Behavioral Economics in Criminal Justice Messaging August 2015

Acknowledgments This report was researched and authored by Sabrina Hassan for The Opportunity Agenda and edited by Julie Fisher-Rowe.

www.opportunityagenda.org

2

Behavioral Economics in Criminal Justice Messaging

INTRODUCTION What behavioral economics does 3 Part I: Select Principles in Behavioral Economics 4 Behavioral phenomena that guide human decisions and supporting research Social Norms 4 Loss Aversion 6 Availability Bias 7 Limited Attention 9 Identity 11 Part II: Applications of BE Principles in Criminal Justice Messaging Illustrations of how to leverage behavioral quirks in three key areas of messaging Registries for Sex-Related Offenses 13 Drug Policy 18 Racial Profiling 21 CONCLUSION 23

3

INTRODUCTION While I worked on this project, I sat in a coffee shop. The inside of the door to the shop was fixed with a vertical handlebar that one could grasp to open the door. Time and again, I watched departing customers grasp the handle and pull in vain on the door that swung outward, even though it was marked with a sign above the handle that read “PUSH” in plain block letters. Oftentimes despite the availability of certain information, we rely on our split-second intuition to make decisions. We are humans. We are biased. Sometimes we are lazy or inattentive. Instead of computing all of the stimuli available to us, we often act on impulse instead of considering all of our options. 1 Behavioral economics (BE) allows us to predict what people actually do in decision-making contexts instead of depending on people to behave like rational, controlled, forward-thinking computers. By understanding what influences real choices, we can design choice settings that guide people to choose in a certain way. We can fix the inside of an outward-swinging door with a flat metal plate instead of asking, “Can’t they read?” Behavioral economists have developed ways to “nudge” people into better choices, which is to encourage selection of certain options without eliminating or taxing alternatives.2 For example, putting healthy food at eye level in a school cafeteria encourages students to eat more nutritious meals; choosing junk requires reaching for a different shelf. A state can increase its number of organ donors by instituting a default rule of presumed consent to donate; opting out of donation requires unchecking a box. This paper introduces behavioral economics as a way to improve criminal justice messaging. Specifically, Part I of the paper introduces a few key concepts of behavioral economics to consider when designing messages. Part II suggests specific ways to apply those concepts in messages dealing with each registry regarding sex-related offenses, drug policy, and racial profiling.

1 Richard H. Thaler and Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (New Haven & London: Yale University Press, 2008). 2 Ibid.

4

PART I : Select Principles in Behavioral Economics Social Norms We tend to follow the herd. People demonstrate a tendency to conform to perceived behavior of others in their groups. In other words, a person is likely to do what she thinks “everyone else” is doing. Researchers have harnessed the positive potential of this phenomenon, motivating desirable behavior by publishing prosocial group norms to group members. The decision settings in which people gravitate toward perceived social norms span a range of variables including whether to vote,3 whether to litter or commit crimes,4 whether to react to an emergency,5 and how quickly to finish easy tasks.6

In an often-cited intervention, the power company OPOWER let its customers know how their household energy usage compared to that of their neighbors by periodically mailing them Home Energy Reports.7 These reports showed customers a graph comparing their energy usage to the median and 20th percentile usage among their neighbors with similarly sized homes. As a result, customers using more energy than average decreased their consumption. However, the norm works both ways—consumption by customers using less than average increased. Like other experiments investigating the power of social norms, these results show that people tend to migrate

3 Alan S. Gerber and Todd Rogers, “Descriptive Social Norms and Motivation to Vote: Everybody’s Voting and so Should You.” The Journal of Politics 71, No. 1(2009): 178–191. http://scholar.harvard.edu/files/todd_rogers/files/descriptive_social_norms_and_motivation_to_vote-everybodys_voting_and_so_should_you.pdf 4 See James Q. Wilson and George L. Kelling ,“Broken Windows: The police and neighborhood safety.” The Atlantic (March 1982). http://www.theatlantic.com/magazine/archive/1982/03/broken-windows/304465/ 5 Broadly speaking, the bystander effect is the decreased likelihood that an individual will act among a group of non-actors. John M. Darley and Bibb Latane, "Bystander Intervention in Emergencies: Diffusion of Responsibility." Journal of Personality and Social Psychology 8 (1968): 377-383. http://psycnet.apa.org/journals/psp/8/4p1/377.pd 6 See Robert B. Zajonc, “Social Facilitation.” Science 149, No. 3681 (1965): 269–274. http://www2.psych.ubc.ca/~schaller/Psyc591Readings/Zajonc1965.pdf 7 Hunt Allcott, “Social Norms and Energy Conservation.” Journal of Public Economics 95 (2011): 1082–1095. http://www.sciencedirect.com/science/article/pii/S0047272711000478

5

toward social norms not only when their behavior is less socially beneficial than normal behavior, but also when it is more so.

Energy customers reduced consumption after learning that their neighbors

consumed less than they did.

