How Scarcity Affects the Working Poor

All too often, the working poor are unfairly put down as miscreants, reprobates, and degenerates that impede our otherwise thriving Western society. The poor are first marginalized as defective, inadequate and flawed consumers, and then stigmatized for not participating in socially relevant consumer practices. Recent global economic downturns serve as a key source to a greater diversity in the poverty experience. Research suggests that these plights have precipitated the advent of the nouveaux pauvres (middle-class consumers whose social and cultural capital has decreased) and the working poor (consumers that work yet fail to pull above the poverty line or make ends meet) (Hamilton, Piacentini, Banister, et al., 2014).

This article aims to dissuade readers from typecasting the working poor as apathetic and incapable. Specifically, insights from behavioral economics are used to explain why the poor are not poor simply by virtue of their bad decisions. Instead, it is suggested that people make bad decisions because they are poor. Together, capitalistic structures at the macro-level and impaired decision-making at the micro-level render the working poor’s consumer behavior all the more faulty and unstable.

Macro-Level Influences

At the macro-level, marketing makes the poor even poorer because aggressive campaigns target the poor with a range of products — chiefly cigarettes, alcohol, fast-food, lotteries, pawn shops, casinos, predatory mortgages, fringe-banking schemes, payday loans, rent-to-own and high-interest credit cards. These products are patently crippling to low-income consumers’ health and wealth. Still and all, market forces never cease to ‘slay the slain’. Unlike the poor, the rich have systems such as attractive ‘no-fees’ options, automatic deposits, and reminders that are designed to shelter higher-income consumers from errors. Overall, much less is done by marketers to aggressively campaign positive options such as healthier diets, non-profit services, union banks, and prime-rate lenders to low-income consumers (Bertrand, Mullainathan & Shafir, 2006).  

The credit card industry serves as an illustrative example to show the inequity on how low-income as compared to higher-income consumers are treated. A recent report by FRONTLINE and The New York Times cites some shady tactics that the industry uses to incite consumers to take on more debt. A set of code words within the industry’s vernacular follows the insidious theme. Consumers who pay off the full balance on-time are referred to as ‘deadbeats’, whereas low-income consumers that carry monthly debt are called ‘revolvers’. Deadbeats are lost causes, whereas revolvers are cash cows. Revenues are generated by tactics that include hidden fees, default terms, penalty fees and higher rates that are often triggered by marginal errors (e.g., payment that arrives only an hour late; a charge that exceeds the limit by a few dollars) (Bertrand, et al., 2006).

Micro-Level Influences

Clearly, the poor become poorer because of manipulative and exploitative marketing tactics. A deeper level of analysis reveals that the working poor are also more likely to fall for these tricks because of impaired decision making perpetuated by the state of poverty itself.

According to Dr. Sendhil Mullainathan, a Professor of Economics at Harvard University, the poor necessarily require higher levels of self-control and restraint. In his book, Scarcity, Dr. Mullainathan explains that scarcity of financial resources affects the poor as they cannot afford to waste a dime never less shell out wads of cash to splurge on non-essential wants. The working poor are constantly trying to stretch their dollar so they can scrape by and fit the bare necessities in their tight budgets.

Dr. Mullainathan explains that in these circumstances, there is a profound psychological dynamic at work called the bandwidth tax. The single-mindedness and tunnelled vision caused by scarcity leads to reduced mental functioning in both fluid intelligence (solving problems and reasoning logically), and executive control (planning and controlling impulses).


The AI Governance Challenge

We can all think of an instance when increased cognitive load inhibited problem-solving. How many of you have struggled to find your car keys while running late to work in the morning only to realize they were in your pocket the whole time? In this simple example, we understand that the scarcity of time leads to an increase in bandwidth tax, which then impairs the capacity to search the most obvious place wherein which the keys might be.

For the working poor, budgeting or planning errors can occur even after meticulous tracking because their minds are constantly ruminating about financial struggles. For example, a high-level of bandwidth tax may lead to more forgetfulness causing missed or late payments, which then lead to further penalty fees. Also, the working poor may experience reduced ability to solve unexpected problems like a bounced cheque, or a failed transaction. The outcome — they dig themselves into a deeper hole. In this way, scarcity not only raises the cost of mistakes, it also provides more opportunity to make mistakes by reducing mental capacity. As a result of this constant state of anxiety, unease, and impaired mental functioning, the working poor may become more vulnerable to shady payday loans, fringe-banking schemes, and predatory mortgages.

Science Denial Isn’t Only A Conservative Problem

Do you believe in climate change, and that humans have contributed to it? If you lean left, chances are the answer is yes. Scientists have often counted on liberals and Democrats to support their political causes, including climate legislation, stem-cell research, and the teaching of evolution in schools.

Yet it may surprise many liberals to recognize that science denial isn’t purely a symptom of the political right.

For instance, much of the debate over genetically modified organisms (GMOs) is between scientists and liberals. Approximately the same amount of scientists believe that climate change is mostly due to human activity (87%)* as believe that genetically modified foods are safe to eat (88%), yet scientists face an uphill battle in convincing their usual political allies about the science of GMOs.

