The Actual Number is Almost Surely Higher

Harrison Nathan
44 min readDec 9, 2016

An Evaluation of Effective Animal Activism

This paper analyzes efforts to apply the principles of Effective Altruism to advocacy on behalf of animals. A particular focus is on the work of Animal Charity Evaluators (ACE), formerly known as Effective Animal Activism, a meta-charity whose stated purpose is to promote effective animal advocacy by using “science to analyze the impact of interventions (i.e. tactics) to help animals.”[1] The paper also briefly discusses related organizations which avow a similar philosophy, including the Effective Altruism Foundation and the Open Philanthropy Project. It is not possible to cover every aspect of the subject.

Table of contents.

1 The hypothesis that current meat reduction interventions have an effect on behavior is not supported by the available evidence

2 ACE and Nick Cooney have distorted research results

3 ACE’s estimates of the effectiveness of other interventions may lack any empirical basis

4 Effective Altruist analyses of the direct benefit of cage-free reforms do not accurately reflect the relevant science

5 Charity evaluation may damage organizations’ actual effectiveness

6 The current Effective Altruist thinking on social norms and social movements lacks rigor

7 The current Effective Altruist thinking on wild animal issues lacks rigor

8 Effective Animal Activism is fraught with severe conflicts of interest

9 Conclusion

Section 1: The hypothesis that current meat reduction interventions have an effect on behavior is not supported by the available evidence

Animal Charity Evaluators has studied the impact of three interventions aimed at reducing meat consumption: online ads, humane education, and leafleting. The studies which influence their analysis are briefly reviewed here. Some additional studies by The Humane League are also discussed.

Online ads

The only controlled study of the impact of vegan online ads was commissioned by Mercy for Animals, which invests heavily in this intervention, and run in the summer of 2015.[2] The experimental group, which saw a pro-vegan video, later reported eating slightly more pork, dairy, eggs, and chicken than the control group, but the reported differences in behavior were not statistically significant. Mercy for Animals pointed out reasons why the negative result should not be taken as evidence against the effectiveness of online ads; namely, that self-reporting of one’s diet is known to be extremely unreliable and subject to bias, and that, in general, infeasibly large sample sizes are necessary in order to measure the effectiveness of online advertizing. These concerns, while entirely valid, should lead us to question why the study was commissioned at all.

Humane education

ACE cites two studies on the effectiveness of humane education. The first, conducted by Justice for Animals in summer 2013, was an online survey filled out by 114 out of 2200 students across several U.S. states who had viewed various humane education talks, and had no control group. 16 of the 114 respondents said they became vegetarian as a result of the talk, and ACE noted that this figure might be skewed upwards by social desirability bias.[3] ACE conducted another small study in fall 2013, which did use a control group. Its analysis of the self-reported dietary changes concluded that “The differences in percentage change are small and in most cases are in the opposite of the direction we would have predicted if humane education has an effect.” No statistically significant impact on behavior was observed. Curiously, ACE cites the uncontrolled Justice for Animals study as providing “the most relevant data we have,” and reasons that its results suggest that the number of students who become vegetarian as a result of humane ed talks is between 0.7% (16 out of 2200) and 14% (16 out of 114).[4]


ACE appears confident that leafleting is a cost-effective intervention, writing that “The existing evidence on the impact of leafleting is among the strongest bodies of evidence bearing on animal advocacy methods.”[5] Apart from anecdotal evidence, ACE cites two quantitative studies. The more recent one was conducted by ACE itself in fall 2013, and attempted to use a control group and to measure social desirability bias using RAND SDRS-5. However, only 23 respondents reported receiving the control flyer and not the Vegan Outreach leaflet. ACE still compared respondents who reported having received a Vegan Outreach leaflet against those who reported having received no leaflet at all, but in their words: “We did not find support for claims that distributing leaflets from Vegan Outreach (or similar leaflets published by other groups) causes the overall population who receives the leaflets to reduce their meat consumption. Rather, we found an overall pattern of reported reduction in meat consumption throughout the sample.”[6]

The results of this study do not figure into ACE’s quantitative cost-effectiveness estimates for leafleting. Those are based entirely on a Fall 2012 study by Farm Sanctuary and The Humane League (THL) which attempted to measure the relative effectiveness of two different leaflets: Compassionate Choices (a Vegan Outreach staple) and Something Better (the then-new predecessor to Your Choice, which is currently the most common Vegan Outreach leaflet distributed at US colleges.) Students were surveyed two to three months after their schools were leafleted, and asked what dietary changes they made in response to the leaflet they received. There was no control group, and no data was collected from students who did not report receiving one of the leaflets. Since the study did not use a control group, its results cannot be taken as evidence for the effectiveness of Vegan Outreach leaflets. It nevertheless continues to be the basis of such claims, and is discussed in detail in the next section.

Other studies

Animal welfare groups, especially The Humane League, have produced a number of other studies which use similarly flawed methodology, which I do not analyze in detail. Broadly speaking, these studies presume that the general types of outreach preferred by these groups are effective at getting people to change their diets, and attempt to compare marginally different messages. Because these types of studies have failed to establish that any outreach method has an effect at all, little or nothing of value can be discerned from their results. One of them, reported in May 2014, compared individuals’ self-reported dietary changes after receiving several different leaflets, and found that the control group, which received no leaflet at all, reported the greatest dietary change.[7] Another, reported in September 2015, attempted to determine whether it is better to ask people to “cut out or cut back on” meat, “eat less” meat, go vegetarian, or go vegan. The researchers reported that the difference between the first and fourth of these suggestions was “trending toward statistical significance.” Once again, the control group, which received no request at all, showed by far the greatest reduction in meat consumption.[8] In spite of this, researchers continue to regard basic scientific standards as optional. For example, a Vegan Outreach study conducted in Summer 2016 compared four different leaflets but did not use a control leaflet due to cost concerns.[9]

It is interesting to note that the controlled studies often seem to show a pattern of reported reduction in consumption of all animal products, and that the vegetarian messages often appear to lessen this effect, as though they caused people to eat more meat. It’s equally interesting to note the manner in which THL has reported such results: it has described them as “anomalies” or “noise.” The May 2014 report tells us that

“We do not believe that booklets promoting vegetarian eating actually decrease people’s likelihood of changing their diet. Other studies which intentionally examined this question have found that people who receive a pro-vegetarian booklets are more likely to reduce their meat consumption than those in a control group.”

But the link is to a blog post about the botched Animal Charity Evaluators leafleting study from Fall 2013, which attempted to use a control group but got insufficiently many respondents. The comment also obfuscates the purpose of a control group: if the study fails to show that the experimental group performs better than the control, we cannot claim the study provides evidence that there is an effect, let alone evidence that one method is more effective than another.

It’s cause for concern that these non-results have been widely repeated as justification for a particular kind of messaging. For example, Mercy for Animals cites them as part of its reasons for avoiding “the V word.”[10]

Section 2: ACE and Nick Cooney have distorted research results

Animal Charity Evaluators’ cost-effectiveness estimate for leafleting rests entirely on one uncontrolled study (as mentioned above) which was conducted by Farm Sanctuary and The Humane League in Fall 2012, and involved giving a questionnaire to college students two to three months after their school was leafleted.