The practical application of social norms is straightforward: one need simply publicize evidence that the vast majority of a group to which a given audience belongs (Democrats, seniors, minorities) does something the audience should be doing. A key to using this principle is knowing the composition of the audience. Announcing what 80% of Hispanics think to an audience of mostly Asians may be ineffective or even counterproductive. Also, take care not to emphasize undesirable majority behaviors of a group with whom the audience identifies. Though it may seem rational to advertise the severity of a behavior problem as a way to spur a counter-movement, people tend to follow the herd.8 See applications of social norms in criminal justice messaging on p. 19. Further reading: Alter, Adam. Drunk Tank Pink. New York: The Penguin Press, 2012. (See Part II: The World Between Us.) Gerber, Alan S. and Rogers, Todd. “Descriptive Social Norms and Motivation to Vote: Everybody’s Voting

and So Should You.” The Journal of Politics 71. 1 (2009): 178–191. Online. (Available at

http://scholar.harvard.edu/files/todd_rogers/files/descriptive_social_norms_and_motivation_to_vote-everybodys_voting_and_so_should_you.pdf)

Hunt Allcott. “Social Norms and Energy Conservation.” MIT and NYU. (2010) Online. (Available at http://opower.com/uploads/library/file/1/hunt_allcott__june_2010_-_social_norms_and_energy_conservation.pdf)

Thaler, Richard and Sunstein, Cass. Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven & London: Yale University Press, 2008. (Chapter 3.)

8For example, advertising low recent voter turnout produced lower reported likelihood of voting compared to advertising high recent voter turnout. Gerber and Rogers, “Descriptive Social Norms and Motivation to Vote.”

6

Loss Aversion9 We favor avoiding losses over acquiring gains. People are loss averse, meaning they prefer avoiding losses to making gains. The psychological pain of losing X exceeds the pleasure of gaining X. Take the following example10 for illustration: Problem 1: Which do you choose? Get $900 for sure OR 90% chance to get $1,000 Problem 2: Which do you choose? Lose $900 for sure OR 90% chance to lose $1,000 Most people are risk averse in Problem 1, choosing the certain $900 over the very likely $1,000 to avoid the 10% risk of getting nothing. But in Problem 2, the opposite is true. Although the amounts are identical, most people choose to gamble on the 90% probability that they will lose an additional $100; they risk losing the larger amount rather than admitting a loss of 10% less. These results demonstrate loss aversion, our built-in distaste for losses. Loss aversion also explains why people buy insurance they don’t need and refuse to settle a lawsuit that they will likely lose.11 We detest losses so much that we pay extra for a narrow possibility of escaping them.

Loss aversion prevents a defendant from settling a lawsuit he or she is likely to

lose.

Loss aversion is a useful tool in designing choices because whether something is perceived as a loss depends on how it is framed. Just as a glass of water can be deemed half empty or half full, a policy can be described in terms of its costs or its benefits. In addition, the consequence of a choice can take the form of a probability (2% chance of death) or a loss from the status quo (for every 100 people, two will die). Thus in an experiment asking participants to choose between two hypothetical aid programs to combat a lethal pandemic, most people chose the program with more certain deaths when the programs are framed in terms of lives saved (200 of 600 saved versus a one-

9 Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux: 2011). See Chapters 26–28. 10 Ibid., 280. 11 Ibid., 318 (re insurance) and 320 (re risk-seeking defendant).

7

third probability of saving all lives), but they choose the program with a greater risk of all patients dying when the very same programs are framed in terms of lives lost (400 lives lost versus a two-thirds probability of losing all 600 patients.)12 We can help an audience perceive flaws in a policy by framing a policy concern as a loss of value from a reference point. See applications of loss aversion in criminal justice messaging on pp. 15 and 19. Further reading: Kahneman, Daniel. Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. 2011. (See Chapters 26-28.)

Sunstein, Cass. “Moral Heuristics.” Behavioral and Brain Sciences 28.4 (2005): 531–573. Online.

(Available at http://wrightjj1.people.cofc.edu/psychology/Moral%20Heuristics%20Published%20August%202005.pdf.)

Availability Bias13 We estimate probability (of an event or scenario) by how easily examples come to mind. When asked whether the letter “K” is more likely to appear as the first or third letter of a word, most people choose the first. When asked whether death is more frequently caused by asthma or by tornadoes, most people choose tornadoes.14 Most people are wrong on both counts. These misjudgments illustrate the availability bias: we overestimate the occurrence of events for which it is easy to recall examples. It is difficult to think of words whose third letter is “K,” and it is easy to think of killer tornadoes, which get far more media attention than deaths by asthma. When estimating frequency of an event or size of a category, we replace the question of “how often?” with the easier question of “how easily can I think of examples?” The more easily we can think of instances in a category, the more frequent we estimate its incidence to be.

12 Daniel Kahneman and Amos Tversky, “Asian disease problem,” summarized in Cass R. Sunstein, “Moral Heuristics.” Behavioral and Brain Sciences 28.4 (2005): 531–573. http://wrightjj1.people.cofc.edu/psychology/Moral%20Heuristics%20Published%20August%202005.pdf 13 Amos Tversky and Daniel Kahneman, “Availability: A Heuristic for Judging Frequency and Probability.” Cognitive Psychology 4(1973): 207–232. http://psych.colorado.edu/~vanboven/teaching/p7536_heurbias/p7536_readings/tversky_kahn_1973.pdf 14 Kahneman, Thinking Fast and Slow, 138.

