Limits of the Effect of Social Information on the Voluntary Provision of Public Goods: Evidence from Field Experiments

Article excerpt

Social scientists of all ilks have examined why individuals make charitable contributions and voluntarily provide public goods. In economics, models of altruism (Becker 1974), warm-glow and warm-glow altruism (Andreoni 1989, 1990), conditional cooperation (Fishbacher, Gachter, and Fehr 2001), and reciprocity (Sugden 1984) have been developed and tested in the experimental laboratory (Croson 2007).

Recently, field experiments have been introduced as a research tool in studying public goods provision and charitable contributions in economics. (l) For example, List and Lucking-Reiley (2002) find that increasing the proportion of seed money in fundraising increases participation rates and the average amount contributed. Eckel and Grossman (2005) find that matching contributions and rebates solicit about the same number of contributions, but that matching generates higher amounts contributed. Karlan and List (2007) show that while matching contributions increase donations, the exact amount of the match is irrelevant. These articles change the financial decision facing the potential donor.

A second stream of literature investigates the social situation facing the donor. Falk (2007) uses gift exchange and reciprocity to increase giving. Landry et al. (2006) show that the attractiveness of the solicitor significantly increases giving in door-to-door fundraising.

Still other literature directly demonstrates the impact of social information, knowing whether and what others will contribute, on one's own contribution. Frey and Meier (2004) show that as more individuals contribute, the probability of a target individual contributing is increased. Shang and Croson (forthcoming) show that as other donors give more to a public radio station, the target donor also gives more. Croson and Shang (2008) show that social information can be used to both raise and lower contributions made via mail as well as on the phone. Shang, Reed, and Croson (2008) demonstrate that the social similarity between the other who has made the contribution and the target donor affects the strength of the social information effect. Martin and Randal (2008) show that individuals donate more when a museum donation box contains higher-denomination bills. Chen et al. (2010) show that below-median contributors significantly increase their volunteering activities when informed of the median activity of others. (2)

While this previous work demonstrates that social information can affect individual contributions, in this experiment we identify the boundary conditions of the social information effect. We ask the question: at what point will social information cease to be effective? Our setting is an on-air fundraising campaign for a public radio station. We begin with some background, describe our experiment and its results, and finally conclude.

I. BACKGROUND

In this article, we examine the limits of the influence of social information on giving behavior. While previous research has demonstrated that higher social information increases contributions (and lower social information decreases them), one might imagine a limit to this effect. At the limit, if a non-profit organization has received an especially large donation, the targeted donor may conclude that the organization is sufficiently funded so as to not need their money. At more moderate levels (like the ones we test), large donations may simply be disregarded by potential donors as "irrelevant" for their own decisions. Indeed, research from psychology and social influence suggests that in order for social information to influence behavior, the information must be judged relevant by the donor (Cialdini 1998). We hypothesize that as the donation of another party increases, the judgment of its relevance (and thus its effectiveness) decreases.

To test this conjecture we ran a field experiment with a public radio station. In public radio, each individual has an incentive to free ride, listen to the station, and not contribute to its continued functioning. …

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.