The tragedy of little Madeleine in 2007 did not only catch the headlines of the public media but was also followed with interest by researchers after the remarkably successful donation acquisition of the desperate parents. To finance a costly search for their taken daughter, they asked the British population for donations. Presumably unconsciously they benefited of the so called "identifiable victim effect" because society is willing to spend far more money to save the lives of identifiable victims than to save statistical victims (Jenny und Loewenstein, 1997). In the literature this effect has up to now been investigated in laboratory experiments and in surveys. Studies indicate that identifiable victims evoke more intense feelings than more inclusive ones. In our study we investigate the effect in a field experimental setting and are especially focussing on measuring support in terms of monetary donations. For the experiment we cooperated with a German charitable organization which promotes primary medical health care in five developing countries. Together with the organization we framed two versions of a solicitation letter that was sent to 57,325 households as part of the yearly winter mailing campaign. Within the "baseline"-group recipients were asked - as in previous mailings - to donate any desired amount to the organisation. Potential donors in the "choice"-group were beyond that given the possibility to select a particular country (or more countries) as donation recipient. In line with the effect of the identifiable victim we expect higher donations when one explicit object of benevolence can be chosen for a donation. The reference group can thus be reduced from the whole population of five countries to one particular country. Our hypothesis was that, although the victim is in this case not as identifiable as in the case of Madeleine, this weak identifiable victim effect could lead to higher donations. Overall, 6,709 study relevant donations were received by the organization in the observation period between December 3rd, 2007 and January 31st, 2008, adding up to an amount of more than 1 million Euro. The response rate was 11.7% for both groups. Within the "choice"-group, 3.4% of the donors made use of the selection possibility and donated to a particular country. Such donors donated with an average amount of 160 e significantly more than those donors who did not select a country for their donation (135 e). The organisation additionally provided us with data of their two previous winter mailing campaigns that allowed us to observe the donation behaviour of households over time. Under inclusion of the donation history we deduced that households that donated to a particular country in 2007 did not donate differently in previous years than did those donors who did not select any country for their donation in 2007. This supports our hypothesis that the different average amounts in 2007 stem from our treatment manipulation and not from random or selection biases. Our study emphasizes that charitable organisations can benefit from giving donors more precise information of how the donated money will be used. The probable reason is that altruistic action seems to be mediated by aroused empathetic emotions. People give in order to do something good for the victims and/or themselves. But the intensity of the emotions which might be expressed by the height of the donation seems to depend on factors that can be influenced, e.g. in terms of more detailed information with respect to the donation purpose.

Aretz, Bodo and Sebastian Kube (2010), Choosing Your Object of Benevolence – A Field Experiment on Donation Options, ZEW Discussion Paper No. 10-016, Mannheim. Download


Aretz, Bodo
Kube, Sebastian


charitable giving; identifiable victim; field experiment; altruism; contingent valuation