8

Most people estimate deaths by tornadoes to be more frequent than deaths by asthma, although the latter are 20 times more likely.

An important concept related to availability is representativeness, which is how closely an example fits a stereotype. In an early 1970s experiment demonstrating the influence of representativeness, participants received a description, which they were told was based “on psychological tests of uncertain validity,” of a man named Tom W.15 Tom was described as intelligent, lacking creativity, needing order and clarity, a dull writer who likes corny puns, having little sympathy, and self-centered but moral. Participants were asked to rank nine given fields of study by the likelihood that Tom was a graduate student in those fields. Given the dubious nature of the description, a rational agent simply would have correlated the likelihood that Tom studied in a field to the relative size of each field. But the average participant ranked computer science first and humanities and education, a far more populous field, eighth. The “uncertain validity” detail had no effect on responses. In fact, the above experimental results replicated answers from students who were asked not how likely Tom was to study in each field but “how similar the description of Tom W [was] to the typical graduate student” in each field! As with the killer tornado question, participants in the Tom W study substituted a question of probability with a question that was easier to answer: representativeness. “How closely does Tom resemble the stereotypical student in each field?” A plausible narrative trumped statistical reasoning. Availability and representativeness are partly to entirely responsible for many popular misconceptions. We can leverage these behavioral phenomena by presenting and popularizing true individual narratives that represent the facts, making new examples easy to recall or envision. See applications of availability bias in criminal justice messaging on p. 22. Further reading: Kahneman, Daniel. Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. 2011. (See Chapters

12–16.)

15 Ibid., Chapter 14.

9

Limited Attention When our attention is depleted, we don’t make good decisions. Our performance suffers when we attempt to do multiple things at once16; texting and driving is one palpable example. And even when we are focusing on only one activity, we draw from an exhaustible (though renewable) resource and will eventually fatigue. In other words, our attention is limited. Each person’s pool of mental energy is shared by cognitive, emotional, and physical efforts.17 Consider how hard it is to concentrate on work after receiving shocking news or to perform complex mental math while walking. (If you’re like most people, you’ll stand still to finish computing.)18 When a person exerts self-control by thinking hard, exercising vigorously, or suppressing emotion, she will eventually experience a diminished capacity to regulate her thoughts, feelings, and actions.19 Poor decisions ensue. The limits of our attention can have dramatic life consequences. Consider, for example, the forms a college hopeful must complete in order to apply for financial aid. A few years ago, the Free Application for Federal Student Aid (FAFSA) was a complex series of 153 questions that took hours of focus to complete.20 Researchers suspected that the demands of the application were inhibiting college enrollment. In the study21 that confirmed the researchers’ hunch, some participants received help completing the FAFSA from an H&R Block accountant, who pre-populated most fields of the questionnaire using information from the applicant’s tax return and interviewed the applicant to answer the remaining questions before offering to submit the FAFSA. The accountant also offered these participants applicant aid eligibility estimates and informed them of local tuition costs.

16 Kahneman describes the limited capacity to perform simultaneous tasks as “cognitive busyness.” Ibid., 41. 17 Roy F. Baumeister and John Tierney, Willpower: Rediscovering the Greatest Human Strength (New York: The Penguin Group, 2011) 22–39. See Kahneman, Thinking, 41. 18 Kahneman notes this tendency in Thinking, 39–40. 19 Baumeister coined the term “ego depletion” to describe this phenomenon. Willpower, 28. 20 Executive Office of the President, Council of Economic Advisers, National Economic Council, “Simplifying Student Aid: The Case for an Easier, Faster, and More Accurate FAFSA.” September 2009. https://www.whitehouse.gov/assets/documents/FAFSA_Report.pdf 21 Eric P. Bettinger et al., “The Role of Simplification and Information in College Decisions: Results from the H & R Block FAFSA Experiment,” National Bureau of Economic Research. September 2009. http://www.nber.org/papers/w15361.pdf

10

Participants who received help completing the FAFSA were up to almost three times as likely to submit a FAFSA, and such participants who lacked college experience were 20 to 30% more likely to enroll in college than other participants. Those who received only information about aid eligibility and tuition costs, by contrast, were no more likely to submit the FAFSA or to enroll in college than those who received neither information nor application assistance. Thus it seems the burden of the application, rather than ignorance about aid availability, was to blame for non-submission.

Individuals who received help completing their lengthy federal financial aid

applications were over 20% more likely to enroll in college.