This study is the basis for the commonly cited statistics that every two leaflets distributed spares one animal from a factory-farmed life, and that one in fifty leaflets distributed produces a vegetarian or pescatarian. Nick Cooney ebulliently reported the results in a January 2013 blog post: “The data is in. The facts are there. College leafleting is an absurdly effective activity….”[11] After reporting that fifty animals can be saved by distributing one hundred leaflets, he remarked “The actual number is almost surely higher.”[12] In fact, the study had suggested that 100 leaflets would save 141 animals, making the famous 2:1 figure a “conservative estimate” (in Cooney’s words.)

The study’s lack of controls and failure to account for endogeneity and observer bias were noticed by some members of the community, and criticized in blog posts.[13][14] However, because the Effective Animal Activism community lacks any rigorous standards, these criticisms did not lead to a retraction. It unfortunately appears necessary to debunk Cooney’s claims in some detail.

On the survey sheet, the options “I eat a little less” and “I eat a lot less” lie in the middle of the range of responses, likely prompting respondents to regard them as typical. Worse still, the leaflet covers (which are, after all, vegan propaganda) appear on the same sheet, carrying messages such as “making a difference for animals” and “millions of people are changing what they eat.” Not to mention, eyes. In Change of Heart, Nick Cooney cites research showing that images of eyes can have a dramatic effect on ethical behavior.[15] Moreover, the question is phrased so that it doesn’t merely ask how much meat the respondent eats, but rather, how much the booklet influenced him to change. This presumes that it ought to influence him to eat less meat, and also confuses two issues: whether the respondent thinks it’s a good idea to eat less meat, and whether he actually does.

Plainly, a large number of factors prompt respondents to answer in a certain way, making social desirability bias a major concern for this study.

In his January 2013 blog post reporting the results, Cooney did attempt to account for bias. He first claimed that the study had no non-response bias, since no one who reported having received a leaflet declined to fill out the survey. This is a dubious assertion, as it neglects the effect of students simply not recalling the leaflet. After all, what is the chance of remembering the cover of a piece of paper received on the street three months ago and promptly thrown away? The raw data for students who received Compassionate Choices shows that fewer than 10% said they viewed the leaflet for “less than 10 seconds,” and based on my experience leafleting, I think this figure is unrealistically low. What’s more, Vegan Outreach conducted a pilot study in Winter 2014 wherein 167 out of 223 readers of Your Choice (which is similar to Something Better) responded to a follow up survey after one month, and only 14% of them correctly identified which leaflet they had read.[16]

Cooney’s method of dealing with social desirability bias is more interesting. He claims that students who reported having received a leaflet before should have the same “level” of social desirability bias as those who didn’t. Since the study found that repeat-leafleting is less effective, the difference between the two groups must not be due to social desirability bias. The excess animals spared are attributed to a real effect, and this is the basis of his “conservative” estimate of 50 animals spared per 100 leaflets.

Selected results from the Fall 2012 FS/THL study. (Formatting retained from the original.)

However, it’s actually not the case at all that repeat-recipients reported less desirable results. Looking at the data for those who received Compassionate Choices either once or more than once, we see that similar numbers in each group reported no change in behavior. Almost all of the difference can be attributed to a shift among repeat-recipients from “I eat less X” to “I already did not eat X.” It’s because of this shift that repeat leafleting is judged to be less effective — after all, 16% of those who receive a second leaflet already do not eat any beef or pork! This is a more desirable response, not less. The difference is certainly not explained by people actually giving up these products, since it is out of proportion with the small number of first-timers who reported that they stopped eating meat.

ACE’s interpretation of this study, while relying on entirely different reasoning, was no less heterodox than Cooney’s. ACE tallied the respondents who said they had eliminated chicken, beef/pork, and fish, and took the minimum of these numbers (which happened to be 5) as “the number of vegetarians that can be “patched together” from these results.” They then fit this figure to a beta distribution and came up with the probability distribution Beta(5.5,484.5), which indicates an estimate of 1.1% (5/489) with a standard deviation of 0.5%. Their cost effectiveness estimate for leafleting is based exclusively on this.

In the many, many times I have heard statistics quoted concerning how many vegetarians are created per leaflet, however, it never occurred to me to ask whether they were patchwork vegetarians. After reading this, I wondered how many of the 489 individual human respondents indicated they became vegetarian?

Just one.[17]

Section 3: ACE’s estimates of the effectiveness of other interventions may lack any empirical basis

Perhaps the most transparent example of ACE’s dubious research is offered by a post on the website of the small research group Faunalytics (formerly Humane Research Council) titled Did Faunalytics Save a Million Animals This Year?[18]

This page contains four requests for donations, including the donation button in the tool bar. The article advertizes ACE’s analysis of Faunalytics, which received the status of a “standout charity,” but takes issue with the estimate for the number of animals which were saved by this organization, whose main contributions are posting links to research papers produced by academia, as well as occasional original essays and a very limited amount of original empirical research. It also does consulting for advocacy groups. According to ACE, these activities saved only 415,800 animals in 2015, whereas Faunalytics suggests that they might have saved 1,039,500.

Plainly, examining the uses of research and consulting is an extraordinarily complex task. Nobody understands the vast number of social, political, and economic forces which go into determining how much meat people eat and other effects of advocacy. Presumably, the largest effects of advocacy research are diffuse and long-term. Since we have no idea what sort of trajectory the future will take, trying to estimate the number of animals saved is plainly futile. Even the notion that there is such a number is strange, given the vast amount of counterfactual reasoning it involves. However, thanks to ACE’s laudable commitment to transparency, we can examine its reasoning. The relevant information is in ACE’s cost effectiveness estimate for Faunalytics under the “program effectiveness” tab.

ACE has the following to say about how it evaluated the impact of Faunalytics’ independent studies.

“To make the estimate of animals spared per organization affected, we use a hypothetical case of an organization that (i) distributes 1,000,000 leaflets per year as their sole activity and (ii) has a 50% likelihood of improving the efficacy of these leaflets by 5%. Given our estimate of animals spared per leaflet at 1.4, this implies an impact of 17,500 more animals spared by the organization. We are extremely uncertain about this estimate and use the cost-effectiveness estimates of these activities as only one small component in our evaluation of Faunalytics.”

There is no suggestion of which organizations are affected or how. ACE does not suggest that the organizations affected even engage in leafleting, nor does it examine any specific way the research might have influenced a change in policy. Instead, the impact of organizations is abstracted in terms of leaflets, and an estimate of the improved efficiency of the unknown organizations’ abstract leafleting is invented.

I find myself at a loss to convey the problems with this estimate without sounding derisive. The suggestion that organizations’ efficiency was improved by 5% is not based on anything whatsoever. It is a made-up figure. Notice further that, while ACE could have simply made up a figure of 2.5% for the hypothetical organizations’ improvement at a hypothetical intervention, it instead made up two figures: first it made up a 5% improvement, and then it made up a 50% probability of such an improvement. This presumably allows ACE to claim it took into account the possibility that Faunalytics’ work had no impact. Finally, note that the entire estimate is based on the results of the leafleting study discussed in Section 2. This study, and ACE’s unusual analysis of it which involved “patching together” vegetarians, are the whole basis of these animals-per-dollar estimates.

Faunalytics’ other interventions are similarly abstracted into leaflet terms, despite having nothing to do with leafleting, and similarly involve invented probabilities of invented percentage improvements in efficiency.

