{ "name": "encyclopedia", "description": "Encyclopedia entries.", "license": "CC BY 4.0. Cite “Charitology (charitology.org)”.", "generated": "2026-10-26", "data": [ { "id": "beneficiary-feedback", "path": "/record/encyclopedia/beneficiary-feedback/", "title": "Beneficiary feedback", "summary": "Systematically asking the people a program serves what they think of it, and acting on what they say. It is the most direct way to restore part of the feedback loop that charity is missing.", "aliases": [ "constituent feedback", "constituent voice", "client feedback", "listening to the people served" ], "conditions": [ "missing-feedback-loop", "missing-voice" ], "related": [ "/diagnosis/missing-feedback-loop/", "/open-problems/closing-the-feedback-loop/", "/record/encyclopedia/participatory-grantmaking/", "/record/principles/08-feedback-and-voice/", "/record/lexicon/#feedback-loop" ], "sources": [ { "title": "Listening to Those Who Matter Most, the Beneficiaries", "url": "https://ssir.org/articles/entry/listening_to_those_who_matter_most_the_beneficiaries", "publisher": "Fay Twersky, Phil Buchanan and Valerie Threlfall, Stanford Social Innovation Review", "year": 2013 }, { "title": "Feedback Labs", "url": "https://feedbacklabs.org/", "publisher": "Feedback Labs" }, { "title": "Fund for Shared Insight", "url": "https://fundforsharedinsight.org/", "publisher": "Fund for Shared Insight", "year": 2026, "note": "The collaborative concluded in 2026 after 12 years. Its site is now an archive." }, { "title": "Six Practices of Trust-Based Grantmaking", "url": "https://www.trustbasedphilanthropy.org/practices", "publisher": "Trust-Based Philanthropy Project" } ], "review": "draft", "body": "A business that ignores its customers soon finds out: sales fall. A charity that ignores the people it serves may never find out, because those people are not the ones paying. **Beneficiary feedback** is the practice of asking them anyway, in a systematic way, and changing the work in response.\n\n## Why the signal is missing\n\nIn a 2013 article, Fay Twersky, Phil Buchanan and Valerie Threlfall argued that funders and nonprofits consult experts and crowds but rarely ask the people they serve. They pointed to the reasons: the people served often have few alternatives, funders rather than beneficiaries pay the provider, and incentives favor the funder’s view. So a flawed program can continue for years. This is the [Missing Feedback Loop](/diagnosis/missing-feedback-loop/) seen from the ground.\n\n## What listening can show\n\nThe authors drew on fields that already listen. In health care, they cited research in which patients’ perceptions of their care, especially nursing care, predicted clinical quality and lower inpatient death rates for heart attack patients. In education, the YouthTruth initiative had gathered feedback from about 142,000 students, and 98 percent of participating school leaders said they used or planned to use the results. Their broader point was that the views of the people served can be a leading indicator, showing a problem before it shows up in outcomes.\n\nThey were also clear about the limits. Feedback is expensive to collect. People with low incomes or low literacy are hard to survey well, and people may fear that criticism will cost them help. Feedback works best alongside other evidence, not instead of it.\n\n## Who is building it\n\nFeedback Labs supports funders and nonprofits to listen to people and act on what they hear. Fund for Shared Insight, a national funder collaborative that ran for 12 years before concluding in 2026, launched Listen4Good, a program offering nonprofits tools and coaching to collect and use feedback. The Trust-Based Philanthropy Project counts “solicit and act on feedback” among its six practices, applied to the relationship between funders and grantees.\n\n## The open question\n\nFeedback that is collected but never moves money only half closes the loop. What would let the people served send a signal that changes where money goes, the way prices do in a market? That is [Open Problem 5, Closing the Feedback Loop](/open-problems/closing-the-feedback-loop/). [Participatory grantmaking](/record/encyclopedia/participatory-grantmaking/) and [cash transfers](/record/interventions/cash-transfers/), where recipients choose what to buy, are two partial answers." }, { "id": "charity-as-a-system", "path": "/record/encyclopedia/charity-as-a-system/", "title": "Charity as a system", "summary": "Charitology’s model of giving as one connected system. Givers respond to signals, money moves through channels to the work, the work produces outcomes, and feedback can carry results back. The size of any gap is an open empirical question.", "aliases": [ "the charity system", "the system map", "giving system" ], "conditions": [ "missing-feedback-loop", "missing-profession", "fragmentation" ], "related": [ "/diagnosis/", "/record/principles/02-the-charity-system/", "/record/encyclopedia/source-depth/", "/open-problems/closing-the-feedback-loop/", "/tools/giving-dashboard/" ], "sources": [ { "title": "Giving USA: $617.2B Was Given To Nonprofits During 2025", "url": "https://thenonprofittimes.com/npt_articles/giving-usa-617-2-b-was-given-to-nonprofits-during-2025/", "publisher": "The NonProfit Times, reporting Giving USA 2026", "year": 2026 }, { "title": "Exempt Organizations Business Master File Extract", "url": "https://www.irs.gov/charities-non-profits/exempt-organizations-business-master-file-extract-eo-bmf", "publisher": "Internal Revenue Service", "year": 2026, "note": "The page listed a record count of 1,975,975 organizations when checked in October 2026. The file covers every kind of tax-exempt organization, not only charities." }, { "title": "DAF Assets Soar 30% to $326 Billion", "url": "https://www.philanthropy.com/news/daf-assets-soar-30-to-326-billion/", "publisher": "The Chronicle of Philanthropy, reporting the Annual DAF Report 2025 by the DAF Research Collaborative", "year": 2025 }, { "title": "The Role of Nonprofit Enterprise", "url": "https://yalelawjournal.org/article/the-role-of-nonprofit-enterprise", "publisher": "Henry B. Hansmann, Yale Law Journal 89", "year": 1980 }, { "title": "Annual DAF Report 2025, full report", "url": "https://scholarworks.indianapolis.iu.edu/server/api/core/bitstreams/8f309b9f-da38-433e-8c33-bc1b3d1f9a1a/content", "publisher": "Donor Advised Fund Research Collaborative, Indiana University repository", "year": 2025 } ], "review": "draft", "body": "Charitology studies the **system** that connects them all. Like any system, it produces the results it is built to produce, so to change the results you have to understand the structure.\n\n## Five stages\n\n1. **Givers.** The people and institutions who give. In the United States, Giving USA estimates that $617.2 billion was given in 2025. Individuals gave 64 percent of it, foundations 19 percent, bequests 10 percent and corporations 7 percent.\n2. **Signals.** What givers see when they decide: stories and faces, overhead ratios, rankings, prestige and tax breaks.\n3. **Channels.** Where money is held and routed: charities, foundations, donor-advised funds and giving platforms. The IRS master file of tax-exempt organizations of every kind listed nearly two million records in October 2026, and donor-advised funds alone held about $326 billion in 2024.\n4. **The work.** What the money pays for, described by where it acts on the [depth scale](/record/encyclopedia/source-depth/), from relief to source.\n5. **Outcomes.** What changes in people’s lives, in institutions and in the future.\n\n## The missing loop\n\nIn a market, the person who pays is usually the person who benefits, and prices and reviews carry information back. In charity, the payer and the person served are usually different people. The legal scholar Henry Hansmann called the general version of this *contract failure*: the people paying for a service cannot easily judge its quality. This can make results harder for donors to assess. The frequency and allocation effects of the gap are not measured by Hansmann’s theory. That is [the Missing Feedback Loop](/diagnosis/missing-feedback-loop/), and Charitology treats it as the root of the [Diagnosis](/diagnosis/).\n\n## The field around it\n\nAround all five stages is the field: the knowledge, methods and people that could steer the system as a whole. Researchers, evaluators, advisors and philanthropy educators already study parts of this system. Charitology proposes a shared craft connecting their work across causes and frameworks. That is [the Missing Profession](/diagnosis/missing-profession/).\n\n## Where the conditions sit\n\nEach condition in the Diagnosis acts at one point in the system:\n\n- **Givers:** [Short Horizon](/diagnosis/short-horizon/), [Proximity Pull](/diagnosis/proximity-pull/)\n- **Signals:** [Overhead Myth](/diagnosis/overhead-myth/), [Story Bias](/diagnosis/story-bias/), [Prestige Economy](/diagnosis/prestige-economy/)\n- **Channels:** [Restriction Reflex](/diagnosis/restriction-reflex/), [Fragmentation](/diagnosis/fragmentation/), [Parked Money](/diagnosis/parked-money/)\n- **The work:** [Symptom Trap](/diagnosis/symptom-trap/), [Single Lens](/diagnosis/single-lens/)\n- **Outcomes:** [Measurement Gap](/diagnosis/measurement-gap/), [Silence About Failure](/diagnosis/silence-about-failure/)\n- **Feedback:** [Missing Feedback Loop](/diagnosis/missing-feedback-loop/), [Missing Voice](/diagnosis/missing-voice/)\n\nThe map is a model, not a measurement. Its value is that it shows how a problem at one stage shows up at another. Signals that reward what is visible shape what channels fund, and short, restricted grants favor work with quick, countable outputs. Charitology’s working hypothesis is that fixing one stage while the loop stays open changes little. See [Principles, Chapter 2](/record/principles/02-the-charity-system/) for the full account." }, { "id": "cost-effectiveness-analysis", "path": "/record/encyclopedia/cost-effectiveness-analysis/", "title": "Cost-effectiveness analysis", "summary": "Estimating how much good a program produces per dollar, so that options can be compared. The estimates are rough, and the differences they reveal can be enormous.", "aliases": [ "CEA", "cost per outcome", "cost per DALY", "cost per life saved" ], "conditions": [ "overhead-myth", "measurement-gap" ], "related": [ "/tools/cost-effectiveness-model-kit/", "/tools/impact-calculator/", "/record/encyclopedia/expected-value/", "/open-problems/the-exchange-rate-problem/", "/record/lexicon/#cost-effectiveness" ], "sources": [ { "title": "The Moral Imperative toward Cost-Effectiveness in Global Health", "url": "https://cgdev.org/sites/default/files/1427016_file_moral_imperative_cost_effectiveness.pdf", "publisher": "Toby Ord, Center for Global Development", "year": 2013 }, { "title": "Cost-Effectiveness", "url": "https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness", "publisher": "GiveWell", "year": 2025 }, { "title": "How Much Does It Cost to Save a Life?", "url": "https://www.givewell.org/how-much-does-it-cost-to-save-a-life", "publisher": "GiveWell", "year": 2024 }, { "title": "Errors in DCP2 cost-effectiveness estimate for deworming", "url": "https://blog.givewell.org/2011/09/29/errors-in-dcp2-cost-effectiveness-estimate-for-deworming/", "publisher": "The GiveWell Blog", "year": 2011 }, { "title": "Why we can’t take expected value estimates literally (even when they’re unbiased)", "url": "https://blog.givewell.org/2011/08/18/why-we-cant-take-expected-value-estimates-literally-even-when-theyre-unbiased/", "publisher": "Holden Karnofsky, The GiveWell Blog", "year": 2011 } ], "review": "draft", "body": "**Cost-effectiveness analysis** estimates the cost of producing one unit of a result: a life saved, a case of blindness cured, a year of healthy life added or a year of extra income. It answers the question in the [Giver’s Checklist](/method/#checklist): is this the cheapest way to get this result?\n\n## Why it matters\n\nBecause the differences can be huge. In a 2013 essay for the Center for Global Development, the philosopher Toby Ord looked at 108 health interventions costed in *Disease Control Priorities in Developing Countries*, a standard reference. Their estimated cost-effectiveness ranged from 0.02 to 300 disability-adjusted life years (DALYs) per $1,000, with a median of 5. Moving money from the least to the most cost-effective would produce about 15,000 times the benefit, and moving it from the median to the best about 60 times. Ord noted that real differences may be smaller than the estimates, and argued that even a much smaller spread would still matter morally.\n\nA **DALY** counts the years of life lost to a condition plus the years lived with a disability, weighted by how severe the disability is.\n\n## How it is used\n\nGiveWell calls cost-effectiveness the single most important input in its evaluations, but says it does not judge programs on it alone. Strength of evidence and track record also carry weight. Its April 2024 explanation reported typical modeled costs of $3,000 to $5,500 per life saved for grants to its top charities, while cautioning that future costs could rise. This is a dated model range, not a universal or current guaranteed price.\n\n## How it goes wrong\n\n- **Errors.** In 2011 GiveWell found five errors in the spreadsheet behind the same reference’s estimate for deworming. Together they overstated its cost-effectiveness by a factor of about 100.\n- **Guessed inputs.** An estimate built from many rough guesses can be far off even when no single guess is biased. Holden Karnofsky argued that such estimates should be adjusted toward what similar programs usually achieve, more so the less certain they are. See [expected value](/record/encyclopedia/expected-value/).\n- **Hidden values.** Every model has to say how much one outcome is worth against another, such as income against a death averted. GiveWell calls these *moral weights* and says donors may disagree with its choices.