Why such a shift from the traditional supporters of scientifically informed policy? Part of the reason, as I wrote in my last post, is motivated reasoning. The same ideological narratives – such as “environmental protection” and “keeping corporations in check” – may lead Democrats to believe the science on climate change but reject the science on GMOs. We prefer cohesive and identity-affirming stories more than complicated and nuanced truths, so we sometimes dismiss legitimate data and arguments when they don’t support our previous beliefs.

A related factor that fuels science denial is an increasing distrust in experts and public figures. Trust in many American institutions is at or near all-time lows. While this distrust can be justified, all too often it contributes to us believing we know more than we actually do. In attempting to protect ourselves from misinformation, our distrust can lead us to ignore important information and become entrenched in incorrect beliefs.

Can We Correct Misperceptions?

One of the reasons people rally against GMOs is that they believe modifying an organism’s DNA is unprecedented and unethical. While the naturalistic fallacy certainly plays into people’s confusion on this issue, the truth is that we have been genetically engineering our crops for thousands of years. For example, foods like corn wouldn’t exist had our ancestors not engaged in genetic engineering, however unconsciously. While modern methods are undoubtedly more advanced and can be seen as controversial, the core process is something which has remained unchanged for millennia.

Although we might hope that spreading the facts will increase consensus, the reality is that addressing partisans’ false beliefs often backfires. Correcting factual misperceptions on political issues can fail to convince those who were misinformed and sometimes influences people to harden in their incorrect beliefs. Thus, scientists may not be able to persuade Democrats or Republicans out of their misconceptions.

So what can we do to change people’s minds? It seems that correcting others may be much less effective than getting people to confront their own lack of understanding.

Illusion of Explanatory Depth

To test this idea, Fernbach et al. (2013) asked participants to indicate their levels of understanding and support for these six policy propositions:

(a) Imposing unilateral sanctions on Iran for its nuclear program

(b) Raising the retirement age for Social Security

(c) Transitioning to a single-payer health care system

(d) Establishing a cap-and-trade system for carbon emissions

(e) Instituting a national flat tax

(f) Implementing merit-based pay for teachers

Then, some participants were randomly assigned to give precise step-by-step explanations of how two of these policies would be implemented and affect change. After writing their explanations, participants were asked to rerate their understanding and preference of the two policies and indicate how certain they felt about their positions.

The results showed that, after attempting to explain policies, participants reported decreased confidence in their understanding and more moderate positions on the issues they evaluated. Compared to participants who explained why they supported a policy, those who explained how a policy worked were less extreme and certain in their final positions.

Confronting their own ignorance, and breaking what the authors called the illusion of explanatory depth, humbled the participants and made them more open to other perspectives. The authors suggest that trying to explain how policies work made people feel uncertain about how much they understood the topic, so those participants expressed less certainty and extremity in their views. Alternatively, those who were asked to explain their reasons for support were not led to question their understanding of the topics, so their certainty didn’t shift because they still believed they knew enough to have a confident opinion.

Balancing Our Judgements with Intellectual Humility

Studies like Fernbach et al. (2013) illustrate the need for us to cultivate intellectual humility to recalibrate our overconfident policy evaluations. Intellectual humility has been defined as “having insight about the limits of one’s knowledge, marked by openness to new ideas; and…the ability to present one’s ideas in a non-offensive manner and receive contrary ides without taking offense,” (Davis et al., 2014). Like the participants who realized that they couldn’t adequately explain the policies they were writing about, we need to accept that we may not always possess the information necessary to be definitively confident in our opinions.

Recognizing our limitations requires us to trust others if we want to obtain knowledge, and preliminary data from Davis et al. (2014) suggests that intellectual humility is indeed related to trust. Psychological scientists are just beginning to study how this trait impacts one’s worldview and decision-making, but it seems clear that when we trust others, especially those with greater knowledge than ourselves, we have a better opportunity to reach valid conclusions about the state of the world.


The AI Governance Challenge

That being said, cultivating intellectual humility doesn’t require us to abandon our skepticism altogether. Experienced clinicians, lawyers, and political analysts are all susceptible to errors and biases, and experts in all professions deserve our scrutiny. These individuals surely possess important knowledge that most of us do not, but they can be biased and overconfident in their assessments like the rest of us. Maintaining skepticism allows us to hold others to the same standards of intellectual humility that we should strive for ourselves.

So while GMO skeptics should accept the current evidence that GMOs are safe, there’s no reason why they can’t also demand increased oversight and data collection until more long-term data are collected. This moderate stance, which acknowledges the concerns of both GMO skeptics and supporters, seems to be the best way forward on this issue.

Experts, including scientists, deserve our skepticism, but we are not likely to remedy any of our complex policy problems unless we also appreciate their perspective and insights. And, if partisans develop more intellectual humility, we will hopefully see greater acceptance of facts on both sides of the aisle.

*It is important to note that 97% of climate scientists, compared to 88% of scientists at large, agree that climate change is mostly due to human activity. This difference should remind us that experts also forget about the limits of their expertise.

Is it too late for Trump and Clinton to become more likeable?

This article originally appeared in [] and belongs to the creators.

According to the old adage, one never gets a second chance to make a first impression. Might that hold true for the presidential candidates?