What to make of ACE’s final claim, that it is “extremely uncertain” of its result? Given that the “result” is entirely fabricated, one would hope this. It’s plain that ACE uses such statements to deflect criticism. It often refers to “rough estimates” and “back-of-the-envelope calculations,” and about lacking “robust evidence,” when really it means that it lacks any evidence whatsoever. In an April 2016 blog post, ACE researcher Jacy Reese attempts to address concerns over their cost effectiveness estimates (CEEs), remarking:

“We’ve found CEEs to be the most debated and controversial part of our evaluations, with both our readers and the organizations we review. We think this comes partly from a gut discomfort with assigning numbers to important or hard-to-quantify figures like how much dietary change you expect from receiving a leaflet. If we assign a value of 10% to a figure, some can easily interpret this as our having good reason not to assign 9% or 11%, which implies a level of certainty we rarely achieve.

We’ve heard from some people who interpret our CEEs more literally than we believe they should be, for example, by using our CEE results from top charity reviews to identify one as the most effective. We worry about doing this ourselves, especially given that CEE results are a concise way of explaining the huge impact donors can have in animal advocacy.

To help with this issue, we plan on presenting our cost-effectiveness estimates as ranges rather than point estimates in the future.”[19]

All such deflections carry an implication that the estimates have some basis in reality, which they do not. The problem is not that the estimates are not accurate enough, but that there are no grounds on which to base any estimate at all. Hedging cannot possibly change this fact. A claim that the Earth is 7,000 years old is pseudoscientific, and a claim that it is probably between 6,000 and 12,000 years old is still pseudoscientific. No amount of bayesian analysis can make such a baseless and ideologically-motivated estimate anything other than complete nonsense.

Then there is ACE’s argument that quantitative estimates only form part of its evaluations. But why should anyone trust the other subjective value judgments of people who conduct blatant pseudoscience?

This impugns Faunalytics also. Faunalytics took issue with ACE’s analysis, and suggested that ACE could have arranged for it to have saved a larger number of animals. ACE, for its part, then published Faunalytics’ improved analysis of itself. This supposedly justifies Faunalytics’ post, with four donation requests, musing about how many hundreds of thousands of animals it saved. ACE is aware that organizations are using its estimates of number of animals spared per dollar to deceive donors, and cannot claim it bears no part of the blame for this. Intelligent people should not be deceived by tactics such as saying things while denying saying them, or phrasing lies in question form.

Section 4: Effective Altruist analyses of the direct benefit of cage-free reforms do not accurately reflect the relevant science

Farmed animal welfare groups, particularly The Humane League (THL) and the Humane Society of the United States (HSUS), have obtained an impressive string of successes in recent years by targeting large corporations which buy eggs with campaigns demanding they stop sourcing them from conventional (“battery cage”) facilities, and instead switch to buying eggs from aviary (“cage-free”) facilities. Many of the resulting commitments made by egg buyers have garnered significant media attention. For example, McDonald’s’ commitment was reported on by NPR, TIME, the Wall Street Journal, the New York Times, the Chicago Tribune, and numerous other news organizations.

Many advocates are understandably enthused by this progress. Animal Charity Evaluators is particularly so, estimating that THL’s cage-free corporate campaigns spared 14.5 animals per dollar in 2015.[20] To derive this estimate, ACE guessed that THL was responsible for a 60% share of corporate campaigns which got 5 million hens “out of cages.” No hens were actually freed of course — these are estimates based on promises to transition to cage-free some number of years in the future. However, this is not the main problem with the figure. The biggest problem lies in ACE’s valuation of “equivalent animals spared per animal helped through cage-free corporate commitment,” which it sets at 0.1. This means that ACE would be willing to increase the number of factory-farmed hens by up to 10% in exchange for a switch to cage-free production.

Researchers at ACE have indicated their interest in using welfare science. In a 2016 response to a paper in Animal Sentience, they write:

“Welfare biologists can also help advocates work in the most effective ways, by identifying the incremental changes, such as larger cages or less harmful breeding, that would most benefit animals as we work towards more fundamental improvements.”[21]

It stands to reason, then, that ACE’s estimate of the value of switching to cage-free in terms of hens’ lives should have been informed by welfare science. It was not, in any respect. In fact, it is simply another completely invented figure.

There is good reason for this: the literature does not support the notion that aviary systems provide laying hens with improved welfare. Whether they are worse is not clear. A 2011 review article in Poultry Science expressed what seems to be the consensus when it said “It appears that no single housing system is ideal from a hen welfare perspective.”[22]

Recent debate on this subject in the Effective Animal Activism community stems largely from an informal memo by Direct Action Everywhere, which reviewed recent literature and concluded that cage-free reforms likely harm laying hens.[23] Lewis Bollard, program officer for Farm Animal Welfare at the Open Philanthropy Project, responded to their claims, referencing a 2006 paper which constructs a computer model called FOWEL,[24] which he calls “the most comprehensive review I know of.”[25] This is in spite of the more recent review cited above. This paper also appears to be the only one published in the past decade which makes the claim that aviary systems provide an overall welfare improvement, and the computer model seems not to have gained acceptance.

Mortality of hens in the CSES study. From Karcher, D. et al. (2015). Impact of commercial housing systems and nutrient and energy intake on laying hen performance and egg quality parameters. Poultry Science, peu078.

The most comprehensive study to date was conducted by the Coalition for a Sustainable Egg Supply (CSES), and detailed information is provided at the CSES website.[26] The research results report appendix is particularly valuable, as it shows the relative incidence of various horrors in the three types of system studied — conventional (battery cage), enriched cage (colony), and aviary (cage-free.)[27] The most important and informative figure, however, is the mortality rate. Consistent with the 2011 review article, the study found significantly higher mortality in aviary systems, with 11.7% of birds in the such systems dying before the end of the production cycle.

Many of the excess deaths in the aviary systems were due to cannibalism and vent-pecking, wherein a hen’s cloaca (perhaps the most sensitive part of her body) is pecked out until she dies. Additionally, far more of the hens necropsied in the aviary systems were found to be emaciated. Ammonia levels were also higher, because the birds live in their own feces and kick them into the air. A much larger number of hens in the aviary systems were found to be “dirty.”

The study also found that it takes more hens to produce the same number of eggs in an aviary system, evidently because more of them are dead.

Lewis Bollard disputed whether there is a consensus in the literature that mortality is higher in aviary systems, as the Direct Action Everywhere memo argues. But the first paper he cites as disputing this also found that mortality rates were higher in the aviary systems studied.[28]

With scientists unwilling to take a clear position on whether these systems provide a welfare improvement, it’s strange that Effective Altruists should appear so confident in doing so. The assessments by ACE and OPP of the benefits of the ongoing cage-free transition make less sense as an effort to dispassionately analyze the evidence than as an effort to justify the actions of The Humane League, Mercy for Animals, and The Humane Society of the United States, which are heavily invested in these reform efforts.

Section 5: Charity evaluation may damage organizations’ actual effectiveness

It’s obvious that a charity evaluator could do harm to non-profits simply by providing bad advice, but there is also a more serious danger: as argued in a 2012 paper by Baur and Schmitz, pressure to achieve according to artificial metrics may motivate non-profits to seek immediate results by entering into disadvantageous negotiations with corporate targets.[29] The key word here is co-optation. According to Baur and Schmitz

“Social movement scholars view co-optation as a form of institutionalizing social protest that is engineered by more powerful groups to demobilize the opposition and ensure that their demands are watered down.”