\n\n## The limits\n\nCost-effectiveness works best when options share a unit. Comparing a life saved with a law changed or a species protected needs an exchange rate, and each candidate rate carries a moral framework. That is [Open Problem 2](/open-problems/the-exchange-rate-problem/). Source-level work is hardest of all to cost, because its effects are diffuse and slow. That is [Open Problem 3](/open-problems/measuring-the-source/). To build your own model, start with the [Cost-Effectiveness Model Kit](/tools/cost-effectiveness-model-kit/)." }, { "id": "discounting", "path": "/record/encyclopedia/discounting/", "title": "Discounting", "summary": "Giving less weight to benefits and costs the further in the future they fall. Some reasons for discounting are about uncertainty and growth. One, pure time preference, is the moral claim that later wellbeing matters less just because it is later.", "aliases": [ "discount rate", "time discounting", "pure time preference", "social discount rate" ], "conditions": [ "short-horizon" ], "related": [ "/open-problems/discounting-the-future/", "/open-problems/the-payout-question/", "/record/principles/07-time/", "/tools/legacy-planner/", "/record/lexicon/#pure-time-preference" ], "sources": [ { "title": "Ramsey and Intergenerational Welfare Economics", "url": "https://plato.stanford.edu/entries/ramsey-economics/", "publisher": "Partha Dasgupta, Stanford Encyclopedia of Philosophy", "year": 2019, "note": "First published in 2019 and revised in 2023. Quotes Frank Ramsey’s “A Mathematical Theory of Saving” (1928)." }, { "title": "A review of GiveWell’s discount rate", "url": "https://rethinkpriorities.org/research-area/a-review-of-givewells-discount-rate/", "publisher": "Jenny Kudymowa, James Hu, Melanie Basnak and Tom Hird, Rethink Priorities", "year": 2023 }, { "title": "The “Stern Review” on the Economics of Climate Change", "url": "https://nber.org/papers/w12741", "publisher": "William D. Nordhaus, NBER Working Paper 12741", "year": 2006 } ], "review": "draft", "body": "A **discount rate** shrinks future benefits and costs by a percentage for each year they lie ahead. At 4 percent a year, a benefit 18 years from now counts for about half as much as the same benefit today. The choice of rate can decide whether a long-term program looks brilliant or pointless, so it is worth asking what the rate stands for.\n\n## Three reasons to discount\n\n**Uncertainty.** A benefit promised for the future may never arrive. The program may fail, or the world may change in ways that make it irrelevant. In the extreme, a risk that humanity does not survive acts like a discount rate on all expected future wellbeing.\n\n**Growth.** If people in the future are richer, an extra dollar does less for them than it does for someone poor today. Economists combine this with a time discount rate in what is called the Ramsey rule, after the philosopher and mathematician Frank Ramsey.\n\n**Pure time preference.** Valuing later wellbeing less only because it comes later. Ramsey himself rejected it. In 1928 he wrote that discounting later well-being is “ethically indefensible and arises merely from the weakness of the imagination.”\n\nThe first two reasons are about facts: what will happen and how much it will help. The third is a moral claim about whether time itself makes a life count less.\n\n## An example\n\nA 2023 review by Rethink Priorities described GiveWell’s then-current 4 percent rate for future income and consumption gains, with a lower rate for health benefits. It broke the 4 percent into parts: 0 percent for pure time preference, 1.4 percent for temporal uncertainty, 1.7 percent for improving circumstances as people get richer, and 0.9 percent for benefits that compound over time. Every part except the first is a claim about the world that evidence can improve.\n\n## Why it is contested\n\nThe rate can settle large questions on its own. William Nordhaus argued in 2006 that the Stern Review’s call for urgent, deep cuts in greenhouse gas emissions depended decisively on its use of a near-zero social discount rate, and would not survive rates consistent with market interest rates. Disagreements like this mix facts, such as how fast economies will grow, with values, such as whether time itself matters. Separating the two is the first step to settling them.\n\nIn Charitology’s terms, the [Future generations](/tools/values-compass/) framework treats the claims of people not yet born as central. The [Short Horizon](/diagnosis/short-horizon/) condition is the hypothesis that the charity system’s timescale, set by annual budgets and reports, is shorter than the timescale of its problems. How much the future should count is [Open Problem 13](/open-problems/discounting-the-future/)." }, { "id": "counterfactual-impact", "path": "/record/encyclopedia/counterfactual-impact/", "title": "Counterfactual impact", "summary": "The real effect of a gift is the difference between what happens with it and what would have happened without it. Estimating that difference is one of the hardest and most important problems in giving.", "aliases": [ "counterfactual", "additionality", "replaceability", "funging" ], "conditions": [ "missing-feedback-loop", "measurement-gap" ], "related": [ "/open-problems/the-counterfactual-problem/", "/record/encyclopedia/neglectedness/", "/record/principles/05-counterfactuals-and-neglect/", "/record/lexicon/#counterfactual", "/tools/impact-calculator/" ], "sources": [ { "title": "Counterfactuals and how they change our view of what does good", "url": "https://80000hours.org/articles/counterfactuals/", "publisher": "Benjamin Todd, 80,000 Hours", "year": 2021 }, { "title": "Doing good together: how to coordinate effectively and avoid single-player thinking", "url": "https://80000hours.org/articles/coordination/", "publisher": "Benjamin Todd, 80,000 Hours", "year": 2018 }, { "title": "Revisiting leverage", "url": "https://blog.givewell.org/2018/02/13/revisiting-leverage/", "publisher": "James Snowden, The GiveWell Blog", "year": 2018 }, { "title": "Our Criteria", "url": "https://www.givewell.org/how-we-work/criteria", "publisher": "GiveWell" }, { "title": "“Scared Straight” and other juvenile awareness programs for preventing juvenile delinquency", "url": "https://www.cochrane.org/evidence/CD002796_scared-straight-and-other-juvenile-awareness-programs-preventing-juvenile-delinquency", "publisher": "Anthony Petrosino and colleagues, Cochrane Database of Systematic Reviews", "year": 2022 } ], "review": "draft", "body": "A **counterfactual** is what would have happened otherwise. The counterfactual impact of a gift is the difference between the world with the gift and the world without it. A program’s results and a gift’s results are not the same thing. If the program would have been funded anyway, the gift’s real effect lands somewhere else.\n\n## Three ways the counterfactual changes the picture\n\n**Funding gaps get filled.** Benjamin Todd of 80,000 Hours warns against treating a charity’s funding gap as if you were the only donor. If other donors would have filled it, your gift does less than it seems. The opposite mistake is also possible: assuming your gift does nothing because others would give. Filling the gap frees those donors for other work and saves the charity fundraising time, so the honest answer depends on what everyone else does.\n\n**Other funders respond.** GiveWell uses two words for this. *Leverage* is a charity’s influence on how others, such as governments, allocate their money. *Funging* is when spending on a program crowds out money that would have come from somewhere else. Since late 2017 GiveWell has estimated both for each charity it recommends, and it says the estimates involve substantial guesswork. Its test of “room for more funding” asks the same question from the charity’s side: what would extra money achieve, beyond what the charity would raise anyway?\n\n**The comparison group decides.** A program can look like a success until you see what happened to similar people without it. A Cochrane review of nine trials of “Scared Straight” programs, which organize prison visits for young people at risk of offending, found that they increased offending compared with doing nothing at all. See [the case study](/record/case-studies/scared-straight/).\n\n## Shared credit\n\nMost outcomes have many causes. Todd points out that when two groups are each necessary for a result, both can be counterfactually responsible for all of it, so claimed impacts can add up to more than 100 percent. That is not a logical error. The practical risk is that funders, each counting the whole result, collectively pay more than it is worth.\n\n## Why it is hard\n\nThe counterfactual is never observed. Randomized trials estimate it for the people in a study. Ordinary gifts get no estimate at all, which is part of why the [feedback loop](/diagnosis/missing-feedback-loop/) is missing. Producing counterfactual estimates cheaply and at scale is [Open Problem 4](/open-problems/the-counterfactual-problem/)." }, { "id": "donor-advised-funds", "path": "/record/encyclopedia/donor-advised-funds/", "title": "Donor-advised funds", "summary": "Charitable accounts held by a sponsoring public charity, which the donor advises on grants and investments. They held about $326 billion in 2024, and how fast that money should reach working charities is one of the live debates in philanthropy.", "aliases": [ "DAFs", "DAF", "donor-advised fund" ], "conditions": [ "parked-money", "short-horizon" ], "related": [ "/diagnosis/parked-money/", "/open-problems/the-payout-question/", "/record/encyclopedia/private-foundations/", "/record/lexicon/#donor-advised-fund", "/record/lexicon/#payout-rate" ], "sources": [ { "title": "Donor-advised funds", "url": "https://www.irs.gov/charities-non-profits/charitable-organizations/donor-advised-funds", "publisher": "Internal Revenue Service" }, { "title": "DAF Assets Soar 30% to $326 Billion", "url": "https://www.philanthropy.com/news/daf-assets-soar-30-to-326-billion/", "publisher": "The Chronicle of Philanthropy, reporting the Annual DAF Report 2025 by the DAF Research Collaborative", "year": 2025 }, { "title": "The Annual DAF Report 2025", "url": "https://johnsoncenter.org/resource/report-annual-daf-report-2025/", "publisher": "Johnson Center for Philanthropy", "year": 2025 }, { "title": "Annual DAF Report 2025, full report", "url": "https://scholarworks.indianapolis.iu.edu/server/api/core/bitstreams/8f309b9f-da38-433e-8c33-bc1b3d1f9a1a/content", "publisher": "Donor Advised Fund Research Collaborative, Indiana University repository", "year": 2025 } ], "review": "draft", "body": "A **donor-advised fund** (DAF) is an account for charitable giving. The IRS describes three features. It is a separately identified fund or account made up of a donor’s contributions. It is maintained by a 501(c)(3) organization, called the sponsoring organization, which has legal control of the money once the donor contributes. And the donor keeps advisory privileges over how the money is granted and invested. The donor recommends grants, and the sponsor makes them.\n\n## How big they are\n\nThe Annual DAF Report 2025, produced for the first time by the DAF Research Collaborative after the National Philanthropic Trust compiled it for nearly two decades, found that in 2024:\n\n- DAFs held about **$326 billion** in assets, up from about $256 billion in 2023.\n- Donors contributed **$89.6 billion**, the highest of the past five years.\n- DAFs granted **$64.9 billion** to charities.\n- The aggregate payout rate was **25.3 percent**, calculated as grants divided by prior-year assets.\n\nThe payout rate is not the share of every account spent that year. The report uses prior-year assets as its denominator and aggregates across sponsors.\n\n## Where they sit in the system\n\nOn the [system map](/record/encyclopedia/charity-as-a-system/), a DAF is a channel. It sits between the giver and the work and holds money until the donor decides where it should go. That can be useful. A donor can give when they have money and decide carefully later, and one account can handle many gifts. It also means the money can wait.\n\n## The debate\n\nThis is [Parked Money](/diagnosis/parked-money/), and both sides have numbers. Defenders point to the aggregate payout rate. It is higher than the roughly 5 percent legal distribution requirement for many [private foundations](/record/encyclopedia/private-foundations/), but those measures use different denominators and are not a direct comparison. Critics point to the size of the pool, and note that an aggregate rate can hide accounts that rarely grant at all. The figures above describe DAFs as a whole, so they cannot settle how money moves through individual accounts.\n\n## What we don’t know\n\nIs that aggregate rate fast or slow? The answer depends on how quickly the best giving opportunities are being used up, how much invested money grows while it waits and how much the future should count. That is [Open Problem 12, the Payout Question](/open-problems/the-payout-question/), and [discounting](/record/encyclopedia/discounting/) is part of it." }, { "id": "effective-altruism", "path": "/record/encyclopedia/effective-altruism/", "title": "Effective altruism", "summary": "A philosophy and social movement that uses evidence and careful reasoning to find the ways of helping others that do the most good. Charitology shares its rigor and starts from a different place.", "aliases": [ "EA" ], "conditions": [ "single-lens", "proximity-pull" ], "related": [ "/record/encyclopedia/scale-neglectedness-tractability/", "/record/encyclopedia/moral-uncertainty/", "/record/disagreements/near-or-far/", "/record/principles/01-what-charitology-is/", "/record/encyclopedia/systemic-change/" ], "sources": [ { "title": "Famine, Affluence, and Morality", "url": "https://www.jstor.org/stable/2265052", "publisher": "Peter Singer, Philosophy & Public Affairs 1(3)", "year": 1972 }, { "title": "About GiveWell", "url": "https://www.givewell.org/about", "publisher": "GiveWell" }, { "title": "Our history", "url": "https://www.givingwhatwecan.org/about-us/history", "publisher": "Giving What We Can" }, { "title": "The history of the term ‘effective altruism’", "url": "https://forum.effectivealtruism.org/posts/9a7xMXoSiQs3EYPA2/the-history-of-the-term-effective-altruism", "publisher": "William MacAskill, Effective Altruism Forum", "year": 2014 }, { "title": "Open Philanthropy Is Now Coefficient Giving", "url": "https://forum.effectivealtruism.org/posts/vkvtu6xbvfkHPhJkC/open-philanthropy-is-now-coefficient-giving", "publisher": "Effective Altruism Forum", "year": 2025 }, { "title": "What are the most pressing world problems?", "url": "https://80000hours.org/problem-profiles/", "publisher": "80,000 Hours" }, { "title": "FTX bankruptcy also endangers founder’s philanthropic gifts", "url": "https://www.news4jax.com/business/2022/11/13/ftx-bankruptcy-also-endangers-founders-philanthropic-gifts/", "publisher": "Thalia Beaty and Glenn Gamboa, Associated Press", "year": 2022 }, { "title": "Effective Altruism and Its Critics", "url": "https://doi.org/10.1111/japp.12176", "publisher": "Iason Gabriel, Journal of Applied Philosophy 34(4)", "year": 2017 }, { "title": "Stop the Robot Apocalypse: The New Utilitarians", "url": "https://www.lrb.co.uk/the-paper/v37/n18/amia-srinivasan/stop-the-robot-apocalypse", "publisher": "Amia Srinivasan, London Review of Books 37(18)", "year": 2015 } ], "review": "draft", "body": "**Effective altruism** asks a simple question and tries to answer it rigorously: how can I do the most good with the time and money I have? It answers with evidence, cost-effectiveness and comparisons across causes, and it asks people to act on the answers.