High-Leverage Research Questions

In my previous post, I discussed three high-level principles I believe will generally promote the most valuable giving research. In this post, I’ll describe three specific research questions I believe are particularly important to answer because of their relevance to improving real world giving decisions:

1. How can we design intermediaries and resources that promote better giving decisions AND that donors will want to use?

2. How do personal differences impact the way people think about giving?

3. Which core beliefs lead to good giving behavior, and which core beliefs obstruct it?  

In the spirit of the principle of paddling downstream, which certainly includes not reinventing the wheel, I’ve included (non-exhaustive) links to research relevant to each topic.

Question 1: How can we design intermediaries and resources that promote better giving decisions AND that donors will want to use?

Donor intermediaries (such as charity evaluators or aggregators, workplace giving platforms, donor advised funds, etc.) will likely play a large role in any efforts to improve giving quality. These are places where (often captive) donors choose between charities, and therefore natural places to try to influence which organizations they give to (and where data could be collected to measure results). Moreover, it’s unreasonable to expect most donors to make informed and impactful giving decisions without having access to the type of  information provided by these intermediaries.

Some intermediaries provide services that donors really like to use, and some intermediaries provide services that help donors significantly improve the quality of their giving. Where those sets intersect lies an enormous opportunity.  This suggests two prongs of research into intermediaries:

The first prong is studying what will get people to use intermediaries in the first place. What qualities would they like in intermediaries? What information do they want to see? How should that information be presented? To answer these questions, we’ll need to both listen to what people say and observe what they do, as these will not always be the same.

The second prong is studying how intermediaries can “nudge” people to give more impactfully, i.e. to make good giving as close to an easy default choice as possible. Intermediaries provide the “choice architecture” in which many donors make their decisions. By changing the “choice architecture” (the context in which people make a decision), we can presumably change their giving behavior.

Existing research on this topic includes:

• Camber Collective’s three Money for Good reports (2010, 2011, and 2015) each include significant research (generally from focus groups) on donor preferences regarding intermediaries.

• The Science of Giving, The Economics of Philanthropy, and A Literature Review of Empirical Studies of Philanthropy all provide helpful summaries of giving research.

• The literature on the Identifiable Victim Effect has significant applications for intermediaries.

• The evaluability bias in charitable giving suggests that designs that employ “joint evaluation” frameworks (people explicitly comparing charities rather than assessing each in isolation) may promote better giving decisions.

• Give more tomorrow suggests benefits to spacing out the timing of the decision to donate and the actual payment of the donation.

• The Effect of Effectiveness examines differences in how large and small donors respond to information about impact.

Question 2: How do personal differences impact the way people think about giving?

Presumably some demographics or market segments will be more receptive to effective giving arguments than others. Identifying and deepening our understanding of these segments should enable more efficient outreach work through targeted communications and messaging.

Learning about these differences can also shed insight on the mechanisms underlying good giving. For instance, GiveWell has reported that its donor base skews very heavily toward the finance and tech industries.  This observation could spur research into why those professions seem to be receptive to GiveWell’s message. Hypotheses could include a general quantitative inclination, explicit training in optimization, high earning power, etc. These lessons could then be used to identify high-potential audiences and strategically design messaging.  

A particularly important demographic to understand is high net worth, and especially ultra-HNW, donors.  Improving giving among a small number of people in this demographic could produce outsized social impact.  

Existing research on this topic includes:

• Money for Good 2010’s behavioral segmentation analysis

• Money for Good 2015 contains additional segmentation analysis and tools, including discussion of “opportunities within specific demographic segments, such as Millennials and women.”

• US Trust study of High Net Worth Philanthropy provides a detailed look at the giving habits of this important demographic.

• Understanding Effective Givers examines inclination toward effective giving across demographic variables (discussed more here).

• GiveWell’s 2016 Metrics Report suggests their audience differs significantly from the general population in terms of demographics and occupation.

Question 3: Which beliefs lead to good giving behavior, and which beliefs obstruct it?  

Startups often use a concept called a Minimum Viable Product (MVP), which is the simplest possible implementation of their long-term vision. Facebook’s MVP was the basic site first used by a select group of Harvard students; it had nowhere near the functionality of today’s product but it still demonstrated the essential nature of a social network.

In the context of donor behavior, we want to identify the “Minimum Viable Beliefs/Behaviors” (which I’ll call MVBs) that will produce good giving outcomes. What are the most parsimonious sets of beliefs, attitudes, and behaviors that lead one to give effectively? These sets function as necessary, or even sufficient, conditions for good giving. If we can convince people to adopt these beliefs and behaviors, it should lead them to give effectively.

I expect there will be different MVB “solutions” for different types of donors. The simplistic examples below illustrate how donors with different motivations can arrive at good giving outcomes via different paths.

• Opportunity-motivated donors: “Impact varies greatly across charities” (belief) + “I want to support highest impact charities” (belief) + “Researching highest impact charities” (behavior) => Good giving

• Obligation-motivated donors: “If I can significantly help someone at low cost to myself, I’m ethically obligated to do so” (belief) + “Some charities are impactful and highly cost-effective” (belief) + “Charity evaluators can identify those charities” (belief) => Good giving

• Effort-minimizing donors: “Philanthropy should be the easy thing in my life” (belief) + “Charity evaluators can make it easier to give” (belief) + “Subscribing to evaluator so charity recommendations are pushed” (behavior) => Good giving

As a corollary to this investigation, it’s important to understand the “Least Valuable Beliefs/Behaviors” (LVBs). These are widely held beliefs that conflict with viable MVB solutions. For instance, an obligation-motivated donor might believe they’re obligated to give if it would make a difference, but they still won’t give well if they believe that all charities are wasteful. So the belief that “all charities are wasteful” would be an LVB.

Note that both MVBs and LVBs exist at different levels of specificity. For example, a highly specific MVB might be the belief that “I want to maximize the social impact of my giving”.  A much more broadly defined MVB might be the belief that charities have differing impact. And a related LVB could be the notion that a gift is good “as long as it goes to a good cause.”

Identifying important beliefs and behaviors isn’t enough though. Improving real world giving behavior will require interventions that can change them. So as we improve our understanding of MVBs and LVBs, the natural next area to investigate is how to efficiently propagate MVBs and counteract LVBs. This will be the topic of the next post in this series.

Existing research on this topic includes:

• The Innovations in Fundraising Wiki “aims to collect and present evidence on the best techniques and practices in the field of fundraising, particularly considering the most effective and impactful charities, and to promote innovative research and its application.”

• The Effect of Effectiveness studies “how donors respond to new information about a charity's effectiveness.”

• The 2017 “Effective Altruism Survey” includes a section on how self-identified effective altruists became involved with EA.