There would seem to be plenty of opportunities between now and Election Day – including nonstop coverage of the horse race, policy statements, debates and live campaign events – for the candidates to share their views and values and for voters to analyze them in order to make an informed choice.

But past research suggests that one of the most influential predictors of how we vote might have already been determined: our first impressions. According to psychological studies, voters’ initial impressions of candidates are an important determinant of how they evaluate them and eventually cast their ballots.

For example, social psychologist Alex Todorov from Princeton University examined voting data in elections for the U.S. Congress and found that whether a candidate’s facial features conjured up a sense of competence predicted approximately 70 percent of voting behavior. (This finding has been replicated with French, Japanese and Bulgarianpoliticians.)

Findings like these and others imply that what may matter most to voters isn’t the candidates’ voting record or what they say during the campaign but simply what face they were born with.

Given this influence of first impressions on voting behavior, is it too late for a candidate with a high unfavorability rating – notably, both Donald Trump and Hillary Clinton – to change voters’ minds?

Implicit versus explicit

Although there is no strict definition of what exactly counts as a first impression, it is generally understood to be our first sense of a person as good or bad, which in turn can be based on a variety of sources of information.

In addition to facial characteristics, these could include a person’s group membership (for example, one’s gender) as well as the kinds of behaviors (mannerisms, humor, arrogance, etc.) the person displays when we first encounter him or her (for example, how Bernie Sanders behaved during the first national Democratic presidential debate).

Cognitive and social psychologists have argued that there are two types of impressions: explicit and implicit. Whereas our explicit impressions are those that we can easily feel and report to others (e.g., “I really love Hillary!”), our implicit impressions are those that we may or may not be conscious of.

Psychologists measure implicit impressions by administering tests in which subjects unknowingly and unintentionally reveal their preferences. For example, instead of asking voters how much they like Hillary Clinton on a scale from 1 to 10, an implicit measure would assess how much voters display positivity or negativity toward photos of the former secretary of state presented rapidly on a computer screen.

The respondents on this kind of measure do not realize that their impressions of Clinton are even being assessed. (Ideally, this kind of measuring eliminates the self-editing that subjects do when they describe their own impressions, and reveals preferences and biases that might not otherwise be identified.)

Implicit impressions, as measured by psychologists, have been shown to uniquely predict people’s decisions and behaviors, including voting behavior. So when it comes to choosing a candidate, our conscious thinking and feeling about a candidate might not be the only thing that drives our behavior and choice. Our implicit impressions also seem to play an important role.

Minds made up

Many findings to date have shown that implicit impressions are especially difficult to change once they have been formed.

For example, in one series of experiments, participants learned about two novel groups, one of which was described as evil and the other as benevolent. The researchers then measured participants’ implicit first impressions of the groups. As expected, participants implicitly evaluated the “evil” group as negative and the “benevolent” group as positive.

Then the researchers tried to get the participants to revise their first impressions – to change their implicit mind about the two groups. They told participants that they had accidentally mixed up the information about the groups and that the one that had been described as evil was actually good and the one described as good was actually evil.

The researchers also tried to give participants a long and detailed story about how the groups eventually changed their moral character (e.g., the formerly evil group eventually started to regret their behavior and tried to make amends, etc.). The participants’ explicit impressions changed in line with the new information. No matter what the researchers tried, however, the subjects’ implicit impressions did not budge. These findings have been interpreted to mean that we might not be able to move beyond our implicit first impressions of others, even if we say differently.

So should we give up on reading policy papers from the candidates? Or listening to the debates? Can we sit back and assume that our (implicit) mind is made up?

A little wiggle room

Recent work in our research lab suggests that the answer is no.

We demonstrated that implicit first impressions in fact can be changed, even immediately and in a lasting manner. Research I conducted with Williams College social psychologist Jeremy Cone and Cornell University graduate student Thomas Mann showed that learning even just one new and highly diagnostic piece of information about someone can swiftly lead to corrected implicit impressions.

Instead of examining attitudes toward groups, as some prior research did, we focused on individuals, and on how a single, extremely diagnostic piece of information might influence first impressions of them. Whereas people may not believe that entire groups can shift their moral character, they do seem to believe that individuals can sometimes behave in surprising ways and that an extreme behavior reflects something meaningful about the person and should be used to update first impressions.

In some of our studies. All participants formed strongly positive implicit first impressions of Bob after learning about these behaviors.

But then participants learned about one additional behavior that was very negative and rare and therefore extremely diagnostic of Bob (for example, he had been convicted of mutilating a small animal). Contrary to the predominant view that implicit evaluations are almost impossible to update on the basis of a single new piece of evidence, participants instantly moved from a strongly positive implicit view of Bob to a strongly negative one.

Across six experiments with over 1,250 participants, we found replicable and strong evidence that people rapidly reversed their implicit first impressions after learning a single piece of new information.

So perhaps we can move beyond an initial positive implicit impression to a negative one, but can we move from an initial negative implicit impression to a positive one?

Set in stone?

Lots of psychological research has found that negative (vs. positive) information has a greater influence on our judgments of others.

