6 Distance, awareness/consideration
Connection to terms/concepts in other work
- Caviola, Schubert, and Nemirow (2020) (literature review): Motivational obstacles, especially…
- ‘personal connection’,
- ‘narrow moral circle’
6.1 Description and relevance to Effective Giving
If we focus on ‘enhancing current human well-being and avoiding suffering’, the evidence strongly suggests that the most effective charities operate in developing countries. All but one of GiveWell’s top charities support people in sub-Saharan Africa (“Top Charities” n.d.).2 Moreover, people in wealthy developed countries are more likely to be in a position to give a significant amount (Macaskill 2015). However, the large cultural and physical distance between those in developed countries and those in developing countries may be an important barrier to giving and a reason why people prefer to give locally.
Similar issues of ‘distance’ also applies to other causes with strong arguments for being effective and impactful (see, e.g., ‘EA funds’ recommendations), particularly supporting (farmed and wild) animal welfare and the preserving the long-term future of humanity. Most potential donors do not see inside of factory farms3 and obviously no one can visit the far-future.
6.2 Theoretical and conceptual underpinning
This section may be considered an ‘aside’ for now, but we do want some discussion of the definition and underpinnings for each barrier.
Construal theory, psychological distance, moral distance/moral circles, parochialism and its evolutionary underpinnings
“Moral expansiveness scale” 4
The importance of sympathy, which may be driven by distance and awareness…6
Jaeger and Vugt (2021) make an explicit connection to evolutionary factors that may be driving parochialism:
For much of their evolutionary history, humans lived in relatively small, close-knit groups. This has led to the emergence of prosocial emotions and intentions towards kin and other members of an individual’s ingroup, but not necessarily towards individuals who fall outside of the group boundaries. (Citing Aktipis et al, 2018; Greene, 2013)
6.3 Overall evidence for…
Some FMRI evidence…
Ana Diez suggests…
fMRI studies … have shown that the MPFC [medial pre-frontal cortex] is active when we think about ourselves. If we think of a family member, it will activate less. And if we are thinking of an unknown person, it will be even less active (Qin & Northoff, 2011). And more interestingly, when we think about our future selves, there is less activation of the MPFC than when we think of our actual self (Hershfield, Hal. 2011), meaning we care little about that person in the future.
6.4 Distance - Spatial/Physical, Social/Cultural: parochial altruism/in-group bias, interpersonal and identity e.g., race, gender, age, etc
Consider:
What is the evidence that we are less empathetic or less generous to those far from us along these margins?
What is the evidence that this is manifested in our actions and choices (political, professional, etc)?
What is the evidence that this is manifested in our charitable giving?
Parochialism may pose a significant barrier to effective giving. Parochialism, as defined by Baron and Szymanska (2011) is the mental process of making a distinction between in-groups and out-groups, where people feel a greater connection with their own groups and weigh their welfare more heavily.
There is some evidence for this in charitable contexts. In a laboratory experiment, Chen and Li (2009) found that when matched with an in-group member, people were significantly more likely to show charitable concern, shown through their allocation of tokens. Sudhir, Roy, and Cherian (2016) suggest that categorizing others as in-group members (along lines such as race, gender and age) increases the sense of responsibility and level of emotion felt, resulting in higher levels of sympathy and consequently a stronger urge to help. They examined this effect by randomizing advertising content in a large scale experiment with charitable mailings in India. When the woman on a flyer advertising the charity in question was from the in-group (in this case a Hindu woman in India, compared to a Christian woman in the out-group), they found a higher rate of donation and greater amounts raised. They suggest that the charity “more than doubles donation dollars by recognizing the identified victim sympathy bias in making its appeals”.
Other potentially relevant work:
Kogut et al 2018:
Abstract: In the present research, we examined how the prospective donorâ’s psychological distance from a given victim may interact with the victim’s identification to determine the donor’s willingness to accede to requests for donations to help the victim in question. In three studies, we measured willingness to donate (Studies 1 and 2) and actual donations (Study 3) to identified or unidentified victims, while measuring (Study 1) or manipulating (Studies 2 and 3) the psychological distance between prospective donors and the recipients. Results indicate that increasing the psychological distance between prospective donors and victims decreases willingness to help but only when the victims are unidentified, not when they are identified. This suggests that victim’s identification mitigates the effect of distance on donor’s willingness to help.
K.F. Law, D. Campbell, B. Gaesser: Biased benevolence: The perceived morality of effective altruism across social distance
Across four studies, the authors demonstrate that donating money to socially distant others (where donations usually have the largest impact) is perceived as less morally acceptable. Actors who helped a distant stranger rather than socially closer others (e.g., family or friends) were also judged more negatively and trusted less in an economic game.