Frequently, co-optation manifests as the manipulation of activists to actually serve corporate ends. Trumpy defines co-optation as “the ability of a corporate target to bring the interests of a challenging group into alignment with its own goals”[30] and Baur and Schmitz also use this definition. The main way in which corporations exploit advocacy groups is by using them to obtain “corporate social responsibility” (CSR), meaning public goodwill arising from the perception that they are accountable and their products are ethical. The corporation succeeds at co-opting the advocates if it obtains CSR through minimally costly (or sometimes even profitable) reforms, which typically fail to address the advocates’ concerns.

In corporate-non-profit negotiations, the businesses typically have the upper hand, which can make it easy for them to obtain CSR from the non-profit at little or no cost to them. Baur and Schmitz argue that watchdog groups such as Charity Navigator and Guidestar may foster co-optation by incentivizing such arrangements. They furthermore note of such watchdog groups that:

“The accountability perspective advocated by these organizations reflects a desire to find “analogues for the commercial ‘bottom line’” (Gray et al. 2006, p. 334), which leads to dysfunctional behavior and distracts from the actual effectiveness of a nonprofit’s activities (Lowell et al. 2005).”

These observations pose major problems from an Effective Altruist perspective, as this philosophy of activism places immense pressure on organizations. It seems that expected utility calculations are more appropriate to charities which obtain results directly, such as disease-control efforts, than to social movement organizations whose actual effectiveness is based on behaving strategically in a complex political and social environment which includes intelligent actors that may oppose or exploit them. From this perspective, attempting to quantitatively measure the effectiveness of corporate outreach becomes not only unwise but absurd. It necessarily entails some sort of counterfactual analysis of how numerous actors would have behaved had the social movement organization acted differently. Anyone attempting such an estimate has to imagine that he is smarter than all the many individuals and corporations who strategically participate in these interactions.

Those who want to qualitatively analyze the effectiveness of a non-profit’s corporate outreach efforts, then, should first and foremost seek to determine whether the non-profit is at risk of being manipulated and exploited. Baur and Schmitz emphasize three factors (among others) that can place a non-profit at risk of co-optation, writing that:

“The most common road to co-optation is corporate sponsoring. Sponsoring is particularly problematic because it can create a resource dependency for NGOs, compromising their ability to challenge corporate behavior.”

and that:

“A second process increasing the risk for NGO co-optation is associated with the rapid expansion of certification and labeling agreements in which NGOs directly or indirectly endorse the products a company sells (Murphy and Bendell 1999).”

and that:

“… there are also increasing personal ties between profit and nonprofit sector that may lead to co-optation (MacDonald 2008). Increasingly, corporate leaders are recruited by NGOs concerned about resource acquisition and management challenges associated with a more competitive fundraising environment.”

What’s more, it’s not only social movement scholars who discuss co-optation. In her book Managing Activism, PR expert Denise Deegan recommends a co-optation strategy based on private sit-down meetings and relationship-building, with specific references to animal rights activists. She indicates that small, radical groups are more effective and harder to work with, and warns that criticism of activists may help publicize them.[31] A natural implication is that corporations should most harshly attack the non-profits that are helping them the most. This strategy will enhance the credibility of the co-opted organizations which they hyperbolically deride as radicals, resulting in more gains in CSR for the token concessions made to them.

Within the context of animal rights, concerns about co-optation are often dismissed as “philosophical” or “ideological.” However, it’s clear that such concerns may be pragmatically founded.

The cage-free corporate campaigns are, in fact, unequivocal examples of co-optation. While members of the Humane Society and The Humane League attribute the recent rash of cage-free commitments to “momentum” and corporate fear of protests, it’s fatuous to believe that large corporations have anything to fear from a handful of animal welfare activists. The rapid-fire successes obtained by The Humane League’s small team of corporate campaigners are not the result of applying pressure; rather, they are mock-protests coupled with closed-door negotiations which enable corporations to acquire CSR at no cost. The corporations don’t have to make any changes at all, since the agreements are generally long term and non-binding, but they do gain positive media coverage immediately for the non-binding commitment to change. The activist group might also get positive coverage, and will couple requests for donations with announcements of “victory.”

Those familiar with the large animal welfare charities will realize that certification and labeling agreements, and personal ties to corporations, are also cause for serious concern. Take, for example, the Global Animal Partnership, a collaboration between the meat industry and several conservative animal welfare groups, whose board consists of three animal welfare representatives, and four animal industry representatives.[32] Through this arrangement, the partners, which include Humane Society of the United States (HSUS) and Whole Foods, make recommendations to meat producers and promote a labelling scheme for humane meat. John Mackey, CEO of Whole Foods, also sits on the board of directors of HSUS.[33]

Whole Foods and HSUS recently produced a documentary film, released in July, 2016. Titled At the Fork, it offers, according to food journalist Mark Bittman, “A beautiful, caring, sensitive and surprisingly non-judgmental look at how most animals are produced in the United States.”[34]

The trailer opens with a feel-good scene of a family barbecue, and explicitly emphasizes the cultural and traditional importance of meat-eating, showcasing an abundance of meat. But the filmmaker decides to give a hearing to his wife, who according to the website, “has been on a plant-based diet for 26 years,” and see for himself how animals are treated in modern agriculture. At the beginning of this adventure, we are introduced to some happy, freely roaming animals. These are then contrasted with confined animals, although there are no images of suffering or death. We are reassured by two separate speakers that Man has dominion over the animals, the second one allowing that “dominion doesn’t mean complete domination.” Temple Grandin makes an appearance, shown heroically walking through a slaughterhouse course of the type she designed. I’m sure HSUS would like to claim that this film at least opposes factory farming, but the trailer includes explicit apologism even for intensive confinement practices: James E. McWilliams assures us that “We have this vision of these fat, evil industrialists, when in fact it was the result of hundreds of thousands of small decisions made by small farmers to produce more with less.”

No rational person can fail to recognize that an ad for clean coal is an ad for coal. Why on Earth, then, is there any question that an ad for humane meat is an ad for meat? Harish Sethu, a board member of The Humane League and current director of Humane League Labs , explains in his blog Counting Animals that a large number of people say things such as “I buy meat only from a small, local, humane farm,” yet the overwhelming majority of meat is factory-farmed.[35] Consumers love the idea of buying meat in a socially responsible way, but they have less inclination to actually modify their consumption pattern. The result is that “humane meat” is a way for people to feel better about regular meat.

When analyzing corporate-non-profit interactions, the correct question to ask is not what the immediate impacts are, but rather, whose purposes are ultimately being served. Could an animal welfare organization with a grocery chain CEO on its board of directors have conflicted interests? Would an animal welfare organization without conflicted interests distribute coupons for bacon? Or organize an event where, each evening, customers eat a different hoofed animal produced by its corporate partners?