Learning that someone who has been characterized as hostile did something heroic just does not have the same weight as learning that someone characterized as nice did something horrific. As today’s campaign managers likely know, it takes only one horrific gaffe to trigger a fall from grace, but an extraordinary amount of laudatory behavior to redeem oneself.

This is true of our implicit impressions as well. Although our recent work shows that we can update our implicit positive first impressions to negative ones, it is harder to move people’s impressions in the other direction.

Is there ever a case in which initial negative implicit impressions can be rapidly and instantly undone? What about when the initial negative evidence is discovered to be unfounded?

False accusations and innuendos typically run rampant in any election season. The National Enquirer recently published a story alleging that Ted Cruz had extramarital affairs with five women, an accusation that Cruz has strongly denied and that hasn’t been confirmed by any other publication or source. How might this story affect voters across the country who may be just starting to pay attention to Cruz? If it turns out that the accusation is indeed false, will those voters nevertheless remain negative toward Cruz because of it, perhaps unconsciously?


The AI Governance Challenge

Our lab recently showed that participants were able to correct even an extremely negative implicit first impression of a target to a strong positive one if they later learned that the person’s initial negative behaviors had been misinterpreted. For example, if someone is told Bob broke into a house – creating a negative impression – and then later informed that the reason was that there was a fire and he was saving his neighbor’s children, it is likely they will change their implicit negative impression into a strong positive one.

This new work shows that earlier studies on changing implicit first impressions were not the final word on this issue, and that – in some cases – the damage from a false accusation can be undone.

Never too late

Yes, first impressions can be powerful, shaping the ways in which we look for, interpret and believe information we subsequently encounter.

A person’s facial features and first behaviors can influence how we interpret their later actions. And, in the absence of any strong evidence to the contrary, a candidate’s electoral success may well be predictable by her or his face.

But our lab’s recent work shows that these first impressions are not unmovable. When we learn something new (and reliable) about someone that is strongly discrepant with our first impressions, we can change even our implicit mind about them.

Although you’ve already formed your first impression of the candidates, there are many months to go in the campaign season. As our research shows, you could still change your mind by Election Day.

Can We Design An Environment That Facilitates Honesty?

I’m writing this from the Environmental Neuroscience Laboratory at the University of Chicago. Here, we study how the physical environment affects affective and cognitive processes, as well as complex human behaviors. Our overarching goal is to quantify the relationships between the physical environment and these psychological phenomena. In this piece, I’ll write about one of the main streams of research I have been working on that has implications for the design of salubrious and cooperative environments (Kotabe, Kardan, & Berman, 2016).

Can we design an environment that causes people to behave more honestly?

First, we are working on a project to quantify the perception of ‘disorder’ in an environment, and furthermore, its effect on a complex human behavior—rule-breaking. Rule-breaking is particularly of interest here because there is an extraordinarily influential social science theory called “broken windows theory” (Wilson & Kelling, 1982) which posits that disorderly environments give rise to disorderly (rule-breaking) behaviors.

This was demonstrated in a series of field experiments reported in the journal Science (Keizer, Lindenberg, & Steg, 2008). Explanations for this theory all have one thing in common—they assume that these effects occur because people in disorderly environments are engaging in some complex social reasoning about, for example, the presence of police, the behavioral norms in the neighborhood, or the prevalence of poverty (social disorder).

Visual disorder and the broken window phenomena

Recent research on visual perception suggested to us that it may be possible that visual disorder could also play a role in broken windows phenomena. We define “visual disorder” as the perception of disorder that can be attributed to basic visual features of a scene such as simple spatial features (e.g., straight lines, symmetry) and simple color features (e.g., hue, saturation, brightness).

In a series of experiments, we carefully examined people’s disorder ratings for hundreds of images of environmental scenes as well as stimuli derived from these scenes that were stripped of scene-level semantics. Through these experiments, we were able to start quantifying visual disorder, and thus we were able to carefully manipulate visual disorder in a second set of experiments in which we tested whether exposure to visual disorder alone, absent of scene-level social cues (see Figure 1), would encourage dishonest behavior.


Figure 1. (a) One of the visually ordered stimuli and (b) one of the visually disordered stimuli used to manipulate visual disorder in the cheating experiments.

In these experiments, we had participants first do a challenging math test. We incentivized performance by telling them that they would be paid for every answer they got right. After the test, they were randomly assigned to be exposed to either visually disordered or visually ordered stimuli for five minutes. Immediately after exposure, they proceeded to the next part of the study in which they were asked to grade themselves on the math test, knowing that they could make extra money for every answer they say they got right … even if they didn’t really get it right.

The results confirmed our hypothesis. Those exposed to visual disorder for just five minutes were, relative to those exposed to visual order, up to 35% more likely to cheat than those exposed to visual order, and the amount they cheated by was up to 87% larger. Even we were skeptical of such robust effects, so we conducted a replication experiment in which we used an even more conservative procedure, and the general pattern of results was the same.

The practical relevance of visual disorder research

The results of this study and our ongoing research have practical implications. What if we start to seriously consider basic visual disorder cues when we design office and school environments? If we could reduce cheating behavior even by just a little, the societal effects could be dramatic. For example, imagine if the amount by which people underreported their taxes decreased by just 1%—billions of dollars would be saved.