\n\n## A short history\n\nThe movement drew on Peter Singer’s 1972 essay “Famine, Affluence, and Morality,” which argued that distance makes no moral difference to the duty to help. GiveWell began as research by a group of friends working in finance, and in 2007 two of them left their jobs to run it full time. Giving What We Can was launched in November 2009 by the Oxford philosopher Toby Ord, working with Will MacAskill, around a pledge to give 10 percent of income. 80,000 Hours, which advises people on high-impact careers, soft-launched in 2011.\n\nThe name came late. When Giving What We Can and 80,000 Hours needed an umbrella organization, their directors held a vote, and on 5 December 2011 “Centre for Effective Altruism” won. GiveWell’s research arm, GiveWell Labs, grew into the Open Philanthropy Project, later known as Open Philanthropy, which renamed itself Coefficient Giving in 2025.\n\n## Ideas and priorities\n\nThe movement’s tools include [scale, neglectedness and tractability](/record/encyclopedia/scale-neglectedness-tractability/), [cost-effectiveness analysis](/record/encyclopedia/cost-effectiveness-analysis/), [counterfactual](/record/encyclopedia/counterfactual-impact/) reasoning and [expected value](/record/encyclopedia/expected-value/). Its priorities have widened from global health and poverty. 80,000 Hours’ problem list includes risks from advanced artificial intelligence and engineered pandemics among its priorities, alongside factory farming and global health.\n\nIn November 2022 the cryptocurrency exchange FTX went bankrupt. Its founder had been a prominent backer of the movement, and the leaders of the FTX Future Fund resigned, warning that many committed grants would not be paid. The collapse prompted a period of public self-examination among the movement’s leaders.\n\n## Critiques\n\nPhilosophers and activists have pressed several objections. Iason Gabriel grouped them under three headings: that the movement neglects justice, including equality, urgency and rights; that its methods are biased toward what can be measured; and that it is poorly placed to bring about systemic change. Amia Srinivasan argued that it works within the existing order rather than asking what produces global misery. Gabriel’s paper also sets out the responses to each objection, and the debate continues. See [systemic change](/record/encyclopedia/systemic-change/).\n\n## Charitology and effective altruism\n\nCharitology shares the movement’s rigor: its insistence on evidence, counterfactuals, honest uncertainty and comparisons across causes. Charitology makes explicit editorial commitments to eight frameworks (welfare, rights, justice, care, faith and tradition, future generations, all life and flourishing), and reports where they agree and part. Effective altruism also contains debates about moral uncertainty and values; it is not one uniform moral theory. And its subject is the charity system itself: the signals, channels, incentives and feedback that decide where all generosity goes, read through the [symptom and source](/record/encyclopedia/symptoms-and-sources/) lens. Charitology does not judge the movement or the people in it. It treats effective altruism as a tradition to learn from, and as one lens among several. The [Single Lens](/diagnosis/single-lens/) condition is a warning against using any one lens alone." }, { "id": "identifiable-victim-effect", "path": "/record/encyclopedia/identifiable-victim-effect/", "title": "Identifiable victim effect", "summary": "The tendency to give more to help one specific, identified person than to help many people described by statistics. It is one of the best-known findings about giving, and recent replications have questioned how large and reliable it is.", "aliases": [ "identified victim effect", "singularity effect", "statistical victims" ], "conditions": [ "story-bias", "proximity-pull" ], "related": [ "/diagnosis/story-bias/", "/open-problems/story-without-bias/", "/record/encyclopedia/psychic-numbing/", "/record/lexicon/#identifiable-victim-effect", "/open-problems/the-replication-rate/" ], "sources": [ { "title": "Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims", "url": "https://doi.org/10.1016/j.obhdp.2006.01.005", "publisher": "Deborah A. Small, George Loewenstein and Paul Slovic, Organizational Behavior and Human Decision Processes 102(2)", "year": 2007, "note": "Freely readable at https://faculty.wharton.upenn.edu/wp-content/uploads/2012/04/Sympathy_and_callousness.pdf" }, { "title": "The identifiable victim effect: a meta-analytic review", "url": "https://doi.org/10.1080/15534510.2016.1216891", "publisher": "Seyoung Lee and Thomas Hugh Feeley, Social Influence 11(3)", "year": 2016 }, { "title": "Revisiting and Rethinking the Identifiable Victim Effect: Replication and Extension of Small, Loewenstein, and Slovic (2007)", "url": "https://doi.org/10.1525/collabra.90203", "publisher": "Maximilian Maier, Yik Chun Wong and Gilad Feldman, Collabra: Psychology 9(1)", "year": 2024, "note": "Open manuscript and publication record at https://wrap.warwick.ac.uk/id/eprint/201115/" }, { "title": "FLoRA Replication Atlas entry for Small, Loewenstein and Slovic (2007)", "url": "https://forrt.org/flora-replication-atlas/doi/10.1016/j.obhdp.2006.01.005/", "publisher": "FORRT" } ], "review": "draft", "body": "The **identifiable victim effect** is the finding that people give more to help one specific person, with a name and a face, than to help a larger number of people described by statistics. It is a central part of the [Story Bias](/diagnosis/story-bias/).\n\n## The classic study\n\nIn experiments published in 2007, Deborah Small, George Loewenstein and Paul Slovic paid participants $5 for an unrelated survey and then gave them a chance to donate some of it to Save the Children. Some saw an appeal about one child with a photograph. Others read statistics about food shortages affecting millions of people in southern and eastern Africa.\n\nIn the first study, people who saw the individual appeal and received no other information gave an average of $2.83. People who saw only the statistics gave $1.17. In a later study, adding the statistics to the individual story *reduced* giving to $1.43, compared with $2.38 for her story alone. And when people were taught about the identifiable victim effect before deciding, they gave less to the individual appeal but no more to the statistical victims. The authors concluded that deliberate thinking dampened sympathy for the identified victim without creating sympathy for the many.\n\n## How big is it?\n\nA 2016 meta-analysis by Seyoung Lee and Thomas Feeley pooled 41 studies and found an effect that was statistically significant but modest. It looked most reliable in particular conditions: a single victim, a photographed child in poverty, a victim who bore little responsibility for their need and a request for money.\n\nIn 2024 Maximilian Maier, Yik Chun Wong and Gilad Feldman published a pre-registered replication of the 2007 studies, using hypothetical donations. They found no support for the effect. They also reanalyzed the earlier meta-analysis, found signs of publication bias and suggested that the pattern may be better described as insensitivity to scale. The [replication record](https://forrt.org/flora-replication-atlas/doi/10.1016/j.obhdp.2006.01.005/) now lists the 2007 paper’s replication as failed.\n\nThe finding is now debated. One difference between the studies is that the 2007 donations were real money and the 2024 donations were hypothetical, and the replication’s authors themselves call for large studies with real donations. Those would go a long way to settling it.\n\n## Why it matters for the system\n\nEven at its most modest, the pattern points the same way as a wider body of research on [psychic numbing](/record/encyclopedia/psychic-numbing/): concern does not grow in proportion to the number of people in need. Charitology hypothesizes that a system relying on faces and stories may favor needs that are easier to portray. The studies here do not measure that system-wide allocation effect. Whether a story can carry scale instead of hiding it is [Open Problem 7, Story Without Bias](/open-problems/story-without-bias/). How often findings like this one hold up is [Open Problem 10](/open-problems/the-replication-rate/)." }, { "id": "impact-investing", "path": "/record/encyclopedia/impact-investing/", "title": "Impact investing", "summary": "Investing with the intention of producing measurable social or environmental good alongside a financial return. It brings investment capital, not only gifts, to bear on social problems, and it raises a sharp counterfactual question.", "aliases": [ "social investing", "mission-related investing", "program-related investments" ], "conditions": [ "measurement-gap", "parked-money" ], "related": [ "/record/encyclopedia/counterfactual-impact/", "/record/encyclopedia/private-foundations/", "/record/encyclopedia/outputs-and-outcomes/", "/tools/impact-measurement-toolkit/" ], "sources": [ { "title": "What you need to know about impact investing", "url": "https://thegiin.org/publication/post/about-impact-investing/", "publisher": "Global Impact Investing Network (GIIN)" }, { "title": "Sizing the Impact Investing Market 2024", "url": "https://thegiin.org/publication/research/sizing-the-impact-investing-market-2024/", "publisher": "Dean Hand, Maddie Ulanow, Hongyu Pan and Kelly Xiao, Global Impact Investing Network", "year": 2024 }, { "title": "The State of Impact Investing", "url": "https://magazine.wharton.upenn.edu/digital/the-state-of-impact-investing/", "publisher": "Jem Hudson, Wharton Magazine", "year": 2014 }, { "title": "Many Paths to One Mountaintop – Antony Bugg-Levine on Building Movements That Shift Systems", "url": "https://www.rockefellerfoundation.org/bellagio-breakthroughs/many-paths-to-one-mountaintop-antony-bugg-levine-on-building-movements-that-shift-systems/", "publisher": "The Rockefeller Foundation" }, { "title": "Program-related investments", "url": "https://www.irs.gov/charities-non-profits/private-foundations/program-related-investments", "publisher": "Internal Revenue Service" } ], "review": "draft", "body": "**Impact investing** puts capital into companies, funds or projects with the aim of doing good as well as earning a return. The Global Impact Investing Network (GIIN) defines impact investments as those made to generate positive, measurable social or environmental impact alongside a financial return. It names four core characteristics: an intention to contribute to measurable benefits, the use of evidence and impact data in designing investments, managing toward the impact intended, and sharing lessons so the field can grow.\n\n## Origins and size\n\nThe term was coined at a Rockefeller Foundation gathering at its Bellagio Center in Italy, which Wharton Magazine dates to 2007. The convening led to the Global Impact Investing Network, whose formation the foundation announced in 2009. In its 2024 market sizing, the GIIN estimated that more than 3,900 organizations managed about $1.57 trillion in impact investments worldwide.\n\n## Returns\n\nThe GIIN says impact investments can target anything from below-market to above-market financial returns. Some investors accept less money back in exchange for more good. Most of the investors the GIIN has surveyed target risk-adjusted, market-rate returns.\n\nFoundations have a tool of their own. A **program-related investment**, in IRS terms, is one made primarily for charitable purposes, where income is not a significant purpose. The IRS’s examples include low-interest loans to needy students and high-risk investments in nonprofit low-income housing.\n\n## The counterfactual question\n\nThe hardest question in impact investing is the one at the heart of [counterfactual impact](/record/encyclopedia/counterfactual-impact/): would the good have happened anyway? If an investment earns full market returns, an ordinary investor might have made it too, and the impact investor’s money changed little. If it accepts lower returns, or takes risks others won’t, it is more likely to have made a difference. Measured outcomes matter too: an investment that reports outputs, such as loans made, has not yet shown [outcomes](/record/encyclopedia/outputs-and-outcomes/), such as incomes raised.