Dishonesty perpetuates everyday life, especially in the form of small cheating that happens on a massive scale. Think of using public transportation without paying, cheating on taxes or stealing from the workplace. Cheating can cause larger scale problems as it can undermine trust, which is well-known to be associated with e.g. economic growth.
But cheating is heavily context-dependent. People are more likely to cheat in certain situations than in others. The question that then arises is in what contexts cheating is more common, what settings lead to more dishonesty.
Houser et al. (2015) address this question via an experiment with parent-child pairs. Parents participate in a coin tossing experiment. They have to toss two blue-green colored coins, and if both coins end up green, they win a gift. The probability of winning is thus 25%. The coin toss is done privately, and the outcome of the toss is entirely self-reported.
There are two twists: (i) half of the parents have to do the coin toss in front of their child, and (ii) in half of the cases, the gift is for the child, in the other half it is for the parent.
The experiment was conducted in Chicago with a sample size of 152.
There are thus four possible scenarios that a parent-child pair can be assigned to as summarized in the table below.
A theory of dishonesty would predict that
- parents cheat more when they’re alone because they’re not subject to scrutiny, and
- parents cheat more when the gift is for the child because they then incur lower moral costs to cheating (one can more easily justify cheating for someone else’s benefits).
Indeed, the authors find evidence for both points. When the child is not in the room, the winning rate is 46% (as opposed to the theoretically expected value of 25%), and when the child is in the room, the rate is only 33%. This difference is statistically significant (p=0.09).
With regard to point (2), when the gift is for the child, the winning rate is 43%, and when it’s for the parent the rate is 36%. This result, however, is not statistically significant.
Going back to our original four scenarios, we can see below that in the context that is most conducive to cheating (parent alone, gift for child; Pa_GiftChild), the winning rate is extremely high at 58%. This is significantly higher than the winning rate in the other three scenarios.
Finally, there are clear gender differences in dishonesty. In fact, other research has shown that men are more dishonest than women. Why is this case? One reason may be that simply parents are more willing to cheat in front of their sons than their daughters. This is illustrated below.
And this difference arises only when the child is present. For some reason, parents do not incur as high of a moral cost for cheating in front of boys as they do for cheating in front of girls.