In designing choice settings, we must remember that giving information isn’t enough to incite action. We must make the effort to simplify complex or lengthy processes that lead to a desired result. Even better, we can “pad the path of least resistance” with default rules that transparently funnel people to one path unless they opt out.22 When we eliminate attention-sucking decision points with a default rule, such as automatic employee enrollment in a retirement plan, people are more likely to subscribe to a choice rather than avoid choosing at all. 23 These principles apply in messaging. By simplifying messages and eliminating decision points—on whether to turn a page, download a document, do mental math—the messenger can increase the chance that an audience will remain attentive enough to absorb her message. See applications of limited attention in criminal justice messaging on pp 13 and 18. Further reading: Bettinger, Eric et al. “The Role of Simplification and Information in College Decisions: Results from the H&R Block FAFSA Experiment.” National Bureau of Economic Research (2009), Working Paper 15361. Online. (Available at http://www.nber.org/papers/w15361.pdf). Baumeister, Roy F. and Tierney, John. Willpower: Rediscovering the Greatest Human Strength. New York: The Penguin Group, 2011. (See Chapters 1, 4.) Thaler, Richard and Sunstein, Cass. Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven & London: Yale University Press, 2008. (See Chapter 5.)

22 Thaler and Sunstein, Nudge, 83. 23 Ibid., 129.

11

Identity We embrace ideas and actions that affirm our self-concept and reject those that threaten it. We are very attached to our identities. I believe that I am X kind of person, and X kind of person believes and does Y. The attachment is so strong that it may account for my inaccurate beliefs and irrational behavior. Many Americans hold factually unsupported beliefs. Sometimes these beliefs become entwined with self-concept and thus are difficult to overturn. For example, despite a dearth of evidence, some parents believe there is a link between vaccinations and autism. In a study attempting to increase such parents’ willingness to vaccinate their children, researchers presented various forms of accurate pro-vaccine information to parents—facts about the lack of correlation between vaccines and autism, facts about the dangers of the diseases vaccinated against, photos of children suffering from the diseases, and a story about one baby who nearly died from one of the diseases.24 None of the information improved the parents’ intent to vaccinate, and some parents became even less inclined to do so.25 Once again, information was not enough to change behavior. Recent evidence suggests two techniques that may overcome the hurdle presented to people who encounter information that conflicts with their identities. 26 The first is presenting accurate information in a graphical, rather than textual, format. Presenting corrective information graphically is shown to be more effective than conveying the same information through words alone. Second, a presenter can leverage the power of self-affirmation—buttressing perceptions of self-worth. Having a person engage in self-affirmation (e.g., by recalling a time she felt good about herself) increases her willingness to admit a position that conflicts with her identity. Self-affirmation can reduce misperceptions even without corrective information. Research supports the effectiveness of both graphical presentations and self-affirmation in reducing misperceptions regarding the success of the U.S. troop surge in Iraq following the 2006 elections, job growth in 2010, and global warming.27

24 Maria Konnikova, “I Don’t Want to Be Right,” The New Yorker (May 19, 2014). 25 Ibid. 26 Bryan Nyhan and Jason Reifler, “Blank slates or closed minds? The role of information deficits and identity threat in the prevalence of misperceptions” (2014). http://www.dartmouth.edu/~nyhan/opening-political-mind.pdf 27 Ibid.

12

Other studies suggest that the need for a positive self-concept can also be used to motivate positive behavior. People not only cling to their identities to deflect factual information, but also in some cases they may engage in behavioral changes for the purpose of aligning with a proposed identity. Labeling an individual as a certain kind of person( e.g., “voter” or “consumer”) can make her act accordingly.28 In a 2011 study, participants received surveys that referred to voting using either a self-relevant noun (e.g., “How important is it to you to be a voter in the upcoming election?”) or a verb (e.g., “How important is it to you to vote in the upcoming election?”). Those who completed the surveys identifying them as prospective “voters” (noun condition) expressed significantly greater interest in registering to vote and significantly increased voter turnout than those in the verb condition.29 Being cognizant of the value an audience places on its identity can thus guide a messenger in multiple ways. The messenger can present identity-threatening information graphically or after priming the audience with self-affirming thoughts, or she can nudge audience members into label-consistent behavior by using a self-relevant noun to which they will conform. See applications of identity in criminal justice messaging on pp. 13 and 21. Further reading: Bryan, Christopher et al. “Motivating voter turnout by invoking the self.” Proceedings of the National

Academy of Sciences 108.31 (2011): 12653–12656. Online. (Available at

http://www.pnas.org/content/108/31/12653.full.) Nyhan, Bryan and Reifler, Jason. (2014) “Blank slates or closed minds? The role of information deficits and identity threat in the prevalence of misperceptions.” Dartmouth College. Online. (Available at http://www.dartmouth.edu/~nyhan/opening-political-mind.pdf.) Bauer, Monika et al. “Cuing Consumerism: Situational Materialism Undermines Personal and Social Well-

Being.” Psychological Science 23.5 (2012) 517–523. Online. (Available at

http://pss.sagepub.com/content/23/5/517.)

Part II: Applications of BE Principles in Criminal Justice Messaging REGISTRIES FOR SEXUAL-RELATED CONDUCT

28 Christopher Bryan et al., “Motivating voter turnout by invoking the self.” Proceedings of the National Academy of Sciences 108.31 (2011): 12653–12656. http://www.pnas.org/content/108/31/12653.full.); Monika Bauer et al, “Cuing Consumerism: Situational Materialism Undermines Personal and Social Well-Being.” Psychological Science 23.5 (2012): 517–523, at 520. 29 Logic tells us that there are limits to this theory (e.g., mislabeling an audience of Republicans as Democrats would not likely sway them to vote against their party).