Section 6: The current Effective Altruist thinking on social norms and social movements lacks rigor

Most members of the Effective Animal Activism community take an incrementalist approach to reducing meat consumption, for example promoting semi-vegetarianism, and there appears to be a widespread belief that this approach is supported by empirical evidence showing that it effects the most dietary change. As discussed in Section 1, however, no evidence actually supports this idea. Largely, these beliefs are based on the selective interpretation of unrelated scientific research in a way that supports a particular ideological position. In recent years it has been fashionable for incrementalists to cite the ideas of Nick Cooney, in particular his book Change of Heart, which synthesizes a wide array of psychology research into a speculative theory of social change. In itself, this is a commendable enterprise — activism can and should be informed by insights from psychology and sociology. Problematically, however, many activists have taken Cooney’s untested speculations as gospel.

An illustrative and important example of this is Cooney’s interpretation of the “foot-in-the-door” approach, a well-studied phenomenon wherein a person who accepts a small request becomes more likely to later accept a larger request of a similar nature. Incrementalists frequently cite this phenomenon as the justification for soft messaging such as Meatless Monday requests, arguing that adopting such a small lifestyle change will set people on a path towards veganism. Skeptics may retort that Cooney also cites another well-studied phenomenon, “door-in-the-face,” wherein a person who refuses a large request becomes more likely to accept a subsequent smaller one. The only meta-analysis comparing the two approaches found no difference in effectiveness,[36] so there doesn’t seem to be any reason to favor either the veganism-only or the meat-reduction approach on the basis of these phenomena alone. What’s more, neither has been studied in the context of advocacy for animals.[37]

Social norms

For the most part, society-level behavioral change is not achieved by reaching out to individuals one-by-one, but by shifting social norms. A good introduction to this subject is provided by a July 2015 article by Gerry Mackie et al. titled What are Social Norms? How are They Measured?[38] This fairly comprehensive review by a top researcher in the field was prepared for UNICEF with the needs of activists in mind, and is specifically intended to inform efforts to change socially accepted harmful behaviors. An important takeaway from this article is that there are different types of social norms — they may be descriptive, that is, based on mimicry, or injunctive, that is, based on fear of judgment by others. Furthermore, different types of social norms are appropriate for changing different types of behaviors. A changing dependent behavior, which is based on observing others without regard for their expectations, may be modeled by the “diffusion of innovations” theory, wherein people adopt a behavior because they witness others adopting it and enjoying success. An example relevant to vegan advocacy might be a health-conscious person witnessing an acquaintance lose weight after going vegan, and thus being inspired to try it himself. On the other hand, interdependent behaviors, which involve mutual cooperation among groups of people, do not change this way, and require attitudinal shifts concerning what others think one ought to do.

Whether a behavior is independent or interdependent also determines the rate at which it will change in response to shifting attitudes. The below graphs, taken from Mackie et al., show the typical progression of a successful effort to introduce a new behavior. The left one shows a mostly independent behavior, the improvement of child nutrition. People generally want their children to be healthy, so they will usually adopt the improved nutrition regime if they believe it is actually superior. However, some people will never be convinced of this and will not change. On the right we see an interdependent behavior, toilet usage. In many parts of India, population density has increased to the point that “open defecation” is a major public health hazard. However the practice continues, strongly supported by tradition. Because of the free-rider problem, people are not inclined to switch to uncomfortable and unfamiliar Western-style toilets unless almost everyone in the community does so. Yet, if a large majority of people are convinced of this solution to the problem, it may be adopted very rapidly, including by those who resent the change.

From Mackie, G., Moneti, F., Denny, E., & Shakya, H. (2015, July 27). What are social norms? How are they measured?. University of California at San Diego-UNICEF Working Paper, San Diego.

It’s important to notice that attitudinal change precedes behavioral change — by a little in the case of an independent behavior, and by a huge margin in the case of an interdependent behavior. While foot-in-the-door is certainly a real phenomenon, and on the individual level behavioral change can indeed precede attitudinal change, this is definitely not the typical case. On the societal level, massive shifts in attitude may be necessary before any change in an interdependent behavior is noticed at all, and then the behavior may change all at once.

This perspective of course leads to the question: is animal use more independent, or more interdependent? I think it is highly interdependent. First of all, farmed animals are killed industrially, rarely by individual consumers, leading to a free-rider problem. Many meat-eaters undoubtedly do not feel responsible for the fact that slaughterhouses continue to operate. Secondly, people have a tendency to eat in groups, and will eat what others are eating. Meat-eating is strongly interdependent in this respect, as abstinence from it creates inconvenience for both oneself and one’s friends and family. These factors suggest that most people may have to be convinced to oppose the commercial use of animals before many stop eating animal products, and that a myopic focus on short-term behavioral change is a bad strategy.

Nick Cooney discusses social norms extensively, but he only ever refers to descriptive norms. This is a conspicuous omission, since the distinction is prominently emphasized in works of Robert Cialdini that Cooney relies upon heavily. Moreover, as Mackie et al. explain, many researchers have focused exclusively on injunctive norms (the ones with normative content!) and this is often what is meant by the term “social norm.” Acknowledging the importance of injunctive norms requires us to take attitudinal change seriously. Efforts to measure attitudes are more likely to succeed than attempts to measure behavioral change, and understanding attitudes may give us more relevant information about the success or failure of interventions.

Social movements

Animal Charity Evaluators’ effort to include relevant social science in its evaluations is represented by its “Social Movements Project.”[39] This wildly expansive project will involve collecting “case studies” of various historical social movements, and then meta-analyzing them in order to glean relevant insights. Why would this be necessary? It is certainly not that social movements haven’t been the subject of a great deal of academic inquiry. But as ACE researcher Allison Smith explains:

“We’ll be using the academic literature on social movements where we can, although our goals and viewpoint are different from those of most academics. We’re focused on applying what we learn to the future of animal activism, not on developing theories that can explain all social change.”

However, a more likely explanation for ACE’s reluctance to use the literature on social movements is that it effectively recommends against their preferred methods of advocacy. Social movement theory emphasizes the importance of ideology and conflict. ACE is certainly aware that critics of soft messaging have accurately pointed this out. In particular, Wayne Hsiung references the mainstream theory dating back to the work of Coleman in the 1970’s, which holds that social movements succeed by establishing new (injunctive) social norms, supported by a small number of “zealots.”[40] But because injunctive norms pose a problem for incrementalists’ ideology, ACE, like Cooney, would prefer to ignore their existence.

Even if ACE decided that the existing sociology literature was not tailored to its needs, and a special analysis was necessary, it is bizarre that it should decide to create an entirely new body of social movement history to begin from. It is not as though there aren’t already respected books on the comparative study of social movements — see, for example Tilly’s Social Movements, 1768–2004,[41] or McAdam, McCarthy, and Zald’s Comparative Perspectives on Social Movements.[42] One may wonder why ACE does not use these as the starting point for its analysis.

Let’s say we accept that ACE wants to produce objective case studies in a uniform way, so it can evaluate them without bias. Each case study will be an enormous project — many books will have been written about each movement studied, and scholars will have dedicated their careers to it. Surely, then, ACE is reaching out to historians and sociologists with relevant expertise, and asking them to prepare these reports? Not exactly. ACE’s blog announces that

“Currently, this project is being conducted mainly by our research interns. If you’d like to get involved, and you have time to do an internship or to volunteer consistently, let us know.”

To deliberately ignore an entire body of literature on an academic subject, only to attempt to re-create it from scratch, demonstrates a pseudoscientific attitude.

