Extremes in spite of us all

According to a study appearing in the latest edition of the open access journal PLoS ONE, displays of spitefulness tend towards extremes where spiteful behaviour is defined in terms of deliberately harming another at no cost to oneself.

In an experiment carried out with 48 student participants, the report’s authors found that, while many people will actively avoid being spiteful at all, those who are spiteful take this behaviour to the extreme by imposing the maximum possible amount of harm. As the experimental technique allowed the team to measure spitefulness, they could demonstrate that the degree of spite exhibited by an individual was largely consistent over time.

Dr Erik Kimbrough, an Assistant Professor in the Department of Economics at Simon Fraser University in Canada, co-authored the paper with Dr Philipp Reiss from Maastricht University in the Netherlands. Funding for the study was provided under Maastricht’s METEOR programme of research in economic behaviour, theory and computing. Dr Kimbrough took the time to answer ScienceOmega.com’s questions about their findings…

What were the motivations behind your study? How does it build on previous similar studies?
Actually, when my co-author Philipp Reiss and I designed this study, our goal was to understand an entirely different phenomenon. We were trying to explain overbidding in auctions relative to a simultaneously available fixed-price option (like a Buy-It-Now price on eBay), so we happened on the spitefulness data almost by accident.

The experiment has participants acting as agents for the experimenter in a series of auctions. We tell them, effectively, "I'm willing to pay you $10 to buy this item. If you can buy it for less, I'll give you $10 and you can keep the difference". In these auctions, the winner is the person who submits the highest bid, but he/she has to pay a price equal to the second highest bid.

The ‘second price auction’ is basically a simplified version of how eBay runs its auctions, and it has been studied extensively in economics. Our innovation was to create second-price auctions that have two stages of bidding, so all the bidders submit an initial bid, and then before the final bids are submitted, we informed everyone of the current highest bid. In a second-price auction, this meant that people could see the bid that was currently winning the auction, and knowing that the price was determined by the second highest bid, they could increase their bid without going over the current high bid in order to raise the price paid by the winner.

This is what we call spite: intentionally raising the price paid by others, at no cost to yourself. And because we know each bidder's initial bid and the high bid, we know the range of bids that could be considered spiteful, so we can actually quantify the extent of spitefulness, roughly measured as the ratio of how much someone actually raised the price divided by how much they could have raised it. We introduced the two stages of bidding because of a model we had in mind of why some people might overbid relative to a posted price, but what it actually allowed us to identify was totally different.

Were you surprised by what your experiment found?
Absolutely. Getting exciting data unintentionally is a pretty wild feeling; I suspect most scientists would agree that it's much more common to get boring data unintentionally – this was definitely a first for me. As far as the actual results, we were really struck by how the data were clustered in the extremes.

Why do you think the results showed such polarised extremes of spitefulness and non-spitefulness?
The honest answer is, I don't know. Our experiment provides a way of clearly measuring the extent spitefulness, and that is our big innovation. Our data don't really speak to the motivations people had for their choices. We asked participants to describe their bidding strategies in writing after the experiment, and some of the spiteful bidders claimed that they had been trying to punish others for bidding too high or teach them a lesson of some sort. The problem with this argument is that there was no way for this ‘punishment’ to alter future behaviour because of the way our experiment was designed. All the first stage bids were submitted for all the auctions before we revealed any information and before any of the final bids were submitted.

What are the implications of this outcome, particularly that finding that an individual’s level of spite is consistent over time?
I suspect my answer is going to be unsatisfyingly conservative, but the implication, to me at least, is that we now have a good way of measuring the extent of someone's spitefulness. To what end this measurement is used remains to be determined, but I hope that future work will explore how spitefulness is related to other important behaviours and how sensitive it is to things like context, information about the others you are harming, and so on.

What are the next steps for your research?
For one thing, my co-author and I now plan to deliberately design an experiment that allows us to measure spitefulness in a similar manner, but without some of the other design features that were only part of this experiment because of our initial research question. Our measurements show a lot of internal consistency as is, but with an experiment designed expressly for the purpose of identifying spite, we could get an even clearer picture of individual spitefulness.

Recently my research has focused on classifying what economists call ‘player types’ in economic decision-making – that is, people who make decisions in specific ways (eg conditional cooperators who will attempt to cooperate in social dilemmas, as long as others do too). Since our spitefulness metric seems to be consistent over time, I plan to add it to the battery of tests we can use to identify types. Then I can ask questions like, "how does the behaviour of a group composed of spiteful 'types' compare to a group composed of non-spiteful 'types' in cooperative environments or competitive environments?" (I put 'types' in quotations, because I want to be clear that I don't necessarily think these are immutable categories.)


Read the full text of the research paper here: 'Measuring the Distribution of Spitefulness'.