13

Research reveals that the benefits of laws relating to sexual-related conduct do not justify their costs. In a nutshell, laws requiring registration, community notification, and residency restrictions for an overbroad category of people convicted of sex-related crimes end life opportunity for non-dangerous people without enhancing community safety.30 Given the ghastly nature of the word “sex offender,” the idea that we should relax monitoring and restrictions of these people is difficult to accept. The goal in communicating about registries is to make it easier for the audience to understand and acknowledge flaws in the current system. At least a few principles of behavioral economics can help activists present the facts in a way that makes them easier to adopt. Limited Attention/Identity: For Hard-to-Believe Figures, Show, Don’t Tell In addition to being hard to recall, some statistics can be hard to believe. Most people likely identify as anti-“sex offender.” Thus bare statistics about the relative harmlessness of people convicted of sex-related crimes can be easy to ignore or reject. Since graphical representations can erode the loyalty people have to their own misconceptions, we should graphically communicate the low risk that people convicted of these crimes pose to the community. Sample communications of actual statistics follow.31

30 See Human Rights Watch, “No Easy Answers: Sex Offender Laws in the US,” (September 2007) and “Raised on the Registry: The Irreparable Harm of Placing Children on Sex Offender Registries in the US (May 2013). http://www.hrw.org/sites/default/files/reports/us0907webwcover.pdf and https://www.hrw.org/report/2013/05/01/raised-registry/irreparable-harm-placing-children-sex-offender-registries-us 31 Although these graphs represent documented research, they are included for the purpose of illustrating a tool; more current/accurate data may be available.

14

32

Most Victims of Sex-Related Crimes Already Know the Person Responsible33

32 Three out of four people convicted of sex-related crimes do not reoffend. See “No Easy Answers,” 27. 33 More than 90% of sex-related crimes against children are committed by known and trusted adults. See “No Easy Answers,” 4.

Most People Convicted of Sex-Related Crimes Do Not Reoffend

Portion of sex offenders whodo not reoffend within 15years

0 20 40 60 80 100

Sexual Assault Cases AgainstChildren Under 18

All Sexual Assault Cases

Person Who Assaulted IsStranger To Victim

Person Who Assaulted IsFamily Member orAcquaintance of Victim

15

People Convicted of Sex-Related Offenses Are Among the Least Likely to Reoffend34

Other candidates for graphical representation of facts about sex-related offenses include the following:

Less than 10% of adults who commit sex-related offenses had committed any sex-related offenses as juveniles.35

Most sex-related crimes are not committed by former offenders.36

Presenting facts that undermine the logic of laws regarding sex-related offenses graphically facilitates support for reform. Loss Aversion: Subtract from a Reference Point Promoting sympathy for people convicted of sex-related crimes is a tall order, even if those people have to wait in a registration line for hours in the blazing heat or freezing cold. Most people don’t value the time or comfort of “sex offenders.” However, people do value their own safety and their (taxpayer) money. A number of facts about laws related to sexual behavior can be framed to elevate concern for these values.

34 “Raised on the Registry” 22, citing Bureau of Justice Statistics Special Report, “Recidivism of Prisoners Released in 1994,” (June 2002). http://bjs.gov/content/pub/pdf/rpr94.pdf. 35 Human Rights Watch, “No Easy Answers,” 9. Ibid., 24.

0

5

10

15

20

25

30

35

40

45

Percentage of Prisoners Re-Arrested for Same Crime

Within Three Years of Release

Homicide 1.2%

Rape 2.5%

Motor Vehicle Theft 11.5%

Robbery 13.4%

Fraud 19%

Non-Sexual Assault 22%

Burglary 23.4%

Public Order Offenses 31.2%

Larceny/Theft 33.9%

Drug Offenses 41.2%

16

The over-inclusion of people on the registry makes it difficult for law enforcement to determine which of them warrant careful monitoring.37

The large volume of non-dangerous people required to register demands hours of administrative police work and sometimes even requires paying for overtime to keep up with demand.38

Residency restrictions push people who have been released from prison away from the supervision, treatment, stability and supportive networks they may need to build and maintain successful, law-abiding lives.39

Residency restriction laws cause police to lose track of people who become transient or drop out of sight to avoid compliance with the law.40

It may be tempting to cite the benefits that would come from reforming the laws. The principle of loss aversion suggests, however, that it is more effective to focus on what current laws cost the public in terms of loss from a reference point. The reference point is desirable laws. The following messages illustrate how to frame the above facts in terms of losses from that reference point. Whenever possible, quantify the loss in absolute terms rather than risk.

Many registration systems are undifferentiated by the danger the registrants pose to the community, if any, and include more registrants than police can monitor. Compared to a registration system that differentiates convicts who pose an elevated risk of danger, an over-inclusive registration system deprives communities of the safety of monitoring potentially dangerous people.