Section 7: The current Effective Altruist thinking on wild animal issues lacks rigor

In part because the number of animals in the wild vastly exceeds the number of domestic animals, issues affecting wild animals are of no small importance. ACE has recognized this. However, while most advocates for wild animals assume that one of the best things that can be done for them is to protect their habitats and counteract the human-driven destructive forces responsible for the ongoing Sixth Extinction, some have taken a more nuanced approach.

ACE’s main source of inspiration on this subject appears to be Brian Tomasik, who was a board member of ACE until late 2015. Four of the six resources on ACE’s “wild animal suffering” page are unpublished papers by Tomasik.[43] Tomasik advocates the position that nature is harmful to animals, and that it would be better to destroy wild animal habitats as much as possible. He explains this in a paper cited by ACE on this page:

“I personally believe that most animals (except maybe those that live a long time, like >3 years) probably have lives not worth living, because I would trade away several years of life to avoid the pain of the average death in the wild, and this is assuming that even their lives are net positive (which is dubious in view of cold, hunger, disease, fear of predators, and all the rest).”[44]

While ACE has approached this subject cautiously, it’s plain that they are sympathetic to Tomasik’s view. A post by ACE researcher Allison Smith examined concerns that the animal movement might promote environmentalism, linking the word “concerns” to an essay by Tomasik[45] which is more explicit.[46] ACE also conducted a study of the effect of various messages on support for policies that would reduce wild animal suffering (WAS), in which respondents got a better “WAS score” if they supported habitat destruction.[47]

Currently, Tomasik is lead researcher at the Foundational Research Institute, one of three “projects” of the Effective Altruism Foundation, the others being Sentience Politics, which forms policy positions based on suffering reduction, and Raising for Effective Giving, which fundraises mainly from online poker players. Simon Knuttson, former chair of the board of ACE, is also currently a researcher at the Foundational Research Institute.

The Foundational Research Institute’s list of open questions reveals a highly ambitious project.[48] Their interests include, but are not limited to, strong AI, utilitarian ethics, epistemology, international relations, moral psychology, and “choose-your-own-topic.” Certainly the most interesting one is “suffering in physics.” The questions they pose in this subject area are copied below.

“ To what extent should we see consciousness-like operations in physics? How much do we care about them?

Are there positive-sum ways to reduce suffering in physics without impinging too much on other values?

How sensitive are our consciousness assessments to fundamental questions in metaphysics, string theory, etc.? For instance, if our ontology changes, do our consciousness assessments also change radically?

If, as it seems, fundamental physics consists of very abstract mathematical structures, can we assess their sentience by analogy with more familiar macroscopic brain processing? Or do we need other ways of directly attributing consciousness to symmetry groups, high-dimensional manifolds, Hilbert space, etc.? This partly depends on our ontological perspective on such mathematical structures — whether they are the universe or just abstractly describe the universe.”

There are too many problems with this to describe. (One of them is that “metaphysics, string theory, etc.” is not a sequence of words that should ever occur in the English language.) I will leave it to the reader to judge the merits of this research program. Likewise the one listed directly below it, which inquires into the suffering which may be experienced or caused by extraterrestrial aliens.

Evidently, Tomasik is not solely interested in the suffering of wild animals. He explains in a blog post for ACE that we should also be concerned with the suffering of future “virtual animals.”[49] Granted, some intelligent people take transhumanism seriously, and the idea of a future civilization running computer programs which simulate subjectively aware beings poses potential ethical problems. However, it’s not clear why this is relevant to present-day animal rights or welfare efforts, or why we should call these hypothetical future beings “virtual animals.”

But Tomasik goes further still. He is not, as it turns out, merely interested in future virtual beings. He is also interested in ones which presently exist. In an essay titled Do Video-Game Characters Matter Morally? Tomasik expresses concern for “non-player characters (NPCs)” such as appear in computer games like Doom 3 and Super Mario RPG, especially because “some NPCs have explicit representations of their “welfare level” in the form of hit points (HP), and the NPCs implement at least crude rule-based actions aiming to preserve their HP.”[50]

Perhaps all this theory is too abstract for most people. In an attempt to more concretely address the issue of alleviating the suffering of actual, living animals, the Foundational Research Institute conducted an analysis of a hypothetical intervention which could prevent elderly elephants from dying of starvation after losing their teeth, a rare cause of death. In a facebook post, Tomasik explained that “the main goal behind this paper is to explore the idea of helping elephants as kind of a symbolic policy proposal in order to oppose blanket sentiments that “there’s absolutely nothing we can do to help wild animals”, etc.”[51] But the researcher, Ozy Brennan, lamentably concluded that:

“Unfortunately, euthanasia of elephants is unlikely to be a useful intervention, because most elephants do not live long enough to die of molar loss.”[52]

The article found some cause for optimism, however:

“One might propose a broader euthanizing of elephants: for instance, euthanasia during droughts or of sick elephants.”

Naïve people might think that a way to help elephants could be to stop the poaching crisis, but for suffering reducers, maximizing utility is largely a matter of minimizing life. The main intervention that interests Tomasik (although not Brennan) is the destruction of habitats.[53]

Not all Effective Altruists who have written about wild animals advocate their annihilation. Oscar Horta, for example, more modestly proposes spreading awareness of the issue and researching new methods of intervening in nature for wild animals’ benefit, citing vaccination, feeding starving animals, and rescue of orphaned or injured animals as examples of current interventions.[54] Just as often, however, their thinking appears wildly out of touch. For example, one of the seven survey questions in ACE’s study on attitudes about wild animal suffering asked “Assuming it could be done cheaply, would you support or oppose terraforming Mars?”

The Effective Altruist work on wild animal suffering and related subject areas is concerning in several ways. First, it reveals an extreme lack of awareness, on the part of many Effective Altruist researchers, of the limitations of their knowledge and ability. Second, it threatens to reinforce the status of animal rights or protection as a fringe movement. Third, should anti-speciesist opposition to nature become more widespread, it could create an irreconcilable rift between some animal protectionists and all environmentalists, further marginalizing efforts to advocate for animals.

One also must wonder why those who support habitat destruction would promote veganism or meat reduction at all? For these suffering reducers, a large reduction in demand for beef would be catastrophic, as it would result in the preservation of vast tracts of nature. In fact, the clear course of action for them is to actively promote beef, which is certainly much easier than making the public care as much about wild animals as they do.

No mainstream animal protection non-profit has openly endorsed the position that destroying ecosystems will benefit wild animals by saving them from life, but this position certainly dovetails with the one espoused by Matt Ball of Vegan Outreach, who cautions against using the environmental argument for veganism because it might cause people to eat less beef. Ball is concerned that reduction in demand for beef will result in increased demand for chicken, and thus a net increase in suffering.[55]

Section 8: Effective Animal Activism is fraught with severe conflicts of interest

For a movement so concerned with objective and impartial reasoning, the Effective Animal Activist community is shockingly tolerant of sources of bias.