Of course, there is a lot of work left to do. We have established an effect of one kind of visual disorder stimuli on cheating behavior in one task. Whether such effects generalize to other categories of stimuli and other activities in which one can break rules are open questions. The theoretical mechanisms we posit would suggest that effects of various visually disordered stimuli would generalize to various rule-breaking behaviors.


The AI Governance Challenge

For example, research on visual perception suggests that processing visually disordered stimuli is more cognitively taxing than processing visually ordered stimuli. Cognitive fatigue is known to have effects across a wide variety of behavioral domains. Furthermore, visual disorder cues may carry semantic information related to linguistic metaphors such as in “he’s bending the rules” and “she’s following the straight path”.

Such metaphors may indicate that the concept of ‘disorder’ is intimately linked to spatial thinking, which could tie perceive disorder to a wide range of complex behaviors. Our future work, and hopefully the work of others inspired by our research, will start to shed light on such inquiries into the order of disorder.

Why Decision Science Matters

This article originally appeared in [] and belongs to the creators.

Nearly everyone is familiar with the DirecTV commercials in which actor Rob Lowe and NFL quarterbacks Eli Manning and Tony Romo appear as unflattering alter-egos of themselves: “peaked in high school” Rob Lowe, “bad comedian” Eli Manning, “artsy craftsy” Tony Romo, etc.

The adoption of data science in companies like Uber, Netflix and Amazon evokes a similarly striking contrast with legacy companies such as Yellow Cab, Blockbuster and Sears/Kmart.

Data science vs decision science

Data science is often used in conjunction with many other science-related terms — algorithms, machine learning, artificial intelligence and predictive analytics. All of these terms are intended to indicate when computers are being used to detect signals or patterns in data that drive better business outcomes.

While data science is perhaps the most broadly used term, “decision science” seems like the more fitting description of how machines are assisting business leaders in solving problems that have traditionally relied on human judgment, intuition and experience.

It may not be the sexiest phrase in the world — I’ve never seen it in any marketing materials — but “decision science” aptly encapsulates how computers are helping to systematically identify risks and rewards pertinent to making a business decision.

Decision science incorporates an economic framework — a consistent, rational and objective system to “price” each possible outcome, taking into account risks and rewards. It is simply a better way to make decisions.

Why use decision science?

Such a framework allows us to separate the bias and pitfalls often introduced by emotion and ego that are otherwise impossible to overcome. Remember the hot water Uber got into last year over its “God View,” which allowed the company’s staff to track both Uber vehicles and customers?

Calculating risks and rewards for all possible outcomes, taking into account all available data, is time-consuming. Computers can do these computations far better than humans. Machine learning techniques are now allowing computers to discerns patterns in very noisy data.

Decision science is most effective when it’s treated as, well, science.

As a result, computers are now better able to cull massive amounts of complex data and act as powerful scenario-generation engines, identifying the risks, rewards and uncertainties in a variety of important decisions. With this analysis at his or her fingertips, a decision-maker can confidently overlay judgment and experience to make better decisions.

Getting the most out of decision science

As an aside, decision science is most effective when it’s treated as, well, science. It’s a mistake for companies to invest massive amounts in building data warehouses to try to get a single “view” of data without really knowing what to do next.

Basic scientific principles call for doing some experiments and testing outcomes before you think about scaling. A scientific approach calls for testing with sample data, validating it and then integrating and expanding.

If you’ve seen the TV commercials featuring Bob Dylan talking to a computer, you’re familiar with a prime example of how computers are getting better at pattern matching and enabling better decision science. The ads are for IBM’s Watson, the “Jeopardy”-winning technology that takes in data from a multitude of sources and interprets it to expose patterns, connections and insights — such as the best course of treatment for a cancer patient.


The AI Governance Challenge

Software company SAS is another notable player in pushing the edges of decision science. Google and Microsoft offer modeling capabilities, though they tend to suffer from requiring too much work to adapt to the context/relevance in a functional area of the enterprise, such as sales.

Thanks to decision science, things that traditionally have been considered difficult or impossible to predict are proving, with the right tools, to be forecastable after all.

Using decision science to predict the unpredictable

Take HR, for example. HR departments typically have lagged behind the rest of their organizations when it comes to harnessing data, but more and more are starting to use data analytics to better find the right person for the job, as well as retain employees. Talent Science is an application vendor in this space.

“Predictive policing” has become one of the hottest emerging areas in law enforcement. According to the National Institute of Justice, “predictive policing leverages computer models, such as those used in the business industry to anticipate how market conditions or industry trends will evolve over time, for law enforcement purposes, namely anticipating likely crime events and informing actions to prevent crime.” Companies like PredPol and HunchLab aim to make real the crime-stopping technologies predicted in the 2002 movie Minority Report.

Simply put, decision science is a marriage of technology and business perspective to solve complex challenges. Executives who don’t catch on may find themselves with a DirecTV-style alter ego: “Hi, I’m yesterday’s business leader.”

Improving Criminal Profiling With Decision Science

This article originally appeared in [] and belongs to the creators.

Forensic psychologists are working with law enforcement officials to integrate psychological science into criminal profiling.

