We like to believe that we’re rational creatures, but we’re not.  In The Honest Truth About Dishonesty: How We Lie to Everyone – Especially Ourselves, Dan Ariely explains what we know about managing our dishonesty in ways that allows us to still believe we’re good, honest people.  This continues on the work in Predictably Irrational.  It follows in the line of Daniel Kahneman’s work in Thinking, Fast and Slow and the work of Robert Cialdini in Pre-Suasion and Influence.

Economics and Criminal Justice

Economics isn’t the study of money.  Likewise, criminal justice isn’t really the study of crime.  Both are really about understanding human behavior.  Economics is human behavior related to money and criminal justice is about human behavior related to crime.

The strict economic model fails to explain why, in the ultimatum game, people would elect to get nothing – as long as they can ensure the other participant gets nothing as well.  (See The Evolution of Cooperation, The Selfish Gene, SuperCooperators, and Does Altruism Exist? for more on why this might develop and The Righteous Mind for how justice is a foundation of morality.)  The ways that we’re willing to accept a cost to ourselves breaks that standard self-interest economic model.

There’s a similarly naive model called the Simple Model of Rational Crime (SMORC).  It posits a cost-benefit analysis that includes the probability of getting caught and the degree of penalty placed against the backdrop of the value of the crime.  The problem is that we know this doesn’t work.  There isn’t a cost-benefit analysis in effect.

To understand why this isn’t true, consider the work of Gary Klein with firefighters and the realization that they don’t make decisions based on some sort of normal, numeric approach to the problem.  Instead, they use recognition-primed decisions to simulate what’s happening and make a working solution.  (See Sources of Power.)  To believe that we make rational decisions based on the impact of getting caught ignores what we learned about D.A.R.E. (see Unsafe at Any Speed) and Scared Straight! (see Science and Pseudoscience in Clinical Psychology) being potentially harmful.

What Leads to Cheating

In The Ethics of Encouraging Dishonesty, I explained some of the things that we know about what causes more cheating.  Ariely expands this to explain how decreasing the chances of getting caught – by eliminating the checks and disconnecting the cheating from direct financial reward – will increase cheating.  However, he also reports curious responses to testing.

One would assume if the reward were higher, there would be more cheating.  However, that’s not what Ariely found.  He found that the degree of cheating went down when the reward went up.  The proposed reason is that people need to feel like they’re honest, so their cheating has to be considered small – by them.  The larger the reward, the harder it becomes to maintain this perspective.

Also, Ariely found that if someone were cheating to help someone else – i.e. they’re partners, and the cheating will help the other person as well – that cheating increased.  Obviously, neither of these results make sense when we evaluate the cheating from a rational model – but they were the results of the testing.

The Cashless Society

Ariely expresses some concern that the move to a cashless society may increase cheating; in his research, when there was an intermediary to money, the cheating increased.  If tokens were exchanged for money, they were seen as somehow not the same as cheating for the money directly.  Certainly, the financial scandals that we’ve seen from organizations (see Moral Disengagement) and the home loan crisis (see The Halo Effect) support Ariely’s concerns.

Grandmother Mortality

Mike Adams is concerned for your grandma while you’re going to college.  He collected data over the years and ultimately demonstrated that grandmothers are ten times more likely to die before a midterm and nineteen times more likely to die before a final exam.  The odds are even worse if the college goer is not doing well.  Students who are failing a class are fifty times more likely to lose a grandmother when compared to non-failing students.

His work is intended to highlight the lying that goes on – and how, while under pressure, students are more likely to make the grandmother death claim.  The obvious sampling error is that Adams data was sourced from student communications – which almost always came with a request for an accommodation.  If grandma dies at a time when the student doesn’t need an accommodation, he wouldn’t get the note.

Still, it might be worth some added protection for grandma if she has a grandchild that’s struggling in their college coursework.

Self-Signaling

The way that we signal ourselves matters.  If you think that you’re wearing genuine designer clothing and accessories, you’ll behave differently than if you believe you’re wearing a knockoff.  It’s not uncommon to find people selling knockoffs of designer clothing and accessories in large cities across the world.  I’ve personally seen them in New York and San Francisco.

The argument for those who buy the knockoffs is that they’re not harming anyone because they can’t afford – and therefore would never buy – the real thing.  This ignores the fact that part of the reason for the high price is exclusivity – something the knockoffs deteriorate.  However, the real problem isn’t the knockoff itself even when considering the brand effects.  The real problem is that it makes you more likely to cheat overall.

Rationalization

In Jonathan Haidt’s view, our consciousness is like a press secretary explaining our behaviors post-hoc.  (See The Righteous Mind and The Happiness Hypothesis.)  That’s make sense when you consider that Richard Nisbett and Tim Wilson laid out four identical stockings and then asked people which ones they liked better.  People made up lots of reasons why they liked one pair over the other – despite the fact they were identical.

In a set of experiments, Michael Gazzaniga studied patients whose corpus callosum was severed.  He found they’d make up stories when shown images in their left eye (processed in the right side of their brain).  The left side of the brain initiated the story creation process and fabricated a post-hoc reason for behaviors triggered by the right side of the brain.  (See The Blank Slate, Noise, and Incognito.)

Gangs

Ariely states, “The act of inviting our friends to join in can help us justify our own questionable behavior.”  This squares with the research of Albert Cohen in Delinquent Boys and Cas Sunstein in Going to Extremes.  The presence of others tends to reinforce and amplify our behavior choices.

Essay Mills and Generative AI

At the time of Ariely’s writing, academia was concerned about the introduction of essay mills.  For a few hundred dollars, one could ask for a paper to be written about basically any topic.  Ariely tested this process by asking for a 12-page paper on the topic of dishonesty.  Ariely’s team paid a few places between $150 and $216 and the results weren’t impressive – such that they decided they weren’t a concern for academia at the time.

Today’s concern is the impact of generative artificial intelligence (GenAI) in the form of large language models.  I fed a slightly modified prompt from Ariely’s original instructions (as recorded in the book) to Bing’s Chat function (based on GPT-4) and received a response that it couldn’t help – but it proceeded to provide a quite useful outline.  I then provided the same prompt to Microsoft Word’s Copilot function.  The result was a 10-page paper – with a few problems.

On the first attempt, it failed to generate, ironically after displaying a reference to The Honest Truth About Dishonesty.  The second attempt, which generated the 10-page paper, also failed to get the references right.  It didn’t make them Word references, and it didn’t supply a complete reference for Leon Festinger’s 1957 title, A Theory of Cognitive Dissonance.  It also generated a reference to Albert Bandura’s work on moral disengagement – but the specific reference wasn’t relevant to the topic.  (However, Bandura’s Moral Disengagement might have been.)

That being said, academic institutions have a valid reason to be concerned about the degree to which they can measure a student’s understanding based on papers.  GenAI solutions can take a lot of the burden off and leave it to the student to do a bit of cleanup instead of a lot of writing.

Congo Is Watching

In the 1995 movie, Congo, the viewer is exposed to the idea that gorillas were watching as other gorillas were mining diamonds.  The idea that someone is watching matters for more than diamond mining.  It matters for honor boxes.  That’s what Ariely and colleagues found as they alternated pictures above an honor box setup between flowers and eyes.  When the pictures of eyes were up, the honor box contained three times more money.

We don’t have to have real people watching us – we just need to be reminded that we could be watched to trigger our honest behavior.  Maybe if you believe you’re being watched, you’ll find The Honest Truth About Dishonesty.