Conflicted interests in ACE evaluations and OPP grants

Sociologist Corey Wrenn has written of Animal Charity Evaluators’ conflicted interests in a peer-reviewed book. She notes that Nick Cooney has contributed to ACE’s social networking, and John Bockman, executive director of ACE, formerly ran a group which distributed Vegan Outreach literature.[56] Animal Charity Evaluators has recommended The Humane League, which Cooney founded, in all five years since its inception, and Mercy for Animals, where Cooney is employed, in all four years that Cooney has been working there. It has also heavily promoted leafleting, and distorted research on its effectiveness. Of the 54 “conversations” which ACE has published on its website, 17 have been with people affiliated with The Humane League, including 4 with founder Nick Cooney and 5 with president David Coman-Hidy.[57] In 2015, The Humane League also received a special “deep review,” which no other organization received.[58] In his book How to Be Great at Doing Good, Cooney goes noticeably out of his way to mention that:

“It’s also worth noting that I am not affiliated with Animal Charity Evaluators, the independent organization that reviewed and ranked these and other animal protection charities. For a thorough review of ACE’s methodology and data, feel free to visit the Animal Charity Evaluators website.”[59]

The Open Philanthropy Project’s work on farm animal welfare is led by Lewis Bollard. According to the website,

“Prior to joining Open Philanthropy, he worked as Policy Advisor & International Liaison to the CEO at The Humane Society of the United States (HSUS). Previously, Lewis worked as a litigation fellow at HSUS and an associate consultant at Bain & Company.”

The Open Philanthropy Project has made $1.5 million worth of grants to the Humane Society of the United States and the Humane Society International, as well as a total of $4 million to The Humane League and Mercy for Animals, restricted to cage-free campaigns, an area in which these organizations collaborate with the Humane Society. Most troublingly, it made a $500,000 grant to the Global Animal Partnership, the industry group discussed in Section 5.[60]

Conflicted interests in research

Much of the research that Animal Charity Evaluators relies upon is produced by organizations on their own methods, and then used to evaluate those same organizations. This is especially true of The Humane League, which has conducted several studies that ACE relies on for its evaluations, including the leafleting study discussed in Section 2.

The Humane League has received criticism in the past for its unscientific and biased research, leading it to attempt some damage control in June 2016. Executive Director David Coman-Hidy wrote in a June 15 blog post: “We are very excited to announce the next phase of Humane League Labs. New and dedicated leadership. New commitments. New research priorities.”[61]

The new director of Humane League Labs is Harish Sethu, who has been centrally involved in designing the studies since its inception. A report by Faunalytics announced the foundation of HLL as follows:

“Calling on the expertise of Harish Sethu from Counting Animals and others, Humane League Labs will engage in research that measures the effectiveness of various veg advocacy approaches and will make the results available to the movement.”[62]

And an October 2014 conversation with THL published by ACE tells us that:

“Usually Nick [Cooney] and Harish Sethu play a role in setting up survey questions and they also get outside feedback and, if possible, involve the team that does the analysis.”[63]

Humane League Labs’s “commitment to ethics” indicates it will “manage conflicts of interest.” But coming from a non-profit organization conducting research aiming to prove the effectiveness of its own methods, this statement is nothing short of bizarre. And while its “new commitments” suggest that its earlier work was not actionable and should not have relied on self-reported data,[64] these statements are no substitute for formal retraction.

The inadequacy of the community’s oversight is made brilliantly clear by a recent post by Harish Sethu on the Animal Charity Evaluators blog, wherein Sethu argues that it is indeed acceptable for non-profits to conduct research on themselves.[65] The very existence of the post indicates that ACE is not sufficiently concerned about appearing too closely connected to the organizations it reviews, and illustrates why it’s necessary for the community to demand nothing less than independent, peer-reviewed research.

The Good Food Institute

If there is any doubt about whether or not ACE’s charity evaluation process truly involves objective assessment of organizations’ performance, it should be put to rest by the recent recommendation of the Good Food Institute (GFI) as a “top charity.”[66] ACE gave its highest accolade to an organization in its first year of existence, passing over at least two organizations which work in a similar domain, New Harvest and the Modern Agriculture Foundation, that have an actual track record of accomplishments including companies founded and scientists funded.

ACE’s review of GFI strains severely to maintain the pretense that the organization performs well on its formal criteria. On Criterion #5, “The Charity Possesses a Strong Track Record of Success,” ACE mostly lists things that GFI plans to do, additionally mentioning trivialities such as “They have developed an email list of more than 50 potential entrepreneurs and scientists,” as well as publicly accessible information that GFI learned:

“Through their research and communication with scientists around the world, GFI has learned about promising technologies that they believe were previously undiscovered or underutilized in the U.S. For example, some European companies are successfully using hemp seeds, oats, lupins, and fava beans in plant-based meats. They also learned of a machine called the “couette cell” that produces plant-based meat using less energy than standard U.S. machinery.”

On Criterion #4, “The Charity Possesses a Robust and Agile Understanding of Success and Failure,” ACE stoops as low as to discuss how GFI used an op-ed to get more signatures for an online petition (which then still failed to persuade its target.) It additionally lauds “their recent decision to work with a recruiter to hire new talent.” It does not even attempt to assess Criterion #2, cost effectiveness, evidently because GFI has not actually done anything.

Why the top rating? The only plausible explanation is provided by a footnote in the review: “GFI was the brainchild of the leadership team at Mercy For Animals (MFA).”

It’s not clear what GFI’s goals are, or how it will operate. Will it focus on development of new technologies, or will it act more like a lobby group for meat alternatives? If the latter, will it favor some companies over others? Will it further entrench potentially damaging corporate-non-profit relationships?

We should ask similar questions about the Open Philanthropy Project’s $1 million grant to GFI. For all its claims of transparency and fairness, Effective Animal Activism suffers from a deficiency of accountability.


As a movement which literally measures good in dollars, and which is willing to base its decisions on complex and unintuitive reasoning, Effective Altruism is particularly vulnerable to fraud. Effective Altruists should be keenly aware of conflicts of interest, and, especially, should not accept research conducted by non-profits on themselves or those closely connected to them. They should also beware of promoting the co-optation of activist groups through the over-emphasis of unrealistic measures of effectiveness.

I have accused organizations including Animal Charity Evaluators and The Humane League of conducting pseudoscience. I mean this in a sense described in The Stanford Encyclopedia of Philosophy: pseudoscience must not only be unscientific, but also be part of a non-scientific doctrine whose major proponents try to create the impression that it is scientific, or that it represents the most reliable knowledge on its subject matter.[67] The research conducted by these organizations is not merely unreliable, but systematically deceptive. The remedy is to insist on peer review, and be skeptical of self-appointed experts.

Quantitative research can inform advocacy efforts, but it cannot determine them. Activists and donors interested in helping animals should consider the opinions of people outside the Effective Altruist bubble, who have not made false claims of being “evidence based,” “data driven,” or “scientific,” but who might have good ideas anyway.