Criminal profiling: the reality behind the myth

For 16 years, “mad bomber” George Metesky eluded New York City police. Metesky planted more than 30 small bombs around the city between 1940 and 1956, hitting movie theaters, phone booths and other public areas.

In 1956, the frustrated investigators asked psychiatrist James Brussel, New York State’s assistant commissioner of mental hygiene, to study crime scene photos and notes from the bomber. Brussel came up with a detailed description of the suspect: He would be unmarried, foreign, self-educated, in his 50s, living in Connecticut, paranoid and with a vendetta against Con Edison–the first bomb had targeted the power company’s 67th street headquarters.

While some of Brussel’s predictions were simply common sense, others were based on psychological ideas. For instance, he said that because paranoia tends to peak around age 35, the bomber, 16 years after his first bomb, would now be in his 50s. The profile proved dead on: It led police right to Metesky, who was arrested in January 1957 and confessed immediately.In the following decades, police in New York and elsewhere continued to consult psychologists and psychiatrists to develop profiles of particularly difficult-to-catch offenders. At the same time, though, much of the criminal profiling field developed within the law enforcement community–particularly the FBI.

Nowadays profiling rests, sometimes uneasily, somewhere between law enforcement and psychology. As a science, it is still a relatively new field with few set boundaries or definitions. Its practitioners don’t always agree on methodology or even terminology. The term “profiling” has caught on among the general public, largely due to movies like “The Silence of the Lambs” and TV shows like “Profiler.” But the FBI calls its form of profiling “criminal investigative analysis”; one prominent forensic psychologist calls his work “investigative psychology”; and another calls his “crime action profiling.”Despite the different names, all of these tactics share a common goal: to help investigators examine evidence from crime scenes and victim and witness reports to develop an offender description. The description can include psychological variables such as personality traits, psychopathologies and behavior patterns, as well as demographic variables such as age, race or geographic location. Investigators might use profiling to narrow down a field of suspects or figure out how to interrogate a suspect already in custody.”In some ways, [profiling] is really still as much an art as a science,” says psychologist Harvey Schlossberg, PhD, former director of psychological services for the New York Police Department. But in recent years, many psychologists–together with criminologists and law enforcement officials–have begun using psychology’s statistical and research methods to bring more science into the art.

How does profiling work?

Informal criminal profiling has a long history. It was used as early as the 1880s, when two physicians, George Phillips and Thomas Bond, used crime scene clues to make predictions about British serial murderer Jack the Ripper’s personality.

At the same time, profiling has taken root in the United States, where, until recent decades, profilers relied mostly on their own intuition and informal studies. Schlossberg, who developed profiles of many criminals, including David Berkowitz–New York City’s “Son of Sam”–describes the approach he used in the late 1960s and 70s: “What I would do,” he says, “is sit down and look through cases where the criminals had been arrested. I listed how old [the perpetrators] were, whether they were male or female, their level of education. Did they come from broken families? Did they have school behavioral problems? I listed as many factors as I could come up with, and then I added them up to see which were the most common.”

In 1974, the FBI formed its Behavioral Science Unit to investigate serial rape and homicide cases. From 1976 to 1979, several FBI agents–most famously John Douglas and Robert Ressler–interviewed 36 serial murderers to develop theories and categories of different types of offenders.

Most notably, they developed the idea of the “organized/disorganized dichotomy”: Organized crimes are premeditated and carefully planned, so little evidence is found at the scene. Organized criminals, according to the classification scheme, are antisocial but know right from wrong, are not insane and show no remorse. Disorganized crimes, in contrast, are not planned, and criminals leave such evidence as fingerprints and blood. Disorganized criminals may be young, under the influence of alcohol or drugs, or mentally ill.

Over the past quarter-century, the Behavioral Science Unit has further developed the FBI’s profiling process–including refining the organized/disorganized dichotomy into a continuum and developing other classification schemes.

“The basic premise is that behavior reflects personality,” explains retired FBI agent Gregg McCrary. In a homicide case, for example, FBI agents glean insight into personality through questions about the murderer’s behavior at four crime phases:

  • Antecedent: What fantasy or plan, or both, did the murderer have in place before the act? What triggered the murderer to act some days and not others?
  • Method and manner: What type of victim or victims did the murderer select? What was the method and manner of murder: shooting, stabbing, strangulation or something else?
  • Body disposal: Did the murder and body disposal take place all at one scene, or multiple scenes?
  • Postoffense behavior: Is the murderer trying to inject himself into the investigation by reacting to media reports or contacting investigators?

A rape case is analyzed in much the same way, but with the additional information that comes from a living victim. Everything about the crime, from the sexual acts the rapist forces on the victim to the order in which they’re performed, offers a clue about the perpetrator, McCrary says.

Psychology’s contributions

Although the FBI approach has gained public attention, some psychologists have questioned its scientific solidity. Ressler, Douglas and the other FBI agents were not psychologists, and some psychologists who looked at their work found methodological flaws.

Former FBI agent McCrary agrees that some of the FBI’s early research was rough: “Early on it was just a bunch of us [FBI agents] basing our work on our investigative experience,” he says, “and hopefully being right more than we were wrong.”