\n\n## In the charity system\n\nImpact investing sits beside the charity system rather than inside it, but it competes for the same intentions. A foundation choosing between a grant and an investment, or a donor between a gift and an impact fund, is making a depth and counterfactual decision. Charitology describes both kinds of capital with the same questions from the [Giver’s Checklist](/method/#checklist)." }, { "id": "expected-value", "path": "/record/encyclopedia/expected-value/", "title": "Expected value", "summary": "The probability-weighted average of an action’s possible results. It is the standard way to compare a sure thing with a long shot, and it misleads when the probabilities are guesses.", "aliases": [ "EV", "expected impact", "probability-weighted value" ], "conditions": [ "measurement-gap", "short-horizon" ], "related": [ "/record/encyclopedia/moral-uncertainty/", "/record/encyclopedia/cost-effectiveness-analysis/", "/tools/forecasting-platform/", "/programs/forecasting-tournaments/", "/record/lexicon/#expected-value" ], "sources": [ { "title": "Why we can’t take expected value estimates literally (even when they’re unbiased)", "url": "https://blog.givewell.org/2011/08/18/why-we-cant-take-expected-value-estimates-literally-even-when-theyre-unbiased/", "publisher": "Holden Karnofsky, The GiveWell Blog", "year": 2011 }, { "title": "Moral Uncertainty", "url": "https://doi.org/10.1093/oso/9780198722274.001.0001", "publisher": "William MacAskill, Krister Bykvist and Toby Ord, Oxford University Press", "year": 2020 }, { "title": "Stop the Robot Apocalypse: The New Utilitarians", "url": "https://www.lrb.co.uk/the-paper/v37/n18/amia-srinivasan/stop-the-robot-apocalypse", "publisher": "Amia Srinivasan, London Review of Books 37(18)", "year": 2015 }, { "title": "A framework for comparing global problems in terms of expected impact", "url": "https://80000hours.org/articles/problem-framework/", "publisher": "Robert Wiblin, 80,000 Hours", "year": 2016 } ], "review": "draft", "body": "The **expected value** of an action is what you get if you multiply each possible result by its probability and add them up. A program that has a 10 percent chance of helping 10,000 people has an expected value of helping 1,000 people, the same as a program certain to help 1,000. Expected value lets a giver compare a sure thing with a long shot on one scale.\n\n## Why it matters for giving\n\nMuch of the most ambitious work in charity is a long shot: research that may find nothing, advocacy that may lose, a campaign to change a norm that may take decades or fail. Without some way to weigh probability against size, a giver can only choose between “certain” and “hopeful.” Expected value turns that into a question with numbers in it.\n\n## Three cautions\n\n**The inputs are often guesses.** Holden Karnofsky of GiveWell argued in 2011 that an explicit expected-value estimate, built from many rough inputs, should not be taken literally even when its errors are not biased in either direction. It should be combined with a prior view of what similar actions usually achieve. The less certain the estimate, the more it should be pulled toward that prior, so a dramatic but weakly supported claim counts for little.\n\n**Tiny chances of huge stakes can take over.** If a result is large enough, even a very small probability can dominate the sum. Philosophers call the extreme version of this *fanaticism*, and MacAskill, Bykvist and Ord treat it as a central problem for anyone deciding under [moral uncertainty](/record/encyclopedia/moral-uncertainty/). The philosopher Amia Srinivasan has argued that expected-value reasoning about speculative risks can end up justifying whatever its users already favored.\n\n**It ignores attitudes to risk.** Maximizing expected value treats a sure good and a gamble with the same average as equal. 80,000 Hours calls ignoring risk aversion a reasonable approximation for its own purposes. Many givers reasonably prefer some certainty. A portfolio can hold both.\n\n## Getting the probabilities right\n\nExpected value is only as good as its probabilities. Forecasters can be scored, for example with a [Brier score](/record/lexicon/#brier-score), and a forecaster whose 70 percent predictions come true about 70 percent of the time is [calibrated](/record/lexicon/#calibration). Charitology’s [Forecasting Platform](/tools/forecasting-platform/) and [Forecasting Tournaments](/programs/forecasting-tournaments/) are meant to make the probabilities in giving decisions testable." }, { "id": "moral-uncertainty", "path": "/record/encyclopedia/moral-uncertainty/", "title": "Moral uncertainty", "summary": "Uncertainty about which moral view is right, as opposed to uncertainty about the facts. Philosophers have developed ways to act well under it, and Charitology’s method of looking through every framework depends on them.", "aliases": [ "normative uncertainty", "deciding under moral uncertainty" ], "conditions": [ "single-lens" ], "related": [ "/open-problems/moral-parliament-in-practice/", "/tools/portfolio-builder/", "/tools/values-compass/", "/record/principles/06-comparing-the-incomparable/", "/record/lexicon/#moral-parliament" ], "sources": [ { "title": "Moral Uncertainty", "url": "https://doi.org/10.1093/oso/9780198722274.001.0001", "publisher": "William MacAskill, Krister Bykvist and Toby Ord, Oxford University Press", "year": 2020, "note": "Open access under a Creative Commons license (CC BY-NC-ND 4.0); freely readable at https://www.moraluncertainty.com/s/Moral-Uncertainty.pdf" }, { "title": "Moral uncertainty – towards a solution?", "url": "https://www.overcomingbias.com/p/moral-uncertainty-towards-a-solutionhtml", "publisher": "Nick Bostrom, Overcoming Bias", "year": 2009 }, { "title": "The parliamentary approach to moral uncertainty", "url": "https://ora.ox.ac.uk/objects/uuid:b6b3bc2e-ba48-41d2-af7e-83f07c1fe141", "publisher": "Toby Newberry and Toby Ord, Future of Humanity Institute Technical Report 2021-2", "year": 2021 } ], "review": "draft", "body": "You can be sure of every fact about a gift and still not know what to do. A donation might save a child’s life or protect a forest for a century, and you might know both outcomes exactly while being unsure how to weigh them. That second kind of doubt is **moral uncertainty**: uncertainty about moral and evaluative questions rather than descriptive ones. In their 2020 book *Moral Uncertainty*, William MacAskill, Krister Bykvist and Toby Ord argue that, given how hard ethics is and how widely people disagree about it, moral uncertainty is the normal condition rather than the exception.\n\n## Ways to decide\n\n**Follow your favorite theory.** Act on whichever moral view you find most plausible. MacAskill, Bykvist and Ord call this *My Favorite Theory* and argue that, simple as it is, it has major problems, starting with what to do when you are split between views.\n\n**Maximize expected choiceworthiness.** Weight each theory’s verdict by how much credence you give it, the way you would weight outcomes by their probability. This works only when theories can be compared on a common scale.\n\n**Borrow from voting.** Some theories only rank options, and some give amounts that cannot be compared with other theories’ amounts. For those cases the book draws an analogy with social choice, the study of how groups vote, and adapts voting rules such as the Borda rule.\n\n**Convene a moral parliament.** In 2009 Nick Bostrom proposed imagining that each theory you take seriously sends delegates to a parliament, in proportion to your credence in it. The delegates bargain and vote, and you act on the result. A theory with few seats can still win on the issues it cares about most by giving way on others. Bostrom developed the idea with Toby Ord, and in 2021 Toby Newberry and Ord argued for a version based on proportional chances voting. They also showed that other approaches to moral uncertainty can be read as parliaments with different rules.\n\n## Hard cases\n\nThe book names two problems any approach must face. *Fanaticism* is the worry that a theory you barely believe could dominate your choices if it says the stakes are enormous. *Infectious incomparability* is the worry that giving any credence to a view on which options cannot be compared spreads that incomparability to every decision you make.\n\n## In Charitology\n\nCharitology’s answer to the [Single Lens](/diagnosis/single-lens/) is to look at every gift through eight moral frameworks at once: welfare, rights and duties, justice, care, faith and tradition, future generations, all life and flourishing. The [Values Compass](/tools/values-compass/) asks how strongly you agree with a statement from each one, and the [Portfolio Builder](/tools/portfolio-builder/) uses a simple credence-weighted method to turn weights like these into an allocation. Whether more sophisticated methods give different, and more defensible, answers is [Open Problem 16](/open-problems/moral-parliament-in-practice/)." }, { "id": "mutual-aid", "path": "/record/encyclopedia/mutual-aid/", "title": "Mutual aid", "summary": "People in a community pooling time, money and skills to meet one another’s needs directly, without a charity standing between giver and receiver. Its structure can differ from giving through a separate service provider.", "aliases": [ "solidarity networks", "community self-help" ], "conditions": [ "missing-feedback-loop", "missing-voice", "proximity-pull" ], "related": [ "/record/encyclopedia/charity-as-a-system/", "/record/encyclopedia/beneficiary-feedback/", "/record/traditions/", "/programs/local-charitology/", "/record/lexicon/#mutual-aid" ], "sources": [ { "title": "Mutual Aid: A Factor of Evolution", "url": "https://www.gutenberg.org/ebooks/4341", "publisher": "Peter Kropotkin (Project Gutenberg edition)", "year": 1902 }, { "title": "We Keep Each Other Safe: Mutual Aid for Survival and Solidarity", "url": "https://bcrw.barnard.edu/videos/we-keep-each-other-safe-mutual-aid-for-survival-and-solidarity", "publisher": "Barnard Center for Research on Women, conversation with Dean Spade, Mariame Kaba and Ejeris Dixon", "year": 2021 } ], "review": "draft", "body": "In **mutual aid**, the people who help and the people who are helped belong to the same community, and often trade places. Neighbors run a grocery delivery list, a congregation keeps a fund for funerals, a tenants’ group covers a member’s rent this month and someone else’s next month. Giving and receiving can overlap, rather than being fixed roles.\n\n## Roots\n\nThe idea is old, and so is its name. In *Mutual Aid: A Factor of Evolution*, published in 1902, Peter Kropotkin argued against the view that competition is the main force in nature and society. Drawing on observations of animals and of human societies, from indigenous communities to medieval guilds, he held that cooperation is just as basic to survival.\n\nOrganizers use the term today too. Dean Spade, author of *Mutual Aid: Building Solidarity During This Crisis (and the Next)*, describes it as collective coordination to meet one another’s needs, usually because existing systems are not meeting them, and points out that those systems have often helped create the crisis in the first place.\n\n## How it differs on the system map\n\nOn the [system map](/record/encyclopedia/charity-as-a-system/), most charity runs one way: from givers through signals and channels to work done for someone else, with results rarely coming back. Mutual aid folds the stages together. The giver can see the need directly, the person helped often has a say in what help looks like, and the result is visible to everyone involved. In Charitology’s terms, direct contact may support a [feedback loop](/diagnosis/missing-feedback-loop/) and [voice](/diagnosis/missing-voice/). That is a feature to examine, not a guarantee that every network shares power well.\n\nThe costs mirror the strengths. Mutual aid reaches the people inside a network, so people outside it can be missed. It runs on volunteer time, which can run out. And because it is close to home, it carries the [Proximity Pull](/diagnosis/proximity-pull/) by design. Under the care framework that is the point; under the welfare framework it is a trade-off worth seeing clearly.\n\n## Depth\n\nMutual aid can provide relief: food, rent and rides this week. But the networks it builds can outlast the crisis that formed them. Humanity’s Checklist names coordination, whether the people who could solve a problem together ever meet, as one of the six sources upstream of world problems. Whether mutual aid networks change coordination in a lasting way is a question the sources here do not settle.\n\n## What we don’t know\n\nMuch mutual aid never passes through a registered organization, so it is largely invisible to the data the rest of the system uses. This entry does not establish its total scale or compare its results with formal charity working on the same needs. The [Giving Traditions Library](/record/traditions/) records older forms of the same practice." }, { "id": "neglectedness", "path": "/record/encyclopedia/neglectedness/", "title": "Neglectedness", "summary": "How few resources already go to a problem relative to its size. Other things equal, an extra dollar tends to do more where few other dollars go, because the best opportunities there have not yet been taken.", "aliases": [ "crowdedness", "uncrowdedness", "diminishing returns" ], "conditions": [ "story-bias", "proximity-pull", "symptom-trap", "single-lens" ], "related": [ "/record/encyclopedia/scale-neglectedness-tractability/", "/record/encyclopedia/counterfactual-impact/", "/record/lexicon/#neglectedness", "/open-problems/the-depth-census/", "/tools/compare-tool/" ], "sources": [ { "title": "A framework for comparing global problems in terms of expected impact", "url": "https://80000hours.org/articles/problem-framework/", "publisher": "Robert Wiblin, 80,000 Hours", "year": 2016, "note": "First published in 2016 and last updated in December 2025." }, { "title": "The Important/Neglected/Tractable framework needs to be applied with care", "url": "https://forum.effectivealtruism.org/posts/74oJS32C6CZRC4Zp5/the-important-neglected-tractable-framework-needs-to-be", "publisher": "Robert Wiblin, Effective Altruism Forum", "year": 2016 }, { "title": "Narrowing down U.S. policy areas", "url": "https://blog.givewell.org/2014/05/22/narrowing-down-u-s-policy-areas/", "publisher": "Holden Karnofsky, The GiveWell Blog", "year": 2014 } ], "review": "draft", "body": "A problem is **neglected** when few people or dollars are working on it compared with how big it is. 80,000 Hours measures neglectedness simply: how many people, or how many dollars, are currently dedicated to solving the problem?