The large number of people required to register diverts a significant amount of police work to registering non-dangerous people. Compared to a system with tailored registration rules, an over-inclusive registration system loses [number] tax dollars that could fund community protection to administrative police work.

Residency restrictions often prohibit people who have been convicted of sex-related offenses from living near friends or family, increasing recidivism risk factors such as isolation, financial hardship, decreased stability, and lack of support. Compared to those who are free of residency restrictions, more people who are subject to residency restrictions will reoffend.

37 “No Easy Answers,” 9. 38 Rob Wildeboer, “Chicago police easing registration for sex offenders.” WBEZ 91.5 (June 23, 2014). http://www.wbez.org/chicago-police-easing-registration-sex-offenders-110392 39 Colorado Sex Offender Management Board, “White Paper on the Use of Residence Restrictions as a Sex Offender Management Strategy,” (June 2009). http://thecrimereport.s3.amazonaws.com/2/db/1/2486/d_co_sex_offender_management_board_white_paper.pdf, at 4. 40 Ibid.

17

Residency restrictions severely limit the locations where people convicted of sex-related crimes can legally live, causing some of them to fail to register. Compared to systems without residency restrictions, systems with residency restrictions eliminate the safety of knowing where offenders live.

Keeping in mind that people prefer avoiding losses to acquiring gains, we can draw opposition to a system by identifying its costs in definite terms.

18

Drug Policy The problem with American drug policy is that it continues the failed war on drugs. Our country has spent more than a trillion dollars fighting drugs in the past four decades41 and incarcerates 10 times as many violators of drug laws than it did in 1980,42 but the ease and frequency of illegal drug use has not decreased. Problems resulting from over-incarceration are numerous and compound both for those convicted of drug offenses and for society at large. Communications on drug policy should aim to garner support for more tolerant laws and less harsh punishment. Limited Attention: Less Is More In a field as well documented as the failed war on drugs, it can be tempting to share heaps of facts and figures with any attentive audience. But it is important to remember that any attention is limited; more information shared doesn’t necessarily translate to more information received. The way to capitalize on the attention we get is to maximize the impact, not the volume, of our messages. Most Americans believe crime is increasing despite the reality that violent crime has been at historic lows for the past few years.43 We thus can’t assume that Americans perceive the gross increase in incarceration as bad or illogical. We must spell out the problem. Fortunately, an audience unfamiliar with or resistant to the glaring flaws in American drug policy needs to accept only a few simple facts to grasp the issue: 1) America incarcerates an astronomical number of people compared to both the American past and to other countries.44 2) The increase in incarceration for drug offenses results from harsher sentencing policy, not from an increase in convictions for drug offenses. 45

41 Drug Policy Alliance, “Wasted Tax Dollars,” http://www.drugpolicy.org/wasted-tax-dollars 42 Ibid. 43 The Opportunity Agenda, “An Overview of Public Opinion and Discourse on Criminal Justice Issues,” (August 2014), 11–13. http://opportunityagenda.org/criminal_justice 44 Rebecca Vallas and Sharon Dietrich, “One Strike and You’re Out: How We Can Eliminate Barriers to Economic Security and Mobility for People with Criminal Records,” Center for American Progress (December 2014), 4. https://cdn.americanprogress.org/wp-content/uploads/2014/12/VallasCriminalRecordsReport.pdf. 45 Changing in sentencing laws and policies are responsible for the rise in mass incarceration and hyper-criminalization. Vallas and Dietrich, “One Strike,” 5

19

3) Treatment delivered in the community costs approximately $20,000 less than imprisonment per person per year.46 Focusing the message on the key points—and refining each of those points using BE tools such as visual representations, detailed illustrative narratives, social norms, loss aversion, etc.—makes these points easier to learn and remember. Social Norms: “People like you do __.” Fortunately, there are several ways in which public opinion supports the movement to reform drug policy. Advertising these popular opinions can increase their popularity via our tendency to follow the herd. It is important to tailor the norms advertised as specifically as possible to the composition of the target audience (e.g. Democrats, New Yorkers, young professionals) because people align their behaviors with groups with whom they identify. Some potentially useful norms follow:

84% of American voters support non-prison alternatives such as drug treatment, community service, or probation for drug and other “victimless” offenses.47

Many American majorities believe too many people are in prison —64% of Democrats, 59% of African Americans, and 58% of Latinos.48

Majorities of American groups would prefer that more money and effort go toward better education and job training, attacking the social and economic problems that underlie crime, instead of toward deterring crime with more prisons, police, and judges—78% of Democrats, 77% of those aged 18-29, and 72% of college post graduates.49

These are merely examples. Majority opinions of peer groups are contagious, so any current statistic that reflects a majority opinion worth spreading can be used as a tool. Loss Aversion: Subtract from a Reference Point People convicted of non-violent crimes such as drug offenses face lifelong barriers to employment, housing, education, and economic security.50 These hardships not only

46 Justice Policy Institute, “Pruning Prisons: How Cutting Corrections Can Save Money and Protect Public Safety.” (May 2009) http://www.justicepolicy.org/images/upload/09_05_rep_pruningprisons_ac_ps.pdf 47 The Opportunity Agenda, “An Overview,” 33, C. Lake, D. Gotoff, and K. Pultorak, “Reducing Incarceration Levels in the U.S.: Opportunities for Reform,” Open Societies Foundation (2013). 48 The Opportunity Agenda, “An Overview,” 31, citing Lake, Gotoff, & Pultorak, 2013. 49 Ibid., 25, citing Hindelang Criminal Justice Research Center, Sourcebook of Criminal Justice Statistics, University of Albany (2013). 50 Vallas and Dietrich, “One Strike.”

