Research demonstrates how much we distrust people who are mean with money

The study participants had no face-to-face contact but played a series of interactive games. They had to make decisions about whom to trust in their dealings with other players, based on information they were given on the level of these other players' generosity in previous games.

The experiments revealed that participants who had been mean with their money were trusted less, and indeed were more likely to be untrustworthy. The findings by researchers from the University of Oxford and the European University Institute, Italy are published in the journal PLOS One.

The experiments involved a total of 265 Oxford University students. First, participants played the Dictator Game, where they were put into pairs, and in each pair one was given £8. These players could then choose the 'mean' option of giving £1 or the 'generous' option of giving £3.50 to the other player.

In the second stage of the experiment, participants played the Trust Game. In this game, the players were again put into pairs but with a different partner. Then, one player could first decide whether to keep or send money to their partner. Sending the money would multiply the amount, but now their partner could either send a larger share back in return, or simply pocket the money. However, before the first player decided whether to send the money, he or she received information about what the other player had done in the Dictator Game. The players could truthfully reveal, hide or lie about whether they had been generous or mean.

Some 65% of participants in the Trust Game sent money to those who had already shown generosity in the earlier game. Only 29% of players sent money to partners after learning those partners had been mean or had remained silent about what they did in the earlier Dictator Game.

While the researchers informed some players about the rules of the Trust Game before they played the Dictator Game, others were kept in the dark. This allowed the researchers to assess whether players might be strategic in choosing to be generous in the first game.

The researchers aimed to discover the signals we rely upon to decide whom we can trust. They found that past displays of generosity are widely regarded as a key attribute when we are looking for indictors of trustworthiness and other information about a person's reputation is not available. The experiment also demonstrates how players recognised the usefulness of displaying past generosity. Around 60% of players who had been mean in the first game and got a chance to say it lied about their meanness in the later game, saying that they had been generous.

Researcher Dr Wojtek Przepiorka, from the Department of Sociology at Oxford University, said: 'When acts of generosity occur naturally with no concern for how they are perceived by others, they can be effective signals of trustworthiness. Charity balls are places where people can openly display their generosity, but in this case, because people know they are going to be observed, this might be a strategic gesture and less telling of their true character. We regard acts of genuine generosity as those produced spontaneously and these are widely seen as a reliable indicator of trustworthiness even when they are small gestures. '

Professor Diego Gambetta, a Nuffield College Official Fellow from Oxford University who is now on leave at the European University Institute, said: 'Our experiments showed that cheating comes in clusters -- a large portion of people who were mean were also prepared to lie about it, and those who lied were much more likely to be untrustworthy, as if one sin promotes another. It appears that people widely regard generosity and trustworthiness as being "cut from the same cloth" as far as human characteristics are concerned. However, some people display generosity when it is likely to be in their own advantage. In our experiments, we also find that those who are guilty of "strategic" displays of generosity themselves are more likely to spot strategic generosity in others.'