\n\n## Why it matters\n\nThe first resources that go into a problem usually go to its best opportunities. Later resources go to the next best, and so on. So returns diminish as a field fills up. 80,000 Hours argues that returns most likely diminish logarithmically, which means each doubling of resources buys roughly the same amount of progress. If that is right, a dollar added to a problem that gets a million dollars a year can matter far more than a dollar added to one that gets a billion.\n\nGiveWell Labs, the project that became the Open Philanthropy Project, made a similar point in 2014 with a different word. When it compared U.S. policy areas, it asked how *crowded* each one was: how much existing advocacy already pushed for the same goals, and where the gaps were.\n\n## Three cautions\n\n1. **Neglected for a reason.** Neglectedness is a good guide only when a problem is neglected for bad reasons. Some problems get little money because they are small or because nobody knows how to make progress on them.\n2. **The factors overlap.** Robert Wiblin of 80,000 Hours notes that a problem can stay important precisely because it is neglected. He suggests judging how important and how tractable a problem would be if nobody were working on it, so that one observation is not counted twice.\n3. **Boundaries move the answer.** A narrowly defined problem almost always looks more neglected than a broad one. How you draw the edges of a problem changes its score.\n\n## Neglect as a property of the system\n\nCharitology adds a system view. Neglect is not random. The [Story Bias](/diagnosis/story-bias/) predicts that problems without faces or headlines will be neglected. The [Proximity Pull](/diagnosis/proximity-pull/) predicts neglect of distant need, and the [Single Lens](/diagnosis/single-lens/) predicts neglect of goods that the dominant moral frameworks cannot see.\n\nCharitology’s founding hypothesis, the [Symptom Trap](/diagnosis/symptom-trap/), is that work at the structure and source levels is neglected relative to relief. That hypothesis is untested. Testing it starts with [the Depth Census](/open-problems/the-depth-census/)." }, { "id": "outputs-and-outcomes", "path": "/record/encyclopedia/outputs-and-outcomes/", "title": "Outputs and outcomes", "summary": "Outputs are what a program does or delivers. Outcomes are what changes for people as a result. Counting outputs alone does not establish outcomes.", "aliases": [ "outputs", "outcomes", "impact", "results chain", "logic model" ], "conditions": [ "measurement-gap", "overhead-myth" ], "related": [ "/diagnosis/measurement-gap/", "/record/encyclopedia/theory-of-change/", "/tools/impact-measurement-toolkit/", "/method/", "/record/lexicon/#outcomes" ], "sources": [ { "title": "Logic Model Development Guide", "url": "https://www.naccho.org/uploads/downloadable-resources/Programs/Public-Health-Infrastructure/KelloggLogicModelGuide_161122_162808.pdf", "publisher": "W.K. Kellogg Foundation", "year": 2004 }, { "title": "Insecticide-treated nets for preventing malaria (Cochrane systematic review)", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6418392/", "publisher": "Joseph Pryce, Marty Richardson and Christian Lengeler, Cochrane Database of Systematic Reviews", "year": 2018 } ], "review": "draft", "body": "Every program can be described as a chain: resources pay for **activities**, activities produce **outputs**, and outputs may or may not lead to **outcomes**. The W.K. Kellogg Foundation’s *Logic Model Development Guide* defines the links this way:\n\n- **Outputs** are the direct products of a program’s activities, usually counted by size and scope: classes taught, meetings held, materials distributed, people served.\n- **Outcomes** are specific changes in participants’ behavior, knowledge, skills, status and level of functioning. The guide expects short-term outcomes within one to three years and longer-term outcomes within four to six.\n- **Impact** is the fundamental change, intended or not, in organizations, communities or systems, which the guide places seven to ten years out, often after the grant has ended.\n\n## An example\n\nImagine a bed-net program distributing 10,000 nets. That is an output. Families sleeping under the nets is an early outcome. Fewer children dying of malaria is the outcome that matters. In the trials pooled by a 2018 Cochrane review, insecticide-treated nets cut child deaths from all causes by about 17 percent compared with no nets. A program that only counts nets handed out cannot say whether it achieved anything like that.\n\n## Why outputs win\n\nOutputs are cheap to count and can be counted in-house. Monitoring outcomes may need follow-up and time. Attributing changes to a program also needs a credible comparison, which does not always require randomization. The risk is that a count such as “10,000 meals served” is read as evidence of a later change without that change being checked. That is the [Measurement Gap](/diagnosis/measurement-gap/). The [Overhead Myth](/diagnosis/overhead-myth/) makes it worse, because measuring outcomes takes staff and systems that donors often see as overhead.\n\n## Outcomes can be negative\n\nThe Kellogg definition of impact includes unintended change. Some programs produce outcomes opposite to the ones intended, which only a comparison shows. See [Scared Straight](/record/case-studies/scared-straight/).\n\n## Outcomes and depth\n\nSome work at the structure and source levels of the [depth scale](/record/encyclopedia/source-depth/) has outcomes far from its activities or takes a long time to show them. Depth does not determine timing by itself. A campaign to change a law has outputs within months and outcomes over years. That is part of why source-level work is so hard to judge. See [theory of change](/record/encyclopedia/theory-of-change/) and the [Impact Measurement Toolkit](/tools/impact-measurement-toolkit/)." }, { "id": "overhead-ratio", "path": "/record/encyclopedia/overhead-ratio/", "title": "Overhead ratio", "summary": "The share of a charity’s spending that goes to management and fundraising rather than to programs. It is easy to find and widely used, but it does not establish what a charity achieves.", "aliases": [ "overhead", "program expense ratio", "administrative costs", "fundraising ratio" ], "conditions": [ "overhead-myth", "measurement-gap" ], "related": [ "/diagnosis/overhead-myth/", "/record/myths/", "/open-problems/the-overhead-replacement/", "/tools/charity-lookup/", "/record/lexicon/#overhead-ratio" ], "sources": [ { "title": "Instructions for Form 990, Part IX, Statement of Functional Expenses", "url": "https://www.irs.gov/instructions/i990", "publisher": "Internal Revenue Service", "year": 2025 }, { "title": "The Nonprofit Starvation Cycle", "url": "https://ssir.org/articles/entry/the_nonprofit_starvation_cycle", "publisher": "Ann Goggins Gregory and Don Howard, Stanford Social Innovation Review", "year": 2009 }, { "title": "3 Major Charity Groups Ask Donors to Stop Focusing on Overhead Costs", "url": "https://www.philanthropy.com/news/3-major-charity-groups-ask-donors-to-stop-focusing-on-overhead-costs/", "publisher": "The Chronicle of Philanthropy", "year": 2013 }, { "title": "Charity Impact Media", "url": "https://theguysaccount.github.io/charity-impact-media/", "publisher": "Charity Impact Media" }, { "title": "The overhead myth: crash course to fundraising transparency", "url": "https://candid.org/blogs/the-overhead-myth-crash-course-to-fundraising-transparency/", "publisher": "Candid", "year": 2022 } ], "review": "draft", "body": "U.S. charities that file the full IRS Form 990 report their expenses in Part IX, the Statement of Functional Expenses, which sorts them into three columns: program services, management and general, and fundraising. The **overhead ratio** is the share that falls in the second and third columns. A charity that spends $100 with $15 on management and fundraising has an overhead ratio of 15 percent.\n\n## Why it became the default\n\nIt is printed on a public form, and it seems to answer the question donors most want answered: how much of my gift reaches the cause? Part IX does not itself measure results.\n\n## What it can’t tell you\n\nThe ratio says how costs are split, not what they buy. Some costs of running an organization well land in those columns, while shared costs must be allocated by function under the IRS instructions, and they are often what make a program work. A charity can spend almost nothing on overhead and achieve nothing, and another can invest heavily in systems and change thousands of lives.\n\nThe ratio also shapes behavior. Ann Goggins Gregory and Don Howard described a *nonprofit starvation cycle*: funders expect low overhead, charities feel pressure to meet the expectation, so they underspend on infrastructure and underreport what they do spend, which keeps expectations low. They cited a study in which more than a third of organizations reported no fundraising costs at all, and a 2001 survey by the Better Business Bureau in which more than half of American adults said overhead should be 20 percent or less.\n\n## An open letter\n\nIn June 2013 the leaders of GuideStar, Charity Navigator and the BBB Wise Giving Alliance, three organizations that donors turn to for information about charities, published an open letter calling the overhead percentage a poor measure of a charity’s performance. They asked donors to look instead at transparency, governance, leadership and results. They also said that very high overhead can still be a sign of fraud or poor management, so the ratio can prompt further questions without being a verdict on its own.\n\n## What could replace it\n\nFinding a simple public measure that tracks results without being easy to game is [Open Problem 6, the Overhead Replacement](/open-problems/the-overhead-replacement/). The raw material exists: [Charity Impact Media](https://theguysaccount.github.io/charity-impact-media/) publishes Form 990 spending breakdowns, with every number linked to its filing and no ratings attached. See also the [Myth-Busting Series](/record/myths/) on the claim that a low overhead ratio means a good charity." }, { "id": "participatory-grantmaking", "path": "/record/encyclopedia/participatory-grantmaking/", "title": "Participatory grantmaking", "summary": "Grantmaking in which funders hand some or all of the decisions about money, including strategy and criteria, to the communities the money is meant to serve.", "aliases": [ "participatory philanthropy", "community-led grantmaking", "shifting power" ], "conditions": [ "missing-voice", "missing-feedback-loop" ], "related": [ "/diagnosis/missing-voice/", "/open-problems/the-agency-premium/", "/record/encyclopedia/beneficiary-feedback/", "/record/lexicon/#participatory-grantmaking", "/record/lexicon/#trust-based-philanthropy" ], "sources": [ { "title": "New report on how funders can engage in participatory grantmaking", "url": "https://www.alliancemagazine.org/blog/new-report-on-how-funders-can-engage-in-participatory-grantmaking/", "publisher": "Alliance magazine, on Cynthia Gibson’s GrantCraft guide “Deciding Together”", "year": 2018 }, { "title": "Humanity’s Checklist: six things to fix so we can fix everything else", "url": "https://humanityschecklist.org/", "publisher": "Humanity’s Checklist", "year": 2026 }, { "title": "Deciding Together: Shifting Power and Resources Through Participatory Grantmaking", "url": "https://learningforfunders.candid.org/wp-content/uploads/sites/2/2018/12/DecidingTogether_Final_20181002.pdf", "publisher": "Cynthia Gibson, GrantCraft, Foundation Center", "year": 2018 } ], "review": "draft", "body": "In a conventional grant process, donors, foundation staff or boards decide where money goes. **Participatory grantmaking** moves some of those decisions to the people who will live with the results. In the version described by Cynthia Gibson’s 2018 guide for GrantCraft, *Deciding Together: Shifting Power and Resources Through Participatory Grantmaking*, funders cede decision-making power about funding to the communities they aim to serve.\n\n## What it looks like\n\nParticipation runs along a spectrum. At one end, community members advise a funder that still decides. Further along, they sit on the panels that review applications, set the criteria or choose the strategy. At the far end, a community-led group controls the money outright. Gibson’s guide, which Alliance magazine described as the first focused entirely on participatory grantmaking, drew on examples from many participatory funders, many of them smaller foundations focused on human rights. The Ford Foundation and Open Society Foundations supported the work.\n\n## The case for it\n\nCommunity members can bring knowledge about needs, risks and trusted organizations that a funder may lack. A process that lets them decide also changes who holds power, which some practitioners see as part of the result rather than a means to it. Humanity’s Checklist names gatekeepers as one of the six sources upstream of world problems: concentrated decision rights may exclude useful ideas. Participatory grantmaking applies the same idea to philanthropy.\n\n## The questions it raises\n\nWho speaks for a community, and how are they chosen? What does it cost in time and money to run a fair process, and who pays participants for their time? How does a funder handle conflicts of interest when panelists know the applicants? Practitioners answer these in different ways, and the answers shape the results.