20

affect the incarcerated and formerly incarcerated individuals facing an uphill battle, but they also damage our society as a whole. Framing the consequences of drug policy by identifying losses— in community safety, the economy, and other valuable societal assets— heightens audience concern. The following examples illustrate the use of loss aversion in promoting concern for drug policy reform:

Post-incarceration employment has powerful anti-recidivism effects.51 By keeping it legal for employers to broadly discriminate against individuals with past drug arrests or convictions, we are increasing the number of repeat offenses in our country.

Using incarceration as the one-size-fits-most response to drug offenses is an enormous financial burden on our state, local, and federal governments. In 2010 we lost more than $80 billion to corrections expenditures.

Using criminal records as employment barriers increased the national unemployment rate by 0.9% in 2008. The unemployment of formerly incarcerated people resulted in an estimated GDP loss of $65B per year.52

There is a long list of drug policy flaws that can be framed in terms of loss to society. What does society lose when we prevent families from reuniting because of housing barriers for individuals reentering society? At what cost do we incarcerate drug users rather than provide mental health and substance abuse treatment? How do we hurt our communities when we economically entrap former prisoners whose child support debt grew while they were behind bars? We are shooting ourselves in the foot with current drug policy, and we can make that clear by linking current practices to the specific economic and social costs we are all paying.

51 Ibid., 10 n48. 52 Ibid., 10 n49 citing John Schmitt and Kris Warner, “Ex-offenders and the Labor Market” (Washington: Center for Economic and Policy Research, 2010), at http://www.cepr.net/documents/publications/ex-offenders-2010-11.pdf.

21

Racial Profiling Racial profiling (i.e., enforcing laws against people based on their race/ethnicity rather than on evidence of wrongdoing) violates core American values. Americans might not oppose racial profiling for at least two reasons. One is that they are not convinced that it really exists, and they think that so many more minorities than whites are arrested simply because that many more minorities commit crimes. Another is that they may not understand what the harm is in racial profiling, even if they acknowledge its existence. Messages about racial profiling should thus focus on convincing people that 1) racial profiling happens and 2) racial profiling is harmful to us all. Identity: Offer an Attractive Label Equality is one of the core values on which our nation was founded. In theory at least, our tolerance of diversity sets the United States apart from other countries. So even though we all have biases beyond our control, most Americans at least aspire to not be perceived or labeled as “racists.” Yet racism, whether individual and overt or structural and subconscious, causes racial profiling. There thus exists a messaging opportunity to gently leverage audience aversion to racism to promote a stance against racial profiling. As discussed in Part I, identity research shows that a person’s behavior tends to conform to a self-relevant noun—like “voter” or “consumer”—with which she can identify. Thus the way to combine Americans’ idealistic aspirations of equality with individuals’ tendency to align their behavior with a relatable label is by offering people a label that puts them squarely in the category of people who do not condone racial profiling. If people will indeed choose a label that is the opposite of “racist” when given the choice, it makes sense to offer the choice. Whereas “voter” and “consumer” are objectively defined terms with little political charge, an analogous term doesn’t exist to describe someone who opposes racism. “Non-racist” is pithy, but by implicitly condemning non-adopters, it is also inflammatory. Picking a political fight would detract from the progress of encouraging behavior that counteracts racial profiling. So, for lack of a better term, the wordy but neutral “supporter of racial equality” is used in the following examples. Posing the label in a prospective rather than retrospective way gives the audience a chance to opt in.

Are you a supporter of racial equality?

The petition for anti-discriminatory practices seeks signatures from supporters of racial equality.

The shooting of another unarmed black man is expected to draw criticism and protests from supporters of racial equality.

Supporters of racial equality will likely back the proposed legislation.

22

In sum, when delivering messages designed to garner opposition to racial profiling, offer a self-relevant noun against such profiling to which your audience can subscribe. Availability Bias: Tell a Story Given the mainstream media’s coverage of crime, it is not difficult to understand how the average American might believe a black person is far more likely to engage in criminal activity than a white person. The media tends to portray blacks and Latinos as criminal and violent,53 and local news disproportionately covers violent criminals who are non-white and male.54 Thus depending on the information to which they are exposed, people may perceive the amount of law enforcement used against minorities as a logical consequence of misbehavior rather than an abuse of justice. Availability bias can be used to combat misconceptions about racial profiling in two ways. One is highlighting stories in which racial profiling obviously occurred—examples of police targeting a minority individual despite the absence of evidence linking that person to criminal activity. Such examples reinforce the notion that minorities are unfairly targeted. Using details such as names and places will make the example a memorable story rather than an abstract concept.55

Pauline Hampton and her niece, both black, were shopping at Dillard’s with their three children. After buying an outfit for one child, they went to the cosmetics counter to redeem coupons they had received with their purchase. A security guard who had been watching them since they arrived approached them, took the shopping bag, searched it, and returned it after seeing that the items matched the receipt. When Ms. Hampton complained about the guard’s behavior, he threatened to have the police remove her from the store.