The AI Governance Challenge

McCrary says he believes that they were right more than wrong, though, and emphasizes that FBI methods have improved since then. In the meantime, psychologists have also been helping to step up profiling’s scientific rigor. Some psychologists have been conducting their own criminal profiling research, and they’ve developed several new approaches:

  • Offender profiling. Much of this work comes from applied psychologist David Canter, PhD, who founded the field of investigative psychology in the early 1990s and now runs the Centre for Investigative Psychology at the University of Liverpool.

Investigative psychology, Canter says, includes many areas where psychology can contribute to investigations–including profiling. The goal of investigative psychology’s form of profiling, like all profiling, is to infer characteristics of a criminal based on his or her behavior during the crime. But, Canter says, the key is that all of those inferences should come from empirical, peer-reviewed research–not necessarily from investigative experience.

For example, Canter and his colleagues recently analyzed crime scene data from 100 serial homicides to test the FBI’s organized/disorganized model. Their results, which will be published in an upcoming issue of APA’s Psychology, Public Policy and Law, indicate that, in contrast to some earlier findings, almost all serial murderers show some level of organization.

Organized behaviors–like positioning or concealing a victim’s body–are the “core variables” that tend to show up most frequently and co-occur with other variables most often, he found. The differences between murderers, the researchers say, instead lie in the types of disorganized behaviors they exhibit. The study suggests that serial murderers can be divided into categories based on the way they interact with their victims: through sexual control, mutilation, execution or plunder.

Canter says that research like this, which uses the statistical techniques of psychology to group together types of offender behaviors, is the only way to develop scientifically defensible descriptions and classifications of offenders.

“Our approach,” he says, “is to consider all the information that may be apparent at the crime scene and to carry out theory-based studies to determine the underlying structures of that material.”

In another study, he and his colleagues collected crime scene data from 112 rape cases and analyzed the relationship among different crime scene actions–from what types of sexual acts the rapist demanded to whether he bound the victim. The researchers found that the types of sexual violation and physical assault did not distinguish rapists from each other; these were the core variables that occurred in most rape cases. Instead, what distinguished the rapists into categories were nonphysical interactions–things like whether they stole from or apologized to the victim.

Canter puts little faith in the investigative experience-derived offender descriptions developed by law-enforcement agents. As he sees it, psychologists need to work from the ground up to gather data and classify offenders in areas as various as arson, burglary, rape and homicide.

  • Crime action profiling. Forensic psychologist Richard Kocsis, PhD, and his colleagues have developed models based on large studies of serial murderers, rapists and arsonists that act as guides to profiling such crimes. The models, he says, are similar to the structured interviews clinical psychologists use to make clinical diagnoses. They come out of an Australian government-funded research program that Kocsis ran, in which he developed profiling methods in collaboration with police and fire agencies.

Now in private practice, Kocsis says crime action profiling models are rooted in knowledge developed by forensic psychologists, psychiatrists and criminologists. Part of crime action profiling also involves examining the process and practice of profiling.

“Everybody seems to be preoccupied with developing principles for profiling,” Kocsis explains. “However, what seems to have been overlooked is any systematic examination of how to compose a profile. What type of information do, or should, profiles contain? What type of case material do you need to construct a profile? How does the presence or absence of material affect the accuracy of a profile?”

He has studied, for example, whether police officers perceive the same profile to be more accurate and useful when they believe it was written by a professional profiler rather than a layperson.

Kocsis agrees that the future of profiling lies in more empirically based research. He also believes, though, that just as some clinicians are better than others, there is also a skill element involved in profiling. Is profiling an art or a science? “Realistically, I think it is probably a bit of both,” he says.

The psychology-law enforcement relationship

Among those in the profiling field, the tension between law enforcement and psychology still exists to some degree. “The difference is really a matter of the FBI being more oriented towards investigative experience than [academic psychologists] are,” says retired FBI agent McCrary.

“But,” he adds, “it’s important to remember that we’re all working toward the same thing.”

In recent years, the FBI has begun to work closely with many forensic psychologists–in fact, it employs them. Psychologist Stephen Band, PhD, is the chief of the Behavioral Science Unit, and clinical forensic psychologist Anthony Pinizzotto, PhD, is one of the FBI’s chief scientists.

The unit also conducts research with forensic psychologists at the John Jay College of Criminal Justice in New York. One recent collaborative study, for example, looked at the relationship between burglaries and certain types of sexual offenses–whether specific aspects of a crime scene differed in incidents that began as a burglary and ended in a sexual offense, as opposed to crimes that began as a sexual offense but included theft. Police looking at the first type of crime might want to look for convicted burglars in the area, Pinizzotto explains. The study will be published in an upcoming issue of Sex Offender Law Report, published by the Civic Research Institute.

One of the FBI’s collaborators at John Jay College is Gabrielle Salfati, PhD, a graduate of the Centre for Investigative Psychology. “Whenever we do research, we try to bring in as many varied points of view as possible,” Pinizzotto says. “Gabrielle Salfati’s expertise on the statistical aspects of evaluating crime scenes is a great contribution.”

More recently, the unit has also begun to collaborate with forensic psychologists at Marymount University in Arlington, Va.–another indication that law enforcement and psychology will continue to work together.