  1. ^ FAQ | Animal Charity Evaluators.
  2. ^ Cruelty on Screen | Mercy for Animals.
  3. ^ JFA Humane Education Study | Animal Charity Evaluators. (
  4. ^ Humane Education | Animal Charity Evaluators. (
  5. ^ Leafleting | Animal Charity Evaluators. (
  6. ^ 2013 Leafleting Study | Animal Charity Evaluators.
  7. ^ Cooney, N. (2014, May 20). Report: What Elements Make a Vegetarian Leaflet More Effective?. Humane League Labs.
  8. ^ Cooney, N. (2015, September 20). Report: Which request creates the most diet change?. Humane League Labs.
  9. ^ Norris, J. and Roberts, E. (2016, July). Booklet Comparison Study — Pay Per Read #2. Vegan Outreach.
  10. ^ The “V” Word: A Note About Terminology. (n.d.). Mercy For Animals. (
  11. ^ Cooney, N. (2013, January 15). The Powerful Impact of College Leafleting (Part 1) | Farm Sanctuary. (
  12. ^ Cooney, N. (2013, January 15). The Powerful Impact of College Leafleting: Additional Findings and Details (Part 2) | Farm Sanctuary. (
  13. ^ Hsiung, W. (2013, September 18). Science or Science-y: Part I. The Liberationist.
  14. ^ Leaflets are Ineffective, Tell Your Friends. (2015, April 24). Aceso Under Glass. (
  15. ^ Cooney, N. (2011). Change of Heart: What Psychology Can Teach Us About Spreading Social Change (p. 84). New York: Lantern Books.
  16. ^ Norris, J. (2014, December). Booklet Comparison Study — Pay Per Read #1. Vegan Outreach.
  17. ^ Cooney, N. (2013, July 19). Report: Which Leaflet Is More Effective?. Humane League Labs. Data available at and
  18. ^ Did Faunalytics Save a Million Animals This Year?. (2015, December 23). Faunalytics. (
  19. ^ Reese, J. (2016, April 21). Some Thoughts on Our Cost-effectiveness Estimates. ACE blog. (
  20. ^ THL Cost Effectiveness Estimate | Animal Charity Evaluators.
  21. ^ Smith, A. M., & Reese, J. (2016). An empirical perspective on animal advocacy. Animal Sentience: An Interdisciplinary Journal on Animal Feeling, 1(7), 12.
  22. ^ Lay, D. C. et al. (2011). Hen welfare in different housing systems. Poultry Science, 90(1), 278–294.
  23. ^ Memo and Review of Literature — Aviary Systems and Corporate Campaigns. (2016, March 14). Direct Action Everywhere.
  24. ^ De Mol, R. M. et al. (2006). A computer model for welfare assessment of poultry production systems for laying hens. NJAS — Wageningen Journal of Life Sciences. 54(2), 157–168
  25. ^ Bollard, L. (2016, March 31). Initial Grants to Support Corporate Cage-free Reforms. Open Philanthropy Project.
  26. ^ Final Results | Coalition For Sustainable Egg Supply.
  27. ^ Research Results Report Appendix | Coalition For Sustainable Egg Supply.
  28. ^ Ahammed, M. et al. (2014). Comparison of Aviary, Barn and Conventional Cage Raising of Chickens on Laying Performance and Egg Quality. Asian Australas. J. Anim. Sci. 27(8), 1196–1203
  29. ^ Baur, D., & Schmitz, H. P. (2012). Corporations and NGOs: When accountability leads to co-optation. Journal of Business Ethics, 106(1), 9–21.
  30. ^ Trumpy, A. J. (2008). Subject to negotiation: The mechanisms behind co-optation and corporate reform. Social Problems, 55(4), 480–500.
  31. ^ Deegan, D. (2001). Managing activism: A guide to dealing with activists and pressure groups (p. 65–66). London: Kogan Page.
  32. ^ Meet The Team | Global Animal Partnership.
  33. ^ Board of Directors and National Council | The Humane Society of the United States.
  34. ^ At The Fork.
  35. ^ Sethu, H. (2014, June 23). Do you know someone who buys meat only from a small local farm?. Counting Animals. (
  36. ^ Pascual, A., & Guéguen, N. (2005). Foot-in-the-door and door-in-the-face: A comparative meta-analytic study. Psychological Reports, 96(1), 122–128.
  37. ^ Taft, C. T. (2016). Motivational methods for vegan advocacy: A clinical psychology perspective. Vegan Publishers.
  38. ^ Mackie, G., Moneti, F., Denny, E., & Shakya, H. (2015, July 27). What are social norms? How are they measured?. University of California at San Diego-UNICEF Working Paper, San Diego.
  39. ^ Social Movements Research Project | Animal Charity Evaluators. (
  40. ^ Hsiung, W. (2015, December 2). Are “Nice” Vegans Making Us Look Bad?. The Liberationist.
  41. ^ Tilly, C. (2004). Social Movements, 1768–2004. Routledge.
  42. ^ McAdam, D., McCarthy, J. D., & Zald, M. N. (1996). Comparative perspectives on social movements: Political opportunities, mobilizing structures, and cultural framings. Cambridge University Press.
  43. ^ Wild Animal Suffering | Animal Charity Evaluators. (
  44. ^ Tomasik, B. (2016, April 22). The Importance of Wild-Animal Suffering.Foundational Research Institute.
  45. ^ Tomasik, B. (2016, May 22). Does the Animal-Rights Movement Encourage Wilderness Preservation?. Essays on Reducing Suffering. (
  46. ^ Additional Analysis of a Study of Vegetarians | Animal Charity Evaluators. (
  47. ^ Hecht, L. (2016, September 5). Wild Animal Suffering Survey Report. Animal Charity Evaluators. (
  48. ^ Open Research Questions | Foundational Research Institute. (
  49. ^ Tomasik, B. (2015, 3 February). Why Digital Sentience is Relevant to Animal Activists. ACE blog. (
  50. ^ Tomasik, B. (2015, October 11). Do Video-Game Characters Matter Morally?. Essays on Reducing Suffering. (
  51. ^ Foundational Research Institute | Facebook.{%22tn%22%3A%22R0%22}.
  52. ^ Brennan, O. (2016, April 29). Euthanizing Elderly Elephants: An Impact Analysis. Foundational Research Institute.
  53. ^ Tomasik, B. (2016, February 7). Habitat Destruction, Not Preservation, Generally Reduces Wild-Animal Suffering. Essays on Reducing Suffering. (
  54. ^ Horta, O. (2015, January 5). Why the Situation of Animals in the Wild Should Concern Us. ACE blog. (
  55. ^ Ball, M. (2016, May 3). Animals as the Bottom Line. A Meaningful Life, A Better World.
  56. ^ Wrenn, C. L. (2015). A rational approach to animal rights: Extensions in abolitionist theory. Springer.
  57. ^ Conversations | Animal Charity Evaluators. (
  58. ^ The Humane League Review | Animal Charity Evaluators.
  59. ^ Cooney, N. (2015). How to be great at doing good: Why results are what count and how smart charity can change the world (p. 77). Jossey-Bass.
  60. ^ Grants Database | Open Philanthropy Project. (
  61. ^ Coman-Hidy, D. (2016, June 15). The Exciting Next Phase of Humane League Labs. The Humane League — Blog. (
  62. ^ Asher, K. (2014, February 21). Animal Rights National Conference 2013: Trading Speculation for Science. Faunalytics. (
  63. ^ October 2014 Conversation with The Humane League | Animal Charity Evaluators. (
  64. ^ Our Statement of Commitment | Humane League Labs. (
  65. ^ Sethu, H. (2016, November 21). Should Animal Organizations Conduct Their Own Research?. ACE blog. (
  66. ^ Review of the Good Food Institute | Animal Charity Evaluators. (
  67. ^ Hansson, S. O., Science and Pseudo-Science, Stanford Encyclopedia of Philosophy.