What Darkness Does to the Mind

In the summer of 2008, I moved from Pittsburgh to Chapel Hill to start my new position as a faculty member at the business school at the University of North Carolina. Although I was sad to leave Carnegie Mellon and my colleagues there, I was excited to meet new ones and to move into our new home. A few months earlier, my husband Greg and I had bought a lovely house surrounded by quiet, leafy streets just a few blocks away from the center of town.

Within a few days of moving in, Greg and I received a letter from Chapel Hill's City Hall welcoming us and informing us that new street lighting would be added in the neighborhood in the follow­ing weeks since that part of town had recently experienced a surge in crime. In addition to raising my fears (and not making me feel any safer), the letter also piqued my curiosity, since it highlighted an intriguing assumption: that lighting would reduce crime.

In a sense, this assumption was consistent with what Ralph Waldo Emerson once wrote: "As gaslight is the best noctur­nal police, so the universe protects itself by pitiless publicity." According to conventional wisdom, darkness conceals identity and also decreases inhibitions; as a result, it may be linked to crime. The idea that darkness promotes unethical behavior dates back to the myth of the "Ring of Gyges," which was recounted by Plato in The Republic (360 BC). In the myth, a shepherd in Lydia named Gyges finds a ring that makes him invisible. He travels to the king's court, seduces the queen, conspires with her to kill the king, and takes control of Lydia. Thus, invisibility corrupted the wearer of the ring. The story leads Plato to ask the following ques­tion: is there anyone alive who could resist taking advantage of the invisibility ring's powers, or is it only others' monitoring that prevents us from committing immoral acts?

From this perspective, by providing anonymity, darkness may facilitate dishonest behavior. When transgressors believe others will not be able to identify them, are they more likely to behave dishonestly? Scholarly work conducted in the 1960s and 1970s found that criminal assaults most frequently occur during hours of darkness and that improving street lighting in urban areas is commonly followed by reductions in crime of between 33 percent and 70 percent -- impressive gains. Although interesting, the sci­entist in me notes that this evidence is inconclusive, as the rela­tionship between darkness and crime suggested by this data could be explained by other factors. I wondered whether there is a direct relationship between darkness and crime rates. Even more inter­estingly, does darkness increase dishonesty?

Soon after Greg and I received our letter from City Hall, Chen-Bo Zhong (a professor at the University of Toronto), Vanessa Bohns (a professor at the University of Waterloo), and I designed a series of experiments to test whether darkness -- or even dim lighting -- would increase dishonesty.

Chen-Bo, Vanessa, and I tested this possibility by conducting an experiment where we manipulated darkness by varying the level of lighting in rooms. Upon arriving at our laboratory, our eighty-four student participants were randomly assigned to one of two rooms (with about half in each room): one of them was well lit (our control condition); the other one was similar in size but was dimly lit (specifically, lit by four fluorescent lights rather than by twelve). Participants in the dim room could see the materials and one another, but the room was more dimly lit than the aver­age room at a university. Participants completed a problem-solving task: they had five minutes to solve twenty problems (which involved finding two three-digit numbers that add up to ten in a matrix of twelve numbers), and were paid 50 cents for each problem they solved correctly. After the five minutes was up, participants in both conditions were asked to self-report their performance on the problem-solving task. They were able to lie by overstating their performance and thus walk away with undeserved money. As in other experiments concerning dishonesty, we tracked whether participants cheated and, if so, by how much. If you were a participant in this experiment, do you think you would cheat by overstating your performance?

Maybe you would stay true to your moral compass. But, as it turns out, many of our participants did not: in fact, on average, about half of them cheated across conditions. More interestingly, the level of darkness in the room dramatically influenced partici­pants' likelihood to lie by overreporting their performance: almost 61 percent of the participants in the dim room cheated, while "only" about 24 percent of participants in the well-lit room cheated. In other words, eight additional fluorescent lights reduced dishonesty by about 37 percent. This is quite a large difference, especially con­sidering that the task Chen-Bo, Vanessa, and I used in the experi­ment was completely anonymous: the only difference between the two rooms was the level of darkness.

These results were consistent with our initial predictions, but we wanted to take them a step further. We reasoned that, beyond simply producing conditions of actual anonymity, darkness may create a sense of what we refer to as illusory anonymity. This type of anonymity is likely to loosen inhibitions surrounding dishon­est behaviors such as lying and cheating. People in a room with slightly dimmed lighting, we reasoned, may feel anonymous not because the relative darkness has reduced others' ability to see or identify them (which it hasn't), but because they are anchored in their own experience of darkness. When people experience impaired vision as a result of darkness, they might unconsciously generalize that experience and expect that others will conversely find it difficult to perceive or see them, even when these others are sitting in a different location (such as another room). Just as small children close their eyes and believe that others can't see them, the experience of darkness, we theorized, would trigger the belief that we are warded from others' attention and inspections. Since peo­ple often have a myopic focus, this reasoning seemed to hold. If it's true, then manipulating darkness in other, more subtle ways than reducing ambient lighting likely would have the same effects on ethical behavior that we observed in our first experiment.

For our next experiment, we invited eighty-three students from the University of North Carolina at Chapel Hill to participate in an experiment for which they would receive a $5 show-up fee and a potential bonus payment of $6. Half of the participants were asked to wear a pair of sunglasses, and the other half were asked to wear glasses with clear lenses. They were then assigned to work with someone they were told was another participant (but was actually the experimenter) in a different room. They would be working with this person by communicating through computers. Participants knew that they would not interact face to face with their partner, nor would they later learn their partner's identity.

Clearly, when you are wearing a pair of sunglasses, no one else's sight is affected, especially when you are not looking at each other. Nonetheless, we expected that the relative darkness caused by wearing sunglasses would trigger a sense of illusory anonymity and influence participants' dishonest behavior. We measured dis­honesty by examining how selfish people were in allocating a sum of money between themselves and their partner.

Each person had $6 to divide between him- or herself and the recipient. The recipient had no choice but to accept the offer, and participants were told they could leave with the money they kept for themselves. Although we told participants that they had been randomly assigned to a role (either initiator or recipient), they all played the initiator against the experimenter. After participants made their choice, they answered a few questions measuring the extent to which they felt anonymous during the experiment.

Participants could offer any amount between $0 and $6. On aver­age, they offered $2.35, a bit less than a 50/50 split. Their offers differed based on whether they wore sunglasses: those who wore sunglasses gave less than $2, on average, while those who wore clear glasses offered an average of almost $3. Participants in the sunglasses condition gave significantly less than an equal division; those in the clear-glasses condition gave significantly more. As we predicted, wearing sunglasses also affected participants' psy­chological state: they felt more anonymous during the study than did those wearing clear glasses. Although darkness had no bear­ing on actual anonymity, it still increased morally questionable behaviors.

This is an excerpt from Sidetracked: Why Our Decisions Get Derailed, and How We Can Stick to the Plan.