Our results, published in the Proceedings of the National Academy of Sciences, show that competitive interactions tend to make people’s behavior slightly less moral, and offer some intriguing clues as to why this might be the case.
We are not the first to take a scientific approach to the question of competition and morality. However, individual tests have returned mixed results, possibly due to differences in the definitions and measures of morality used. Some of the early results were provocative, such as the finding that people in competition were less likely to avoid the death of a mouse. However, these results were difficult to replicate or interpret.
One way to account for differences in the design of individual studies is to perform a ‘meta-analysis’, evaluating and combining the results of many different studies. However, meta-analysis often has its own problems, as selective reporting and publication bias can influence which studies are available for inclusion in the analysis.
What was different in our study:
To really get some reliable results, we went a step further and carried out a ‘prospective meta-analysis’. The “prospective” part means that all the studies to be included in the analysis were registered before they were carried out. This avoids selection of results or bias in the type of results that are published.
Our project involved 45 different experiments carried out by teams from all over the world. Each team independently designed an experiment to test the effects of competition on morality. The results of these studies, which involved observations of more than 18,123 individual participants, were then collated and analyzed.
A small decrease in morality (on average)
The meta-analysis revealed that competition has an overall negative effect on morality, but the effect is very small. (The effect is measured by a number called Cohen’s d. A value of 0.2 is considered a small effect, and the value we found was only 0.1.) As expected, we also observed substantial variation in the effects measured by different experiments. Some were positive, some were negative, and the effect sizes also varied.
So despite the advantages of our new prospective meta-analysis, the jury is still out on the overall effect of competence in morality.
Perhaps the question is too general to answer correctly without particular context. The devil can be in the details. Is the loss, not the competition, to blame? My team (one of 45 involved in the meta-analysis) used a two-person guessing game as the instance of competition. This was followed by an individual game of honesty, which was our measure of the effects on morality.
This individual experiment resulted in a small overall negative effect of competition (d = -0.1) very similar to the meta-analysis, but failed to reach statistical significance on its own. However, an exploratory analysis of our results revealed a potential breakthrough. We found that only the losers of the number-guessing game became more dishonest, with a larger effect (d = -0.34). The winners of the competition stage, on the other hand, showed no change in their honest behavior.
These exploratory results, yet to be replicated, suggest a reason why competition does not affect morality much on average. Perhaps it is being at a disadvantage in a competitive process that corrupts, not competition per se.