\n\n## What we don’t know\n\nThe guide presents principles and practitioners’ experience. It does not establish a general causal estimate of the effect on final outcomes. Do community-decided grants achieve more than expert-decided ones on the same outcome, and in which settings? That is [Open Problem 14, the Agency Premium](/open-problems/the-agency-premium/). See also [the Missing Voice](/diagnosis/missing-voice/) and [beneficiary feedback](/record/encyclopedia/beneficiary-feedback/)." }, { "id": "private-foundations", "path": "/record/encyclopedia/private-foundations/", "title": "Private foundations", "summary": "In U.S. law, charities that do not qualify as public charities, usually because they lack broad public support. They must generally spend an amount tied to 5 percent of their investment assets on charitable purposes each year.", "aliases": [ "foundations", "grantmaking foundations", "the 5 percent rule", "minimum distribution requirement" ], "conditions": [ "parked-money", "short-horizon" ], "related": [ "/diagnosis/parked-money/", "/open-problems/the-payout-question/", "/record/encyclopedia/donor-advised-funds/", "/record/lexicon/#private-foundation", "/record/lexicon/#payout-rate" ], "sources": [ { "title": "Private foundations", "url": "https://www.irs.gov/charities-non-profits/charitable-organizations/private-foundations", "publisher": "Internal Revenue Service" }, { "title": "Minimum investment return", "url": "https://www.irs.gov/charities-non-profits/private-foundations/minimum-investment-return", "publisher": "Internal Revenue Service" }, { "title": "Taxes on failure to distribute income, private foundations", "url": "https://www.irs.gov/charities-non-profits/private-foundations/taxes-on-failure-to-distribute-income-private-foundations", "publisher": "Internal Revenue Service" }, { "title": "Program-related investments", "url": "https://www.irs.gov/charities-non-profits/private-foundations/program-related-investments", "publisher": "Internal Revenue Service" }, { "title": "Key takeaways from Giving USA’s 2026 report on philanthropy", "url": "https://foundationsource.com/blog/key-takeaways-from-giving-usas-2026-report-on-philanthropy/", "publisher": "Foundation Source", "year": 2026 }, { "title": "The 5 Percent Foundation Payout Requirement May Be a Floor, but the Ceiling Is Awfully Low", "url": "https://inequality.org/article/foundation-payout/", "publisher": "Helen Flannery, Inequality.org (Institute for Policy Studies)", "year": 2023 }, { "title": "Gates to give away $200 billon and shutter foundation in 2045", "url": "https://www.alliancemagazine.org/blog/gates-to-give-away-200-billon-and-shutter-foundation-in-2045/", "publisher": "Alliance magazine", "year": 2025 }, { "title": "Phil Trammell on how becoming a ‘patient philanthropist’ might allow you to do far more good", "url": "https://80000hours.org/podcast/episodes/phil-trammell-patient-philanthropy/", "publisher": "80,000 Hours Podcast", "year": 2020 } ], "review": "draft", "body": "Under U.S. tax law, every 501(c)(3) organization is a **private foundation** unless it falls into an excluded category, such as hospitals, universities and organizations with broad public support. Giving USA estimates that foundations gave about $117 billion in 2025, roughly a fifth of all U.S. giving.\n\n## The 5 percent rule\n\nPrivate non-operating foundations generally must make qualifying distributions each year; operating foundations have different rules. The amount is tied to what the IRS calls a foundation’s *minimum investment return*: 5 percent of the value of its assets that are not used for charitable purposes, less debt incurred to buy them. A foundation that falls short owes a 30 percent excise tax on the undistributed amount under Section 4942 of the tax code, and a 100 percent tax if it does not make up the shortfall within 90 days of being notified.\n\nFoundations can also invest for their mission. A **program-related investment**, in the IRS’s terms, is one whose primary purpose is charitable, where earning income is not a significant purpose and which is not used for lobbying or political campaigns. The IRS’s examples include low-interest loans to needy students and high-risk investments in nonprofit low-income housing.\n\n## Floor or ceiling?\n\nThe rule sets a minimum. Whether it works in practice as a maximum is debated. An analysis of IRS data by Helen Flannery of the Institute for Policy Studies found that the median payout rate of private non-operating foundations ranged from 5.2 to 5.6 percent between 2015 and 2021. For foundations with more than a billion dollars, it was usually 5.0 or 5.1 percent.\n\n## Forever or now?\n\nA foundation whose investments earn more than it pays out can, in principle, last forever. Some choose not to. In May 2025 Bill Gates announced that the Gates Foundation would spend about $200 billion over 20 years and close permanently on December 31, 2045, much sooner than its charter had planned.\n\nThere are serious arguments on the other side. The economist Philip Trammell argues that a truly patient funder can invest, let returns compound and give far more later, when people may also know more about how to do good. He also lists the risks: a fund can be lost to war or financial collapse, or drift from its founder’s goals.\n\nWhich approach does more good is not known. It depends on how fast good opportunities are being used up, how fast invested money grows and how much the future should count. That is [Open Problem 12, the Payout Question](/open-problems/the-payout-question/). See also [Parked Money](/diagnosis/parked-money/) and [discounting](/record/encyclopedia/discounting/)." }, { "id": "randomized-controlled-trials", "path": "/record/encyclopedia/randomized-controlled-trials/", "title": "Randomized controlled trials", "summary": "Experiments that assign people or communities to a program or to a comparison group by chance, so that differences in outcomes can be credited to the program. They reshaped development economics, and their limits are still debated.", "aliases": [ "RCTs", "randomized evaluations", "field experiments" ], "conditions": [ "measurement-gap" ], "related": [ "/record/encyclopedia/counterfactual-impact/", "/record/principles/04-evidence-and-its-grades/", "/programs/evidence-reviews/", "/programs/replication-program/", "/open-problems/the-replication-rate/" ], "sources": [ { "title": "The Prize in Economic Sciences 2019", "url": "https://www.kva.se/en/news/ekonomipriset-2019-2/", "publisher": "The Royal Swedish Academy of Sciences", "year": 2019 }, { "title": "About us", "url": "https://www.povertyactionlab.org/about-us", "publisher": "Abdul Latif Jameel Poverty Action Lab (J-PAL)" }, { "title": "Understanding and misunderstanding randomized controlled trials", "url": "https://doi.org/10.1016/j.socscimed.2017.12.005", "publisher": "Angus Deaton and Nancy Cartwright, Social Science & Medicine 210", "year": 2018, "note": "Freely readable at https://pmc.ncbi.nlm.nih.gov/articles/PMC6019115/" }, { "title": "Insecticide-treated nets for preventing malaria (Cochrane systematic review)", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6418392/", "publisher": "Joseph Pryce, Marty Richardson and Christian Lengeler, Cochrane Database of Systematic Reviews", "year": 2018 }, { "title": "“Scared Straight” and other juvenile awareness programs for preventing juvenile delinquency", "url": "https://www.cochrane.org/evidence/CD002796_scared-straight-and-other-juvenile-awareness-programs-preventing-juvenile-delinquency", "publisher": "Anthony Petrosino and colleagues, Cochrane Database of Systematic Reviews", "year": 2022 } ], "review": "draft", "body": "A **randomized controlled trial** (RCT) assigns people, villages or schools to receive a program or not by chance, like a coin flip. Because chance decides, the groups are expected to be alike in everything except the program. The difference in their outcomes then estimates the program’s effect. A trial is a way of seeing the [counterfactual](/record/encyclopedia/counterfactual-impact/): what would have happened without the program.\n\n## In development economics\n\nIn 2019 the Royal Swedish Academy of Sciences awarded the Prize in Economic Sciences in Memory of Alfred Nobel to Abhijit Banerjee and Esther Duflo of MIT and Michael Kremer of Harvard “for their experimental approach to alleviating global poverty.” The Academy said they split poverty into smaller, more precise questions, often best answered by carefully designed experiments among the people most affected. It credited one of their studies with leading to remedial tutoring programs that have reached more than five million children in India. J-PAL, the research center Banerjee and Duflo founded at MIT in 2003 with Sendhil Mullainathan, says its network has led more than 2,700 randomized evaluations.\n\n## What trials can show\n\nTrials can confirm that something works and show how much. A Cochrane review of five trials with about 200,000 participants found that insecticide-treated nets cut child deaths from all causes by about 17 percent, and rated that evidence high-certainty.\n\nThey can also show that something does harm. A Cochrane review of nine trials of “Scared Straight” programs found that they increased offending compared with doing nothing.\n\n## The critique\n\nAngus Deaton and Nancy Cartwright argue that the public, and sometimes researchers, put too much trust in RCTs compared with other methods. Randomization does not guarantee that the groups in any one trial are equal in everything but the treatment. At best a trial gives an unbiased estimate, which they say is of limited practical value on its own. Its result applies to the people in the trial, and carrying it to a new place or population needs a further argument about why the program worked. They see trials as useful inside a cumulative program of research that combines them with other methods.\n\n## Where trials stop\n\nMany of the questions Charitology cares most about cannot be randomized. Some national laws and cultural changes cannot feasibly be assigned at random, though policy rollout or components can sometimes be randomized. Charitology hypothesizes that limits on trial evidence can affect funding for structure and [source](/record/encyclopedia/source-depth/) work. Finding other ways to measure it is [Open Problem 3](/open-problems/measuring-the-source/). How often trial findings hold up when repeated is [Open Problem 10](/open-problems/the-replication-rate/)." }, { "id": "restricted-and-unrestricted-funding", "path": "/record/encyclopedia/restricted-and-unrestricted-funding/", "title": "Restricted and unrestricted funding", "summary": "Restricted gifts must be spent on a purpose the funder names. Unrestricted gifts, also called general operating support, let the organization decide. The balance between them shapes what charities can learn, fix and attempt.", "aliases": [ "general operating support", "flexible funding", "project funding", "multiyear funding" ], "conditions": [ "restriction-reflex", "overhead-myth", "short-horizon" ], "related": [ "/diagnosis/restriction-reflex/", "/open-problems/the-optimal-restriction/", "/record/lexicon/#general-operating-support", "/record/lexicon/#trust-based-philanthropy", "/tools/grant-evaluator/" ], "sources": [ { "title": "New Attitudes, Old Practices: The Provision of Multiyear General Operating Support", "url": "https://www.ncfp.org/download/new-attitudes-old-practices-cep-2020-0", "publisher": "Center for Effective Philanthropy, via the National Center for Family Philanthropy", "year": 2020 }, { "title": "Top Foundations Pledge Flexible Funding to Grantees in Wake of COVID-19 Crisis", "url": "https://www.fordfoundation.org/news-and-stories/news-and-press/news/top-foundations-pledge-flexible-funding-to-grantees-in-wake-of-covid-19-crisis/", "publisher": "Ford Foundation", "year": 2020 }, { "title": "A Call to Action: Philanthropy’s Commitment During COVID-19", "url": "https://www.cof.org/news/call-action-philanthropys-commitment-during-covid-19", "publisher": "Council on Foundations" }, { "title": "Six Practices of Trust-Based Grantmaking", "url": "https://www.trustbasedphilanthropy.org/practices", "publisher": "Trust-Based Philanthropy Project" }, { "title": "The Nonprofit Starvation Cycle", "url": "https://ssir.org/articles/entry/the_nonprofit_starvation_cycle", "publisher": "Ann Goggins Gregory and Don Howard, Stanford Social Innovation Review", "year": 2009 } ], "review": "draft", "body": "A **restricted** gift comes with conditions: it must pay for a named project, often within a set time and budget. An **unrestricted** gift, or **general operating support**, can be spent on whatever the organization judges most useful, including salaries, systems and evaluation. When unrestricted support is promised for several years at a time, it is called multiyear general operating support.\n\n## The case for each\n\nRestriction feels like accountability. A funder can see exactly what its money bought, and it can be sure the gift serves the goal it cares about. That instinct is responsible.\n\nThe case for flexibility is that organizations learn. A charity halfway through a project may find a better approach, or discover that its accounting system is about to fail. A restriction may require the charity to seek the funder’s agreement before changing how the money is used. A grant that does not cover its share of running costs leaves those costs to be funded elsewhere. Ann Goggins Gregory and Don Howard found that foundation allowances for indirect costs averaged 10 to 15 percent of each grant. In four youth-serving nonprofits they studied, actual overhead ran from 17 to 35 percent. They called the result the nonprofit starvation cycle.\n\n## Attitudes and practice\n\nIn a 2020 report by the Center for Effective Philanthropy, 58 percent of foundation CEOs said their foundations gave some multiyear general operating support, usually to only a few grantees. Among the 58 percent that provided it, 11 percent gave it to more than half their grantees. The survey covered foundations giving at least $5 million a year, not all funders.\n\nThen came a natural experiment. In March 2020, as the COVID-19 pandemic spread, a pledge led by the Ford Foundation with the Council on Foundations, and informed by the Trust-Based Philanthropy Project, asked funders to loosen or eliminate restrictions on current grants, make new grants as unrestricted as possible and cut reporting demands. The Council published the signatories to the pledge. The first of the Trust-Based Philanthropy Project’s six practices is to give multi-year, unrestricted funding.