31-year-old Jonny Gammage, a black man, was pulled over while driving his cousin’s Jaguar in a white neighborhood because he repeatedly braked to slow his car while an officer followed him. Backup arrived, and a scuffle broke out after Gammage exited the car holding his phone and datebook. After the officers forced Gammage to the ground and handcuffed him, he died from asphyxiation due to pressure on his chest and neck.

53 The Opportunity Agenda, “An Overview,” 18, citing R.M. Entman and K.A. Gross, “Race to Judgment: Stereotyping Media and Criminal Defendants.” Law and Contemporary Problems: 93–133 (2008). 54 Ibid., 18, citing F.D. Gilliam & S. Iyengar, “Prime Suspects: The Influence of Local Television News on the Viewing Public.” American Journal of Political Science: 560–573 (2000). 55 The following three examples are taken from the American Civil Liberties Union, “Racial Profiling: Definition.” https://www.aclu.org/racial-justice/racial-profiling-definition

23

Amadou Diallo, a 22-year-old immigrant from New Guinea, was standing outside his apartment building when four plain-clothes officers approached him. Diallo ran into his building vestibule, and the police fired 41 shots, killing him. Diallo was unarmed and had no criminal record.

The more easily a person can think of examples of racial profiling, the more frequently she estimates it to occur. If every American were convinced that police unfairly target people of color, the quest to turn everyone against racial profiling still would not be over. People would remain who might logically ask, “Even though it’s bad to target innocent people, aren’t black men primarily responsible for American crime given their proportion of prison population? And doesn’t it therefore make sense to continue being extra vigilant with black men?” In fact, white men are responsible for more crime than the media would have us think. This brings us to the second, more controversial, way to use availability bias to promote opposition to racial profiling—populating an audience’s memory with cases in which white individuals perpetrated crimes. Whereas black people are far more likely to go to prison for drug offenses, white Americans are more likely than blacks to have used most illegal drugs.56 With few exceptions,57 the white men responsible for the 2008 financial crisis got off scot-free. Thousands of mostly white college students vandalized Keene, New Hampshire after the town’s 2014 pumpkin festival, sending at least 20 people to the hospital.58 Most school shooters are white males.59 By profiling stories of white offenders—again, with details that make them memorable—it is possible to erode the notion that people in prison are mostly black because most criminals are black. But given the public’s already-inflated perception of crime frequency and the desire to

56 This is true for cocaine, marijuana, and LSD. Saki Knafo, “When It Comes to Illegal Drug Use, White America Does the Crime, Black America Gets the Time,” The Huffington Post (September 13, 2013). http://www.huffingtonpost.com/2013/09/17/racial-disparity-drug-use_n_3941346.html 57 For example, Jesse Eisinger, “Why Only One Top Banker Went to Jail for the Financial Crisis.” The New York Times Magazine (April 30, 2014). http://www.nytimes.com/2014/05/04/magazine/only-one-top-banker-jail-financial-crisis.html?_r=0 58 Justin Baragona, “Coverage of Pumpkin Fest Riot Compared to Ferguson Protests Exposes Media’s Overt Racism.” Politicus USA (October 19, 2014). http://www.politicususa.com/2014/10/19/coverage-pumpkin-fest-riot-compared-ferguson-protests-shows-medias-overt-racism.html 59 Tiffany Xie, “Mass Shooters Have a Gender and a Race.” Political Research Associates (June 19, 2014). http://www.politicalresearch.org/2014/06/19/mass-shooters-have-a-gender-and-a-race/

24

promote equal opportunity without disparaging any group of people, this technique is one that many groups may be loath to use. CONCLUSION The BE principles described in this paper are powerful tools, but none of them can be understood in a vacuum. The researchers discussed earlier who wanted to reduce opposition to vaccines used loss aversion by describing the horrible symptoms of the diseases vaccinated against and availability bias by profiling suffering children, but they failed to change minds until they catered to the audience’s identity. On the contrary, OPOWER succeeded in changing behavior using social norms, but would they have done so if their Home Energy Reports had been communicated to customers by emails or phone calls instead of in periodically mailed envelopes that customers were in the habit of attending to? There is tension and interplay among the various principles. It would be unwise to view them in isolation. Further, the applications of the principles are dynamic. Their success depends on the relationship between the audience and the message. People in an audience may change how they identify themselves, what they value, and what comprises their memories. Thus knowing an audience is instrumental to a messaging policy’s success. Finally, it is only when you get accurate feedback that you can know whether your message is working. Thus try to measure the effects of your communications whenever possible and tweak as needed. A lot of factors are in play when it comes to changing minds, and it may not be possible to anticipate all of them. By combining knowledge about the basic BE principles with some trial and error, you can communicate more effectively.