MOSAIC Threat Assessment Systems

Articleissues
advert = September 2008
POV = December 2007

MOSAIC Threat Assessment Systems is an American company that devises and administers methods for assessing and screening threats to public figures. It was originated in 1980 by Gavin de Becker & Associates, it has gained wide use by hundreds of law enforcement agencies, universities, schools, and large corporations.

Walt Risler of Indiana University was brought in to assist in development. Robert Martin, the founding commander of the Los Angeles Police Department Threat Management Unit, now a Managing Principal with Gavin de Becker & Associates, heads up the MOSAIC Threat Assessment division.

MOSAIC Background

By the early 1980s, America had learned that one person with some nerve and a lack of concern for others could disrupt the democratic process. Yet comparatively little had been done to learn about that person, particularly considering his (and sometimes her) impact on society. As an example, in the last thirty-five years, more public figures have been attacked in America than in the 185 years before that.

With all the obvious need to better understand violence, our society still knew more about why someone bought a particular brand of shaving lotion than about why he bought a gun. Back then, "threat assessment" was often little more than a matter of someone’s opinion. There was no method, no common language, no pool of experts, little sharing of experience, and almost nothing published on the topic.

When the first MOSAIC methodology emerged almost twenty-five years ago, it sparked a new exploration into the pre-incident indicators of violence. Today's computer-assisted MOSAIC systems have been in wide use by agencies responsible for the protection of high officials in the U.S. Federal Government, and by those protecting many governors, mayors, and judges.

MOSAIC methodology has been made available for several types of high stakes assessments. For example, the MOSAIC for the Assessment of Student Threats (MAST) was developed to help school officials and police manage threats made by students and assess situations most likely to escalate to violence.

What is MOSAIC?

Contrary to a popular misconception, MOSAIC is not a computer program. Rather, it is a method – part of which is a computer-assisted. MOSAIC is an assessment strategy that helps ensure fairness, consistency, and thoroughness in high-stakes matters. It is a tool that helps guide assessors of risk in the criminal justice system, but it is never used to help determine guilt or innocence of crime suspects.

MOSAIC is a way of breaking down a situation to its elements, then organizing and identifying the most important factors. Once a case is thus coded, it can be instantly compared to others where the outcome is known. The case can also be weighed against the opinions of experts in the relevant field. MOSAIC suggests to an assessor those questions determined most valuable to the overall assessment.

All MOSAIC programs ask a series of questions that assessors can answer by drawing on a range of possible answers. This range is the heart of the method, because when questions are asked on a form, giving only YES/NO, or check/no-check answers, it is contrary to the best predictive strategies, contrary to the way natural intuition makes important assessments. Imagine, for example, being asked for your opinion on a movie you just saw, but being required to answer using one of only two options: “Best movie I ever saw,” or “Worst movie I ever saw.”

Clearly, your best assessment could not be gained with that restriction, yet many so-called risk assessment instruments apply precisely that approach. With the Domestic Violence MOSAIC, for example, a police detective asks a battered woman a series of questions. As an element such as the batterer’s alcohol abuse is explored, the detective will have a range of answers to draw upon. More than just checking off “Yes, he drinks,” the detective can select any of several levels of alcohol use, from none at all, to occasional without adverse results, to frequent with adverse results (such as memory loss, using drinking as an excuse for abuse, cruelty when intoxicated, etc.)

When all of the questions are answered, the police detective will have considered more than a 150 variables (because when one answer is chosen, it also means that three or more other possible answers to a given question were not chosen). This allows an apples-to-apples comparison of the instant case to other cases in a department’s experience (or a detective’s experience) where the outcome is known. It also allows a variable by variable comparison to expert opinion.

Historically, assessors’ personal perspectives colored assessments, and just as significant, colored the expression of assessments. As is with MAST, one school official might describe a given student threat situation by calling the student “dangerous,” or calling the case “serious.” Of course, they are all serious, but the school official means “likely to escalate.” Another school official might call the exact same case ”pretty bad,” while still another might say, “typical.” Is that typical as in just like other cases that resulted in death, or typical as in just like other cases that wasted that school official’s time? The point is that even if assessments are uniform and equally valid, even if all school officials were equally experienced and explored the same topics, the ways people express the results are all over the map.

Imagine a doctor trying to diagnose arthritis in an elbow. Ten doctors might look at the same elbow and describe the extent of swelling using any of several terms: extreme, moderate, noticeable, minor, major, etc. It’s the same elbow, but we can’t compare it’s degree of swelling to any other elbow unless we know what a particular doctor means by his of her use of the words.

MOSAIC has provided agencies a shared language, so that words mean the same thing to all involved. This alone -without regard to the quality of assessments by various individuals- has brought substantial improvement to case management.

The MOSAIC method uses artificial intuition to help guide high-stakes assessments. It involves distilled opinions of experts, who suggest, in effect, a series of questions about each case, then offer a range of possible answers.

Artificial Intuition, like human intuition, draws on information from many sources, but unlike human intuition (which can be difficult to express), MOSAIC places information into an organized framework.

Though each assessor brings his or her own intuition and experience, MOSAIC assures that different assessors approach their cases from a shared foundation.

How Does MOSAIC Work?