\n\n## What we don’t know\n\nThe case for unrestricted funding is strong, but most of it rests on what charities say they prefer, not on outcomes. Restriction may help in some situations, for example when a funder brings knowledge the grantee lacks. The sources here do not establish when each works better on final outcomes. The shift of 2020 makes comparison possible: what happened to grantees whose funding became flexible, compared with similar grantees whose funding did not? That is [Open Problem 8, the Optimal Restriction](/open-problems/the-optimal-restriction/)." }, { "id": "psychic-numbing", "path": "/record/encyclopedia/psychic-numbing/", "title": "Psychic numbing", "summary": "The way concern for people in danger fails to grow with their number, and can even shrink, so that a single person in need moves us more than a mass tragedy. The psychologist Paul Slovic has used it to explain how people respond to suffering on a large scale.", "aliases": [ "compassion fade", "compassion collapse", "scope insensitivity" ], "conditions": [ "story-bias", "symptom-trap" ], "related": [ "/diagnosis/story-bias/", "/record/encyclopedia/identifiable-victim-effect/", "/open-problems/story-without-bias/", "/record/lexicon/#psychic-numbing" ], "sources": [ { "title": "“If I look at the mass I will never act”: Psychic numbing and genocide", "url": "https://journal.sjdm.org/7303a/jdm7303a.htm", "publisher": "Paul Slovic, Judgment and Decision Making 2(2)", "year": 2007 }, { "title": "Compassion Fade: Affect and Charity Are Greatest for a Single Child in Need", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC4062481/", "publisher": "Daniel Västfjäll, Paul Slovic, Marcus Mayorga and Ellen Peters, PLoS ONE 9(6)", "year": 2014 } ], "review": "draft", "body": "If every life counts equally, saving a hundred lives should matter a hundred times as much as saving one. Feelings don’t work that way. **Psychic numbing** is the name for the gap: concern fails to keep pace with the number of people suffering, and past a point it can fall.\n\n## The idea\n\nRobert J. Lifton coined the term for the turning off of feeling that let rescue workers keep functioning after Hiroshima. In a 2007 paper the psychologist Paul Slovic used it more broadly, for the pattern in which emotional response shrinks as the number of victims grows, leading to apathy. The paper takes its title from a line attributed to Mother Teresa: “If I look at the mass I will never act.”\n\nSlovic argued that people value lives along a psychophysical curve, the way they perceive other quantities: each additional life registers less than the one before. Few people would feel any difference between saving 87 lives and saving 88. Applied to atrocities, he argued, the pattern helps explain why the world largely stood by as about 800,000 people were killed in Rwanda in roughly 100 days. Statistics about mass killing, however large, fail to produce the feeling that drives action.\n\n## Compassion fade\n\nThe effect may start very early. In studies published in 2014, Daniel Västfjäll, Paul Slovic, Marcus Mayorga and Ellen Peters found that feelings of compassion, and in some studies donations, were greatest for a single child in need and began to decline with the second endangered life. The decline reversed when several people were presented as one unit, such as a family. Some of the studies used real money and some used hypothetical choices.\n\nRelated findings, such as the [identifiable victim effect](/record/encyclopedia/identifiable-victim-effect/), are now being tested again, and at least one prominent replication found no effect. How strong and reliable each finding is remains an open research question.\n\n## Why it matters for the system\n\nA charity system that moves money through human attention inherits the limits of human attention. If concern falls as need grows, the largest problems will draw the least feeling per person affected, and the deepest ones, which are often abstract and statistical, the least of all. This is part of the [Story Bias](/diagnosis/story-bias/), and a reason the [Symptom Trap](/diagnosis/symptom-trap/) is plausible.\n\nKnowing about the bias does not switch it off. Can a story make people feel the size of a problem rather than one face? Formats worth testing include stories of whole communities, visualizations of scale and stories about sources rather than symptoms. That is [Open Problem 7, Story Without Bias](/open-problems/story-without-bias/)." }, { "id": "source-depth", "path": "/record/encyclopedia/source-depth/", "title": "Source depth", "summary": "Charitology’s five-level scale for where a piece of work acts on the chain from symptom to source, from relief through repair, prevention and structure to source. It describes the work. It never grades it.", "aliases": [ "depth scale", "depth", "relief to source" ], "conditions": [ "symptom-trap", "measurement-gap" ], "related": [ "/method/", "/record/encyclopedia/symptoms-and-sources/", "/record/principles/03-depth-from-relief-to-source/", "/open-problems/the-depth-census/", "/open-problems/measuring-the-source/" ], "sources": [ { "title": "Humanity’s Checklist: six things to fix so we can fix everything else", "url": "https://humanityschecklist.org/", "publisher": "Humanity’s Checklist", "year": 2026 } ], "review": "draft", "body": "**Source depth** says where on the chain from [symptom to source](/record/encyclopedia/symptoms-and-sources/) a piece of work acts. Charitology uses a five-level scale, numbered 0 to 4, set out in [the Method](/method/#depth). Every gift, program and organization can be described with it.\n\n## The five levels\n\n| Level | Name | What it does | Examples |\n| --- | --- | --- | --- |\n| 0 | Relief | Meets the need today | A meal at a food bank, emergency surgery, a disaster shelter |\n| 1 | Repair | Restores what was lost | Rebuilding a home, rehabilitation after an injury, a job after a layoff |\n| 2 | Prevention | Stops it happening to these people | Bed nets, vaccines, clean water for a village |\n| 3 | Structure | Changes the rules that produce it | A lead-paint ban, a new public health system, a corporate standard |\n| 4 | Source | Changes what feeds many problems at once | The story a society tells itself, how people coordinate, what a culture rewards |\n\nEach level has a test question. Relief asks whether someone is fed, sheltered or treated right now. Repair asks whether someone is back where they were before the harm. Prevention asks whether these people will avoid the harm altogether. Structure asks whether a law, institution, market or norm changed so the harm stops being produced. Source asks whether something upstream of a whole family of problems changed.\n\n## Depth describes. It never grades.\n\nEvery level is needed. Immediate food matters to a hungry person tonight, and a change at the source can take a generation to arrive. A deeper level is not a better level, and nothing in the scale says a food bank matters less than a policy campaign. The scale exists so that a giver knows what they are paying for, and so the field can ask whether the system’s overall mix is right. That is the question behind the [Symptom Trap](/diagnosis/symptom-trap/).\n\n## How to place a piece of work\n\n- **Classify the activity, not the cause.** One cause has work at every depth. Homelessness has emergency shelters (relief), rehousing (repair), eviction prevention (prevention) and zoning reform (structure).\n- **Classify what the money does, not the mission statement.** One organization can work at several depths. Weight each activity by what it spends.\n- **Say when it is hard to call.** Boundaries between levels are fuzzy. How often two independent people place the same work at the same level is itself worth measuring.\n\n## The top of the scale\n\nLevel 4 comes from Humanity’s Checklist, which argues that world problems grow from six shared sources: the story we tell, how we coordinate, what we can picture, what our feeds reward, what we admire and who gets to build. Charitology treats the claim that changing them would help everything downstream as a hypothesis to test.\n\n## What this project has not established\n\nCharitology has not produced a public estimate of the world’s giving by depth and has not established the current mix. Measuring it is [Open Problem 1, the Depth Census](/open-problems/the-depth-census/). Measuring the effect of source-level work, which is diffuse, shared and slow, is [Open Problem 3](/open-problems/measuring-the-source/)." }, { "id": "scale-neglectedness-tractability", "path": "/record/encyclopedia/scale-neglectedness-tractability/", "title": "Scale, neglectedness and tractability", "summary": "A framework for comparing problems by how big they are, how few resources already go to them and how much extra effort would achieve. It grew out of GiveWell Labs and 80,000 Hours and is widely used in effective altruism.", "aliases": [ "SNT framework", "ITN framework", "importance, neglectedness and tractability", "scale, neglectedness and solvability" ], "conditions": [ "single-lens", "symptom-trap" ], "related": [ "/record/encyclopedia/neglectedness/", "/record/encyclopedia/effective-altruism/", "/open-problems/the-exchange-rate-problem/", "/tools/values-compass/", "/record/lexicon/#tractability" ], "sources": [ { "title": "A framework for comparing global problems in terms of expected impact", "url": "https://80000hours.org/articles/problem-framework/", "publisher": "Robert Wiblin, 80,000 Hours", "year": 2016, "note": "First published in 2016 and last updated in December 2025." }, { "title": "A framework for strategically selecting a cause", "url": "https://80000hours.org/2013/12/a-framework-for-strategically-selecting-a-cause/", "publisher": "Benjamin Todd, 80,000 Hours", "year": 2013 }, { "title": "Narrowing down U.S. policy areas", "url": "https://blog.givewell.org/2014/05/22/narrowing-down-u-s-policy-areas/", "publisher": "Holden Karnofsky, The GiveWell Blog", "year": 2014 }, { "title": "The Important/Neglected/Tractable framework needs to be applied with care", "url": "https://forum.effectivealtruism.org/posts/74oJS32C6CZRC4Zp5/the-important-neglected-tractable-framework-needs-to-be", "publisher": "Robert Wiblin, Effective Altruism Forum", "year": 2016 }, { "title": "Open Philanthropy Is Now Coefficient Giving", "url": "https://forum.effectivealtruism.org/posts/vkvtu6xbvfkHPhJkC/open-philanthropy-is-now-coefficient-giving", "publisher": "Effective Altruism Forum", "year": 2025 } ], "review": "draft", "body": "The **scale, neglectedness and tractability** framework asks three questions about a problem before asking which charity to fund. In the version 80,000 Hours publishes:\n\n- **Scale:** If we solved the problem, how good would it be?\n- **Neglectedness:** How many people, or dollars, are currently working on it? See [neglectedness](/record/encyclopedia/neglectedness/).\n- **Solvability**, often called **tractability:** If we doubled the effort going into it, what fraction of what remains would we expect to solve?\n\nMultiplied together, the three give an estimate of how much good an extra unit of money or work would do. Because the scores are logarithmic, 80,000 Hours adds them instead of multiplying.\n\n## Where it came from\n\nGiveWell Labs, the project that became the Open Philanthropy Project (later Open Philanthropy, and since 2025 Coefficient Giving), compared causes by importance, tractability and how crowded they were. A 2014 post used those three criteria to narrow down U.S. policy areas. In December 2013 Benjamin Todd of 80,000 Hours proposed a cause-selection framework of importance, tractability and uncrowdedness, crediting GiveWell Labs for the names. Robert Wiblin’s 2016 article for 80,000 Hours turned it into the formal version still in use.\n\n## Cautions from its own authors\n\n80,000 Hours lists the limits plainly. The scores are imprecise, and adding them adds uncertainty. Narrowly defined problems tend to score better than broad ones. Trade-offs between the yardsticks rest on contested value judgments. Solvability is the hardest factor to score, and the true differences between problems are probably smaller than the scores suggest. Wiblin also warns that the factors overlap: a problem can stay important because it is neglected, so the same fact can be counted twice.\n\n## How Charitology uses it\n\nCharitology treats the framework as a good tool that needs one more input: a moral framework. *Scale* needs a unit, such as lives, wellbeing, rights secured or species protected, and choosing the unit is a moral choice. A problem can rank first under one lens and last under another. So Charitology asks the questions once per framework. Its frameworks data, used by the [Values Compass](/tools/values-compass/), sketches how each framework tends to rank causes, so a giver can see where the rankings agree. Where they cannot be put on one scale at all is [Open Problem 2, the Exchange Rate Problem](/open-problems/the-exchange-rate-problem/).\n\nThe framework scores problems, not the depth of the work. A problem can be crowded at the relief level and neglected at the source level. Charitology’s working hypothesis is that this pattern is common, which is the [Symptom Trap](/diagnosis/symptom-trap/). This project has not established that pattern." }, { "id": "symptoms-and-sources", "path": "/record/encyclopedia/symptoms-and-sources/", "title": "Symptoms and sources", "summary": "The core distinction of charitology. A symptom is a problem as it shows up in people’s lives. A source is what keeps producing it upstream.", "aliases": [ "root causes", "upstream causes" ], "conditions": [ "symptom-trap" ], "related": [ "/method/", "/record/encyclopedia/source-depth/", "/record/cause-atlas/" ], "sources": [ { "title": "Humanity’s Checklist: six things to fix so we can fix everything else", "url": "https://humanityschecklist.org/", "publisher": "Humanity’s Checklist", "year": 2026 } ], "review": "draft", "body": "A **symptom** is a problem as it shows up in someone’s life: hunger tonight, an untreated fever, a flooded home. A **source** is something upstream that keeps producing problems, often many at once.