“Could switching to Geico really save you 15 percent or more on car insurance? Was Abe Lincoln honest?” So intones the Geico commercial spokesperson, followed by faux vintage film footage of Mary Lincoln asking her husband, “Does this dress make my backside look big?” Honest Abe squirms and shifts, then hesitates and, while holding his thumb and forefinger an inch apart, finally mutters, “Perhaps a bit,” causing his wife to spin on her heels and exit in a huff.

The humor works because we recognize the question as a disguised request for a compliment or as a test of our love and loyalty. According to neuroscientist Sam Harris in his 2013 book Lying (Four Elephants Press), however, even in such a scenario we should always tell the truth: “By lying, we deny our friends access to reality—and their resulting ignorance often harms them in ways we did not anticipate. Our friends may act on our falsehoods, or fail to solve problems that could have been solved only on the basis of good information.” Maybe Mary's dressmaker is incompetent, or maybe Mary actually could stand to lose some weight, which would make her healthier and happier. Moreover, Harris says, little white lies often lead to big black lies: “Very soon, you may find yourself behaving as most people do quite effortlessly: shading the truth, or even lying outright, without thinking about it. The price is too high.” A practical solution is to think of a way to tell the truth with tact. As Harris notes, research shows that “all forms of lying—including white lies meant to spare the feelings of others—are associated with poorer-quality relationships.”

Most of us are not Hitlerian in our lies, but nearly all of us shade the truth just enough to make ourselves or others feel better. By how much do we lie? About 10 percent, says behavioral economist Dan Ariely in his 2012 book The Honest Truth about Dishonesty (Harper). In an experiment in which subjects solve as many number matrices as possible in a limited time and get paid for each correct answer, those who turned in their results to the experimenter in the room averaged four out of 20. In a second condition in which subjects count up their correct answers, shred their answer sheet and tell the experimenter in another room how many they got right, they averaged six out of 20—a 10 percent increase. And the effect held even when the amount paid per correct answer was increased from 25 to 50 cents to $1, $2 and even $5. Tellingly, at $10 per correct answer the amount of lying went slightly down. Lying, Ariely says, is not the result of a cost-benefit analysis. Instead it is a form of self-deception in which small lies allow us to dial up our self-image and still retain the perception of being an honest person. Big lies do not.

Psychologists Shaul Shalvi, Ori Eldar and Yoella Bereby-Meyer tested the hypothesis that people are more likely to lie when they can justify the deception to themselves in a 2013 paper entitled “Honesty Requires Time (and Lack of Justifications),” published in Psychological Science. Subjects rolled a die three times in a setup that blocked the experimenter's view of the outcome and were instructed to report the number that came up in the first roll. (The higher the number, the more money they were paid.) Seeing the outcomes of the second and third rolls gave the participants an opportunity to justify reporting the highest number of the three; because that number had actually come up, it was a justified lie.

Some subjects had to report their answer within 20 seconds, whereas others had an unlimited amount of time. Although both groups lied, those who were given less time were more likely to do so. In a second experiment subjects rolled the die once and reported the outcome. Those who were pressed for time lied; those who had time to think told the truth. The two experiments suggest that people are more likely to lie when time is short, but when time is not a factor they lie only when they have justification to do so.

Perhaps Mary should not have given Abe so much time to ponder his response.