Charitable Giving under Normative Uncertainty: Experimental Evidence on The Behavioral Impact of Normative Expert Advice

“Normative Uncertainty” is a real thing.

Here is some interesting economic research and observations from two researchers across the big pond.  I read it quickly and then again carefully, but I remain normatively uncertain about what I should have learned:  

In this paper, we test both the relative importance of normative uncertainty compared to the well-researched descriptive uncertainty, as well as the impact that expert advice has in comparison to descriptive information. To investigate donor behavior under normative uncertainty, we ran an online experiment with a sample of 1890 US residents. We symmetrically compared the role of the well-established effect of descriptive uncertainty to that of normative uncertainty. In what follows, we use the terms ‘advice’ and ‘information’ to refer to normative advice given by experts and to descriptive information about empirical matters respectively, and use ‘(decision) aid’ to refer to both jointly. Normative uncertainty has substantial importance in real-world philanthropic activity, as donors (both individual and institutional) are plausibly always unsure where they should donate to. For example, Giving What We Can, a meta-charitable organization that aims to find the best charities available for its donors, has to deal explicitly with it. As an organization, they are thus faced with the same normative uncertainty as donors are: Which charities should be recommended? Should human lives be the only thing taking into consideration or should animals count as well? What is the value of life in 1000 years? Organizations and donors remain uncertain over these questions, which plausibly has a strong impact on their behaviour. Having a better understanding of how donors act under this normative uncertainty, and what can be done to reduce this uncertainty thus has substantial importance to the philanthropic sector.

In our study, we aim to provide first data on this question. In it, participants first earned an endowment and were then presented with a choice between five charities that differed along fundamental normative lines. These charities included the National Alliance to End Homelessness, WildAid, GiveDirectly, Cultural Survival, and the Long Now Foundation. These charities were (i) relatively unknown and (ii) supported substantively different causes that draw on a variety of competing normative claims, ranging from domestic homelessness and animal welfare to global poverty, human rights of indigenous peoples, and the long-term future of humanity. All participants were provided with a description of the charity’s mission and some recent successes drawn from the charities’ websites and public material, as well as a normative statement as to why supporting this charity is the right thing to do. These statements drew explicitly on a norm relating to each charity’s subject area, for example claiming that preserving endangered species is important because biodiversity is worth preserving. These statements were added to make salient the normative conflict underlying the choice between the charities. Plausibly, neither the limited information nor the short normative statement provided are sufficient for a choice under certainty. That is, several areas of descriptive uncertainty remain, as the description only covered some aspects of the charity, and the normative statements of the five charities contradicted each other, plausibly leaving participants at least partially uncertain as to what the right thing to do is. We also find this empirically, as most participants self-reported some level of uncertainty about their choice.  

darryll jones