“I think,” says Band, “that there is an incredible value added when applications of professional psychology enter into the mix of what we do.”

Your Vote Counts, But Does It Matter?

As William Vaughan’s wisecrack goes, “A citizen of America will cross the ocean to fight for democracy, but won’t cross the street to vote in a national election.” 

This thought-provoking quip exposes the absurdity within the problem of declining voter turnout. Readers will likely respond to this phrase with either instinctive or systematic responses. Instinctively, some will blame voters for their foolish and lazy nature. Alternatively, others will respond to this allegation with more skepticism. This dichotomy raises important questions in public discourse. Are the voters’ folly, lethargy and irrationality really to blame? Or, can there be another culprit?

In this article, arguments proposed by public choice theory are used to explain why voters are not simply apathetic about the electoral process. To begin, comparisons are drawn between voter and consumer behavior to demonstrate why apathy is not the only source of the problem.

Voters vs. Consumers

If consumers are uninformed, and unable to differentiate good from bad products, then competition in the marketplace will not lead to improvements in product quality. Producers will lack incentives to improve product quality simply because customers will not be willing to pay premium prices. The same is true for politicians and policies. Competition between political candidates will not improve the quality of policies if voters are uninformed. In both cases, our ignorance is incentivizing the negligence of policy and product providers (i.e., politicians and marketers).

However, it would be fallacious to claim that self-interest is unrequited by the public. Everyone has an ‘agenda’ of special self-interests! The public is self-interested in securing the ‘best deals’ in the market and the ‘best representation’ in politics. Meanwhile, politicians and corporations are self-interested in turning a profit and gaining power. The even more harrowing truth is that incentives in the democratic system are designed to reward self-interest! Importantly, there is a key distinction in how self-interest operates with voters and consumers. To understand this, one must simply apply some relative price logic.


In the market, you choose from either X or Y, if you choose X you will get X.


In the ballot box, if you choose X over Y, you may or may not get X.

If the majority chooses X, you get X. If the majority chooses Y, you get Y

Simply put, there is individual inconsequentiality for voters but not consumers. Logically, one can make a case to argue that voters are more vulnerable than consumers by accepting these premises. Individual inconsequentiality is problematic because it negatively impacts motivation. One important mechanism to explain this is instrumental (i.e., benefit to oneself from the functional use of a mean, agent or tool) and procedural (i.e., benefit depending on the actions of oneself) motivations. For example, a voter may vote simply because democracy is an instrument or tool to express one’s political preference. Alternatively, some voters may vote because of the need to fulfill one’s civic duty, participate in the ‘democratic process’, express allegiance or loyalty to a political party, or judge the moral superiority of one candidate. Irrespective of what motivates voters, the ratio of instrumental and procedural benefits is unequal when comparing voters with consumers.


In the market, instrumental and intrinsic benefits trade at 1:1. A dollar’s worth of instrumental benefit is worth a dollar’s worth of intrinsic benefit.


At the ballot box, a dollar’s worth of intrinsic benefit remains a dollar, a dollar’s worth of instrumental benefit shrinks to one dollar times the probability of one’s vote being decisive (i.e., one’s vote decides the electoral result because there is an exact tie among the rest of the voters)

The chances that your vote will be the deciding one in a national democratic election is nearly ~ 0. Most optimistic estimations suggest that there is a 1 in 10,000 chance of your vote being decisive. Thus, the relative instrumental to procedural benefits for voters is 1:10,000, which is drastically lower than the odds for consumers.

Here is a quick thought experiment. Imagine I have hidden a huge sum of money in a secret place. In which of the following two cases do you feel more motivated to find it?


I have hidden 1 million dollars somewhere. The clue to its location lies in this book here.


I have hidden 100 million dollars somewhere. The clue to its location lies in a random page of one book among 54 million other books at the Library and Archives of Canada.

The political attitude of many voters coincides with the latter case. The expected reward does not exceed the costs, so voters think it is not worth the price (i.e., reading through billions of pages to look for a clue). Often, it is not that voters simply do not care. Voters understand that their vote is individually meaningless. Voters are ignorant because acquiring political knowledge is not worth the time and effort since they do not expect a proportionate change in politics.


The AI Governance Challenge

Skepticism vs Cynicism

Cynicism may seem like a logical response to the constant wave of bad news. We read headlines formatted in bold fonts and capital letters that shout about scandals, conspiracies, and unsolvable problems. We hear soundbites on television that fragment viewers and attack opposing networks. We follow tweets, and Facebook statuses that present oversimplifications of complex issues. Amongst all this sound and fury, it is tempting to adopt a cynical approach to politics.

It is enticing to respond with constant distrust and suspicion. One can easily become a critic who either throws rocks from the sidelines or disengages with the public debate altogether. This position requires very little creativity or effort. I urge you to take a different approach! One ought to recognize that so long as the government, media and business are run by imperfect human beings, we will feel disappointed and frustrated. Voters and consumers are not victims! We have the same imperfections and inadequacies as our leaders that are driven by their self-interest, ego, and greed.

To sum up, as George Jean Nathan once said, “bad officials are elected by good citizens who do not vote.” It seems that a voter who puts the power of their single vote in perspective, is truly informed and so most eligible to cast a ballot.