchy. While the criteria for category membership have changed, the underlying categorical logic remains the same — and that logic carries its own significant flaws, which are examined in the sections that follow.
"Empirical critique of risk assessment instruments"
"Landmark cases exposing legal limits of risk assessment"
Kumho Tire Co. v. Carmichael (1999) is one of the most significant risk assessment cases in American legal history. In that case, a tire blowout caused a crash in which one passenger was killed and several others were severely injured. The plaintiff sued, alleging a defective tire. The Court ruled, drawing on the foundation established in Daubert, that causality could not be established because too many other variables — including the driver's own reactions — could have contributed to the accident. The dangerousness of the tire itself could not be isolated and proven.
These court cases illustrate the fundamental fallibility of risk assessment and the methods used to determine dangerousness. In the field of abnormal psychology, the stakes are even higher: an overly harsh assessment of dangerousness can cost a person their freedom, while an underestimate can place the public at serious risk.
The examination of current methods and applications of risk assessment presented in this paper supports the thesis that the prediction of dangerousness is not a science. Current approaches are found to be lacking in key elements of valid empirical research methodology — most critically, the ability to isolate variables and establish causality in the presence of pervasive confounding factors. Risk assessment instruments represent a meaningful improvement over historically class-based judgments of dangerousness, but they remain far from scientifically definitive. Therefore, risk assessment should be treated as one component of a broader evaluative process. Clinical and judicial decisions about dangerousness should never rest on risk assessment and prediction alone.
You’re 52% through this paper. Sign up to read the remaining 2 sections.
Sign Up Now — Instant Access Already a member? Log inAlways verify citation format against your institution’s current style guide requirements.