Recently, the Secretary of the Department of Homeland Security, Jeh Johnson, stated his greatest fear was the threat of “lone wolf” terror attacks. He further clarified that self-radicalized militants are much more difficult to detect for law enforcement and intelligence agencies, versus the bona fide terrorist group member (i.e., ISIS, Al Qaeda, etc.).
As a society, and even within forensic and threat management circles, we remain fixated on Prediction – trying to determine with certainty who will act violently, when they will act, and what they will do. We have spent billions of dollars, extensive human capital, and much of our individual liberty in the pursuit of monitoring foreigners and our own citizenry towards predicting the next attack. Yet we continue to see attacks unfold both here and abroad, by both affiliated terrorists and so called lone-wolves: Reutlingen,
Ansbach, Munich Germany; Nice & Paris France; Orlando, Fl; San Bernardino, CA. This is not an exhaustive list by any means, only focusing on attacks motivated by radical islamist ideology, just within the past year, and just within the U.S. and Western Europe. There are numerous other terror attacks by other groups and ideologies. In response we as a collective have focused heavily on Prediction, and less on Prevention.
The difference is important. The truth of the matter is we cannot predict behavior in general very well, and certainly cannot predict violence very well. A few things contribute to this difficulty.
First, the base rate for violence of any kind is so relatively low in the general population, that it becomes very difficult to predict with any degree of certainty. Accurate prediction requires statistical probability modeling, which in turn requires that you analyze a phenomena that occurs with some regularity. We can predict how often “violence” of different types occurs in the general population, but not specifically who will be violent, when, and in what context.
Second, forensic psychology and sociology research has greatly informed us about violence risk factors, and this research has continued to evolve in improve our understanding over the past 15-20 years. Yet, there is still much we don’t know about why one particular person moves on a trajectory of violence, and another with similar risk factors, features, and personality does not. Frankly, the science is just not there yet. As much as some experts would like to profess, this is not Minority Report and we don’t have pre-cogs working for us.
Third, the developmental concepts of multi-finality and equi-finality impact our accuracy. Multi-finality is when two people with similar genetic loading and/or environmental upbringing show markedly different achievement later in life (i.e., Bill and Roger Clinton). Equi-finality is when two people with very different genetic loading and/or environmental upbringing end up achieving the same later in life (i.e., Bill Clinton and Donald Trump). Multi-finality and Equi-finality occur all the time, and are impacted by a myriad of experiences, events, and influences for which we cannot fully account.
Fourth, violence of any kind is multi-dimensional, dynamic, and transactional in nature. Many factors contribute to one’s violent trajectory, that pathway changes over time accelerated and decelerated by environmental events, and it occurs within the integration between attacker and victim(s). Prediction assumes a static conclusion of what will or won’t happen in the future, yet violence does not play by those rules, it is not a static occurring phenomenon.
Prevention does not require Prediction.
In short, we cannot predict violence at any meaningful level for the individual. Perhaps one day in the distant future, but the state of the science is not there currently. The answer is to instead focus on prevention, and this is not a foreign idea to any of us.
The forensic psychologist Reid Meloy uses a great example to explain prevention without prediction. Consider a cardiologist with a 1,200 patient caseload. The cardiologist cannot “predict” with any certainty which of his patients will have a heart attack in the next month or year. Indeed, we see cases occur all the time that seemingly “break the odds” – the morbidly obese 55yo male patient who smokes a pack a day and drinks a 6-pack of beer daily lives for 15 more years; while the 55yo male patient who eats healthy, runs marathons, does yoga, and meditates unexpectedly drops dead of a major heart attack within 2 months. The cardiologist cannot predict with precision which patient will have the heart attack and who will not, and when the heart attack will occur. Yet, that inherent limitation is no reason to negate trying to prevent the heart attack based on what we do know about the risk factors.
We can still prevent by knowing the risk factors, screening and assessing for those, applying evidenced-based interventions to treat and mitigate those risk factors, and adjust accordingly over time based on whether the risk factors increase or decrease. We treat the risk factors, and thus decrease the risk of heart attack in the aggregate, and in the majority of cases at the individual level as well. It is not a precise science as the above example illustrates, but it is one that is far better than not trying to manage the ris
k factors at all.
THAT is Prevention without prediction, and that is what is involved in evidence-based threat management as well:
- Understand the risk factors for violence
- Clarify the kind of violence you are trying to mitigate (Affective vs. Predatory, to be addressed in a future post)
- Screen and assess for the known risk factors
- Treat (intervene) to mitigate the risk
- Monitor the impact of the intervention, and adjust accordingly
We can and should continue to improve our ability to eventually, one day, predict who will be violent versus those who will not, but that seems a long way off given the complexity involved and the limitations of our current science. Rather, we must continue to use risk estimates based on evidenced-based risk factors, and apply those in a manner so we can intercept a trajectory of violence upstream before violence occurs.