\n\n## Why the distinction matters\n\nSymptoms are urgent, visible and real, and helping with them is good. Some needs recur because the conditions producing them persist. A food bank feeds people tonight, and tomorrow there is a line again. Charitology asks every gift where it acts on the chain from symptom to source, and asks the system as a whole whether its mix is right. That is the [Symptom Trap](/diagnosis/symptom-trap/).\n\n## The six sources\n\nHumanity’s Checklist argues that the world’s problems grow from a few shared sources inside humanity’s collective life:\n\n- **Story**: the story we tell ourselves about what is possible\n- **Coordination**: whether the people who could solve a problem together ever meet\n- **Vision**: whether we can picture the future we want to build\n- **Attention**: what our feeds and media reward\n- **Status**: what a culture admires\n- **Gatekeepers**: who gets to build and be seen\n\nFixing a source, the checklist argues, stops it feeding the problems below it and starts it feeding the cure. Charitology adopts these six as the top of its [depth scale](/record/encyclopedia/source-depth/) and treats the claim that fixing them would help everything downstream as a hypothesis to test.\n\n## Between symptom and source\n\nMost work sits in between. Charitology’s depth scale has five levels: relief, repair, prevention, structure and source. Bed nets are prevention: they stop malaria for the people who sleep under them. A ban on lead paint is structure: it changes the rules that produced the harm. The scale describes where the work acts. It does not say which level is better. A healthy system funds all five on purpose." }, { "id": "systemic-change", "path": "/record/encyclopedia/systemic-change/", "title": "Systemic change", "summary": "Change to the structures that produce outcomes, such as laws, markets, institutions, relationships and beliefs, rather than to outcomes one case at a time. On Charitology’s depth scale it is the work of the structure and source levels.", "aliases": [ "systems change", "structural change", "root-cause change" ], "conditions": [ "symptom-trap", "short-horizon", "measurement-gap" ], "related": [ "/record/encyclopedia/source-depth/", "/record/encyclopedia/symptoms-and-sources/", "/open-problems/measuring-the-source/", "/open-problems/when-charity-harms/", "/record/lexicon/#leverage" ], "sources": [ { "title": "Leverage Points: Places to Intervene in a System", "url": "https://www.donellameadows.org/archives/leverage-points-places-to-intervene-in-a-system/", "publisher": "Donella Meadows, The Sustainability Institute", "year": 1999 }, { "title": "The Water of Systems Change", "url": "https://www.fsg.org/blog/new-article-water-systems-change/", "publisher": "John Kania, Mark Kramer and Peter Senge, FSG", "year": 2018 }, { "title": "Six Conditions for Systems Change", "url": "https://www.complexsystemsframeworks.ca/framework/six-conditions-for-systems-change/", "publisher": "Complex Systems Frameworks Collection" }, { "title": "Effective Altruism and Its Critics", "url": "https://doi.org/10.1111/japp.12176", "publisher": "Iason Gabriel, Journal of Applied Philosophy 34(4)", "year": 2017 } ], "review": "draft", "body": "**Systemic change** aims at the structures that keep producing those outcomes: the rules, markets, institutions, relationships and beliefs underneath. On the [depth scale](/record/encyclopedia/source-depth/) it is the work of level 3, structure, and level 4, source.\n\n## Where to push\n\nThe systems thinker Donella Meadows described *leverage points* as places in a complex system where a small shift can produce big changes. In a 1999 essay she ranked twelve of them, from least to most effective. At the bottom are constants and numbers, such as subsidies, taxes and standards. Further up come feedback loops, the structure of information flows, the rules of the system, its power to reorganize itself, its goals and the mindset it grows from. At the top is the power to transcend paradigms altogether. She observed that almost all attention goes to the numbers, where there is little leverage.\n\nIn 2018 John Kania, Mark Kramer and Peter Senge, writing for FSG, described six interdependent conditions that hold social and environmental problems in place: policies, practices and resource flows, which are the most visible; relationships and connections, and power dynamics; and, deepest, mental models, the beliefs people take for granted.\n\nBoth maps inform Charitology’s scale. They are frameworks for reasoning about change, not evidence that a deeper intervention always achieves more.\n\n## The debate\n\nWhether giving should aim at systems is a live argument. Iason Gabriel summarized a critique of effective altruism along these lines: that its language of charity can hide the structural causes of poverty, and that a movement which changes its priorities as the evidence changes may lack the constancy of purpose that movements such as abolitionism and civil rights relied on. He also noted that effective altruists were increasingly interested in systemic change, and argued for a division of labor and partnerships with committed advocacy groups.\n\nThe case for caution is just as real. Systemic bets are slow, their credit is shared and they can fail. Some do harm: a reform can backfire, and charity can crowd out or delay the public action that would have helped more. That is [Open Problem 20, When Charity Harms](/open-problems/when-charity-harms/).\n\n## How Charitology sees it\n\nCharitology does not treat systemic change as better than direct help. It treats the line between them as a scale rather than a wall, and every level as needed. Its working hypothesis, the [Symptom Trap](/diagnosis/symptom-trap/), is that the system as a whole underfunds the deep end of the scale, partly because systemic change is so hard to measure. Measuring it is [Open Problem 3](/open-problems/measuring-the-source/). The eradication of smallpox is a useful case of a lasting change. Its proposed source-depth classification is Charitology’s hypothesis, open to challenge: see [the case study](/record/case-studies/smallpox-eradication/)." }, { "id": "warm-glow-giving", "path": "/record/encyclopedia/warm-glow-giving/", "title": "Warm-glow giving", "summary": "The satisfaction people get from the act of giving itself, apart from what the gift achieves. The economist James Andreoni built it into a theory of giving in 1990, It is one possible input to Charitology’s hypotheses about giving signals.", "aliases": [ "warm glow", "impure altruism", "joy of giving" ], "conditions": [ "prestige-economy", "missing-feedback-loop", "story-bias" ], "related": [ "/diagnosis/prestige-economy/", "/open-problems/the-status-flip/", "/record/encyclopedia/identifiable-victim-effect/", "/record/lexicon/#warm-glow", "/tools/charitology-wrapped/" ], "sources": [ { "title": "Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving?", "url": "https://doi.org/10.2307/2234133", "publisher": "James Andreoni, The Economic Journal 100(401)", "year": 1990 }, { "title": "Paying Taxes, According To The Brain, Can Bring Satisfaction", "url": "https://www.eurekalert.org/news-releases/465670", "publisher": "University of Oregon, via EurekAlert, on William Harbaugh, Ulrich Mayr and Dan Burghart in Science", "year": 2007 } ], "review": "draft", "body": "Why do people give? One answer is that they care about the result: fewer hungry children, a cleaner river. Another is that giving feels good, whatever the result. The economist James Andreoni called the second motive the **warm glow** and built it into a formal theory in 1990.\n\n## Pure and impure altruism\n\nEconomists had often modeled donors as *pure altruists*, who care only about how much of a public good exists, no matter who provides it. Andreoni’s model of *impure altruism* adds the satisfaction of giving itself, along with motives such as social pressure, guilt and sympathy. He argued that it fits observed patterns of giving better than the pure model, and that it changes the answers to policy questions such as how private giving responds when government funds the same cause and how giving should be taxed.\n\n## Evidence from the brain\n\nIn a 2007 study published in *Science*, William Harbaugh, Ulrich Mayr and Dan Burghart gave 19 women $100 each and scanned their brains while they watched money go to a food bank. Reward areas of the brain responded even when the transfer was mandatory, like a tax, which fits pure altruism. They responded more when the women chose to give, which fits the warm glow.\n\n## Why it matters for the system\n\nThe warm glow is not a flaw. It is part of what makes people generous at all, and Charitology has no interest in removing it. The question is what the glow attaches to. Charitology’s working claim is that today it attaches mainly to the act of giving and to its size, which donors can see, rather than to its results, which they usually cannot. That is one way the [Missing Feedback Loop](/diagnosis/missing-feedback-loop/) shapes behavior: if the reward arrives the moment the gift is made, nothing pulls the donor to find out what happened next.\n\nThe [Prestige Economy](/diagnosis/prestige-economy/) is the social version of the same pattern, and Charitology’s hypothesis is that it pulls money toward visible gifts rather than effective ones. The warm glow itself is well documented. The sources here do not establish whether visible giving shifts money away from better opportunities or by how much.\n\n## Pointing the glow at results\n\nIf the glow can attach to the act, it may be able to attach to the outcome. Ideas worth testing include showing donors what their gift achieved, celebrating results rather than amounts and making effective giving a status signal. Whether any of them work at scale is [Open Problem 15, the Status Flip](/open-problems/the-status-flip/). Charitology’s [Wrapped](/tools/charitology-wrapped/) card is one small experiment in that direction." }, { "id": "theory-of-change", "path": "/record/encyclopedia/theory-of-change/", "title": "Theory of change", "summary": "A written account of how and why a program’s activities are expected to lead to its long-term goal, step by step, with the assumptions that link each step to the next.", "aliases": [ "ToC", "causal pathway", "change model" ], "conditions": [ "measurement-gap", "symptom-trap" ], "related": [ "/record/encyclopedia/outputs-and-outcomes/", "/record/encyclopedia/source-depth/", "/tools/impact-measurement-toolkit/", "/tools/grant-evaluator/", "/record/lexicon/#theory-of-change" ], "sources": [ { "title": "What is Theory of Change?", "url": "https://www.theoryofchange.org/what-is-theory-of-change/", "publisher": "Center for Theory of Change" }, { "title": "TOC Origins", "url": "https://www.theoryofchange.org/what-is-theory-of-change/toc-background/toc-origins/", "publisher": "Center for Theory of Change" }, { "title": "Logic Model Development Guide", "url": "https://www.naccho.org/uploads/downloadable-resources/Programs/Public-Health-Infrastructure/KelloggLogicModelGuide_161122_162808.pdf", "publisher": "W.K. Kellogg Foundation", "year": 2004 } ], "review": "draft", "body": "A **theory of change** explains how a program expects to get from what it does to what it is trying to achieve. The Center for Theory of Change describes it as an account of how and why a desired change is expected to happen in a particular context. It fills in the “missing middle” between activities and long-term goals. The usual method works backward: start from the goal, then name the outcomes that must be in place for it to happen and how each one causes the next.\n\n## Where it came from\n\nCarol Weiss popularized the term in 1995. She sat on the evaluation steering committee of the Aspen Institute’s Roundtable on Community Change, and her essay on theory-based evaluation appeared in the Roundtable’s book *New Approaches to Evaluating Comprehensive Community Initiatives*. Weiss argued that complex community programs were hard to evaluate because the assumptions behind them were poorly spelled out. People were unclear about how change would unfold, so they paid little attention to the early and middle steps a long-term goal depends on.\n\n## Theory of change and logic models\n\nA logic model lays out the chain of resources, activities, outputs, outcomes and impact (see [outputs and outcomes](/record/encyclopedia/outputs-and-outcomes/)). The Kellogg Foundation’s guide notes that outcome-focused models are built with a theory of change in mind, though it is usually not spelled out, and it describes a “theory approach” model that puts the planners’ assumptions up front. A theory of change makes the *why* explicit: each arrow in the chain is a claim that can be checked.\n\n## How to read one\n\nA good theory of change lets a giver ask sharp questions:\n\n- **Which link is weakest?** A chain is only as strong as its least supported step.\n- **What evidence backs each link?** Some steps have trial evidence. Others are hopes.\n- **Where does the goal sit on the [depth scale](/record/encyclopedia/source-depth/)?** Ask how many uncertain links connect the activity to the goal; depth alone does not determine that number.\n- **What would show it is wrong?** A theory that no result could contradict is not doing any work.\n\n## Why it matters for source work\n\nSome structure and source goals involve long chains and many assumptions. A campaign to change a norm may need a dozen links to hold, over many years, with credit shared among many actors. A written theory of change does not make that work easier to measure, but it shows which links to watch. Measuring them is [Open Problem 3](/open-problems/measuring-the-source/). See also the [Grant Evaluator](/tools/grant-evaluator/) and the [Impact Measurement Toolkit](/tools/impact-measurement-toolkit/)." } ] }