MOSAIC is not a computer program, it is an evaluation method that is computer-assisted. MOSAIC prompts the user to address the questions most likely to produce a quality evaluation. Once the questions are answered, the case broken down to its elements, and the most important factors identified, it can instantly be compared to other cases where the outcome is known. The case can also be weighed against the opinions of experts in many relevant fields. Each case is rated on a scale of one to ten, with ten representing the cases most similar to those that have escalated to violence.

Imagine that a student has made a threat that alarms others, and it is your responsibility to evaluate the situation and the student. In a perfect world you’d be able to confer with all the leading experts in threat assessment, law enforcement, psychology, and the behavioral sciences. You might ask:

*What are the most important things I need to learn about this situation?
*What information about the student can most help my evaluation?
*How can I organize the information to weigh it most effectively?
*What factors and warning signs are most relevant to the student’s future behavior?
*How can I effectively express and document my conclusions?

MAST provides the knowledge and guidance of the leading experts, presented in a step-by-step format that is simple to use.

MOSAIC produces an automatic report that documents and presents exactly what questions were asked, how they were answered, and what comments the user chose to add along the way. Both the rating and the MOSAIC process help to inform the school administrator’s evaluation of the situation.

Myths about MOSAIC

There are some people who think that threat assessment, or MOSAIC, is some sort of profiling. The reality is, however, that threat assessment is the exact opposite of profiling. Profiles are hypothetical (e.g., “An unidentified serial killer will turn out to be a male Caucasian in his mid-thirties who likes fast cars”). MOSAIC is never hypothetical. It always concerns actual, identified individuals.

Profiles are often collections of demographic factors. MOSAIC is never concerned with such things as race, appearance, socioeconomic level, or gender. MOSAIC explores only behavior and circumstance.

Profiles are often applied to large groups (e.g., an airline develops a “hijacker profile” and then observes all passengers to see who matches). MOSAIC is never applied to groups, only to those individuals who come to the attention of officials because of self-identifying behaviors, such as making a threat.

MOSAIC seeks to recognize risk in situations, not to identify so-called “dangerous” people.

Controversy

The "Mosaic 2000" pilot profiling program, which began late in 1999 in some 20 U.S. schools in response to school shootings, has drawn some of the heaviest, and most public, criticism. Kirk, Michael (January 18, 2000). [http://www.pbs.org/wgbh/pages/frontline/shows/kinkel/profile/ "Profiling" School Shooters.] "PBS Frontline"]

A 1999 "Seattle Times" editorial stated, "One of the more ludicrous and inevitable solutions for preventing violence in schools is a computer program that tabulates risk. . . random violence in schools has fostered a desperation for predictability and control." Editorial (December 20, 1999). [http://archives.seattletimes.nwsource.com/cgi-bin/texis.cgi/web/vortex/display?slug=A19991221083405&date=19991220&query=violence+in+schools Real Live Adults Better Than 'Artificial Intuition.'] "Seattle Times"] A 1999 "Sacramento Bee" editorial also noted:

There's something terribly wrong with the Mosaic-2000 mind-set: that if we can just find the right software program and feed it some surface characteristics of teens, it will assemble for us the three-dimensional understanding of them that so eludes us. Software is no substitute for the real conversations that need to take place, day to day, in American high schools.. . Profiling is a fine technique for FBI manhunts; it is misplaced in American schools." (November 19, 1999) The War on Teens: Profiling Tactics Have No Place in American Schools. "Sacramento Bee"]

Conservative commentators also raised concerns: "Indeed, applying to children profiling techniques normally reserved for criminals, without any evidence that such methods can successfully predict "future" criminals, is a large step to take in a free society."Thomas, Virginia L. (November 10, 1999). [http://www.heritage.org/Press/Commentary/ed111099.cfm Mosaic 2000: An Educational Dragnet.] Heritage Foundation]

In an article on the method, psychologist Hill Walker of the Institute on Violence and Destructive Behavior told "Wired", "There are some serious validity issues here, some reputation-ruining implications." Forrest, Brett (June 2000). [http://www.wired.com/wired/archive/8.06/mustread.html?pg=5 UltraViolencePredictor 1.0.] "Wired"] Responding to claims that the program amounts to profiling, de Becker replied, "Mosaic-2000 is the opposite of profiling in that it is always applied to an actual known individual, and it always explores actual behavior and circumstance. Mosaic-2000 does not explore age, appearance, race, ethnicity, socioeconomic level, or any other demographic feature; profiling almost always does."de Becker, Gavin (September 2000). [http://www.wired.com/wired/archive/8.09/rants.html Rants & Raves.] "Wired"]

Professor Laurence Steinberg also questioned the need and use of the software for predicting violence:

In the late 1990's, the number of school-age children who died from homicide averaged around 2,500 a year. But fewer than half of 1 percent of them were killed in or around schools. Let's say, for argument's sake, that each of these incidents involved a student perpetrator. In a nation of 90,000 schools, trying to pick out the dozen or so students a year who might commit murder is like looking for a needle in a haystack the size of Kansas.Steinberg, Laurence (April 22, 2000). Software Can't Make School Safe. "New York Times] "]

De Becker responded, "It is not predictive and doesn't claim to be scientific."de Becker, Gavin (May 5, 2000). Threats in School. "New York Times] "]

References

External links

* [http://www.mosaicsystem.com/ MOSAIC Threat Assessment Systems website]