GiveWell: a medium-depth overview

This page supplements our basic About Us page with more information on GiveWell in general. For more information on our specific views on charitable causes and charities, see the "Issues" and "Top Charities" links above. There is also a Frequently Asked Questions page.

What we do

  • GiveWell's mission is to find outstanding giving opportunities and publish the full details of our analysis to help donors decide where to give.
  • GiveWell's vision is a world in which donors reward effectiveness in improving lives.
  • GiveWell is focused on finding a small number of outstanding giving opportunities, not on reviewing as many charities - or as many causes - as possible. We focus the bulk of our research on the causes we find most promising for accomplishing as much good as possible, and within those, on the giving opportunities we find most promising. For more, see our 2011 blog post in which we solidified this commitment.
  • GiveWell does in-depth research and tries to evaluate charities' impact in terms of improving the world. We do not focus on financial metrics such as "percentage of expenses spent on programs vs. administration." More
  • We do not claim to provide "objective" analysis, nor do we believe that "objective" analysis is capable of identifying outstanding giving opportunities. We are open about the fact that we make many judgment calls and make an effort to make these judgment calls transparent so that readers can come to their own conclusions. In short, we care about transparency, not objectivity.
  • We publish as much information as we can - on charities we don't recommend as well as charities we do.

Transparency and credibility

  • GiveWell is committed to extreme transparency in everything we do. We publish a great deal of information about our organization and about the reasoning behind our recommendations.
  • GiveWell has accumulated evidence of credibility. We publish external reviews of our work from people with relevant credentials and experience; we are respected by scholars, nonprofit professionals, and major donors; our research has received significant press and has directed millions of dollars to our top charities. See our Credibility section.

Our process and criteria for finding outstanding giving opportunities

We use a different process and criteria for the two different parts of our research: our traditional work on finding proven, cost-effective, scalable international aid charities and our newer project "GiveWell Labs." We discuss both below.

Proven, cost-effective, scalable international aid charities

GiveWell's general criteria for recommending a charity are

  • Evidence of impact - we look for groups that can produce strong empirical evidence that they're helping people (significantly and reliably), not just appealing ideas and examples.
  • Cost-effectiveness - we seek to maximize "bang for the buck" in terms of good accomplished. We are very mindful of the limitations of cost-effectiveness analysis, which generally cannot come close to capturing the full story of how much a charity is accomplishing. We use cost-effectiveness analysis to look for major differences (e.g., a factor of 10 or more) in how much different programs accomplish for the money.
  • Room for more funding - even if a charity's core programs are outstanding, there is still the question of how additional donations will affect its activities. We feel this issue is the most often overlooked one when discussing where to give.

GiveWell's general process for finding strong charities in an area is to

  • Thoroughly review the independent literature in an area, forming a preliminary idea of promising areas of intervention
  • Cast a wide net for relevant charities, reviewing tax records, foundations' grantees, submissions via our website, referrals and other sources
  • Flag charities that look promising based on heuristics (such as working on a promising program type according to the independent literature, or publishing an unusual amount of substantive information publicly)
  • Have open-ended conversations with promising charities, seeking to evaluate them preliminarily by the above criteria.
  • Conduct very in-depth examinations of the most promising charities, usually including (a) thorough reviews of their evidence of effectiveness and room for more funding; (b) conversations with staff, funders, and other relevant parties; (c) site visits. Create in-depth reviews of these charities discussing their strengths and weaknesses.
  • Rank the strongest opportunities we've found for donors based on the above criteria.

GiveWell Labs

We don't expect to use the same criteria we've used for our traditional work, and the criteria remain a work in progress as of this writing. More

Some core beliefs about good giving and evidence of effectiveness

  • We believe that donors should try to accomplish as much good as possible, which means focusing donations on the most outstanding opportunities rather than spreading them out across many charities. More
  • We approach charities from a skeptical point of view. For some color on why this is, see
  • We are not only skeptical of charities' claims. We also feel it is important not to put too much trust in academic research, and to subject it to the same sort of scrutiny and skepticism.
  • We deeply investigate any claims that are important to us. In 2011 we found serious errors in World Health Organization figures, strengthening our commitment to these investigations.
  • We put heavy weight on empirical evidence of effectiveness, not just conceptual logic of an approach to helping people.
    • Broadly speaking, we look for evidence of effectiveness that is highly specific, representative and credible.
    • Regarding specificity of evidence. Claims like "Our program raises the incomes of farmers by an average of $100 per year" aren't specific enough for us. We try to understand exactly what data a charity has collected to assess its impact and exactly how it has found that data. When we get to this level of specificity - at which there is so little room for interpretation that the charity would be committing fraud if it sent inaccurate figures - we generally take the charity at its word, though we do keep an eye out (on site visits, in conversations with representatives, etc.) for any anomalies, inconsistencies, or other signs that things don't add up.

      In our experience, we rarely see extravagant claims of impact once we get to this level of specificity. It is not practical for us to go to the next level of auditing, which would involve being physically present when data is collected in the field (though we do conduct site visits).

    • Regarding representativeness of evidence. It's common for charities to tell stories of individuals affected by their programs, but most charities serve large numbers of clients, and these sorts of stories could be picked very selectively.

      We are open to all sorts of evidence, including qualitative evidence, if we feel it is representative of the bulk of a charity's activities. In practice, the most common evidence that we think has reasonable claims to "representativeness" usually consists of studies discussing patterns in data.

    • Credibility. Most studies we see are prone to a couple of major problems: selection bias and publication bias. We feel that high-quality studies are relatively rare. We examine studies in depth to determine their credibility, and our writeups include reasons why we find any given study to be more or less credible.
  • In giving, we think it's a mistake to put too much weight on any one criterion, including estimated cost-effectiveness (given the roughness of cost-effectiveness estimates). We collect many different kinds of information on recommended charities and consider them from many different angles.
  • We feel the principle aim of charity should be to empower the people being helped, not to push donors' values upon them. We see some appeal in the idea of giving out cash transfers and often ask why a given charity believes its program is a better use of funds than just giving them out to the poor.
  • We feel that donors should focus on the question of "How can I accomplish the most good?" rather than "How can I address the problem I care most about?" More:
  • We also don't necessarily believe in "helping those who need the most help" - we believe that accomplishing the maximum amount of good is a bit more complex than this.
  • As of September 2013, the most promising area of giving we've found is in developing-world health, though we continue to work to expand our scope via GiveWell Labs. We feel that helping the poor in the developing world is broadly more promising than helping the poor in the developed world (e.g., United States), and that aid has a stronger track record in health than in other areas.

Our strategy for having an impact

Clicky Web Analytics