6.5 Distance: Experiential, Informational, Emotional/Affective7
Experiential distance describes one’s proximity to a particular situation or feeling as a result of having seen something or been a part of it. In particular, a greater experiential distance is seen to make it more difficult to imagine a particular situation or feeling, therefore making it harder to empathize with someone going through it. This could create a barrier to giving effectively as most charitable giving is motivated by empathy and sympathy for victims (Lowenstein & Small, 2007). Experiential distance explains why it is easier for an individual to feel empathy for a victim if they have personally experienced the ailment or someone close to them has. As a result, they do not have to imagine the suffering it may have caused because they have directly or indirectly physically experienced it. In other words, the experiential gap is smaller. For example, people living in wealthy nations are more likely to be affected by cancer than malaria, leading to a greater support for that cause.
This could reflect a form of availability bias, specifically availability-by-recall. This is defined by Pachur et al., (2012): “Recall instances of Risk A and Risk B, respectively, from your social network (encompassing family, friends, and acquaintances), and infer that risk to be more prevalent in the population for which more such instances can be recalled.” An example of this is presented in the study by Lichenstein et al., (1978). They found that participants in their study were more likely to believe that an illness was more deadly if someone they knew (indirect suffering) or they themselves (direct suffering) had experienced the illness.
6.6 Evidence on determinants of scope of moral circle/moral expansiveness and its impact on giving and other-regarding behavior
See: C. R. Crimston et al. (2018) and references cited within
From abstract of this paper: “Moral expansiveness was related (but not reducible) to existing moral constructs (moral foundations, moral identity,”moral” universalism values), predictors of moral standing (moral patiency and warmth), and other constructs associated with concern for others (empathy, identification with humanity, connectedness to nature, and social responsibility).”
6.7 Availability heuristic and media (also see ‘biases’)
Rare events such as natural disasters are particularly vivid and salient, and the availability bias suggests that people will overestimate the probability and value of events if they come to mind more easily (Amos Tversky and Kahneman 1973). For example, people might overestimate the probability of airplane crashes due to their salience.
Loewenstein and Small (2007) argue that vividness is also one of the key determinants of sympathy, and one of the most important demonstrations of this is the identifiable victim effect. Their research suggests that victims who are ‘determined’ received higher levels of aid compared with indeterminate victims, as without identifying them people struggle to empathise, and are as a result less likely to help (Small and Loewenstein 2003).8
Epstein (2006) concluded that natural disasters received a level of donations that was significantly out of proportion to the damage the disasters caused, while other long-term, persistent problems such as AIDS and malaria got much less attention, and consequently much less funding. For example, he discusses the difference in donation rates at the time (2006) between victims of Hurricane Katrina and victims of AIDS, with private donors giving 1,839 dollars per person to those affected by Katrina, and only 10 dollars per person diagnosed with AIDS. He also argued that the huge donations to causes like Hurricane Katrina had the effect of significantly reducing donations to some other charities, although this idea of ‘expenditure substitution’ is hard to measure and is still largely unresolved. See previous section (’Are characteristic in competition)
Even when it comes to relief from the government, Eisensee and Stromberg (2007) suggest that the level of U.S. relief for natural disasters depended on whether the disaster had happened while other widely covered events were also occurring.
Using data on natural disasters from the Emergency Disaster Database (EM-DAT), and disaster responses from the Office of Foreign Disaster Assistance (OFDA), they explored the relationship between news coverage and relief for natural disasters. To examine whether news had a causal effect on relief, they used two instrumental variables. These variables served the purpose of constructing “meaningful and operational measures of the availability of newsworthy material”. First, they used ‘news pressure’, by measuring the median number of minutes news broadcasts devoted to the top three news stories of the day. Secondly, they used the Olympics, chosen as a large media event that isn’t directly related to politics.
They concluded that U.S. disaster relief decisions were dependent on “the availability of other newsworthy material at the time of the disaster”. Their explanation of this effect is that “relief decisions are driven by news coverage of disasters and that this news coverage is crowded out by other newsworthy material”. This potentially poses an issue for effective giving, as the authors suggest that the levels of relief shouldn’t be driven by factors that are not related to the effectiveness of the relief.[They use a very interesting methodology; we might describe it a bit more.][(Are there similar results for charitable giving?)]
6.8 Reference dependence
The issue of the arguably disproportionate attention that natural disasters receive has been previously explored. Epstein (2006) discusses the high levels of support that natural disasters get. He drew on the work of Spence (2006), who reported that there was little relationship between the degree of need and the level of donations. This was illustrated by the private donations (at the time - 2006) of USD 1839 per person impacted by Hurricane Katrina, and USD 10 per person impacted by AIDS. Small (2010) posed potential questions that this might raise:
Although the response to dramatic events showcases a great human capacity for caring, the relative neglect of ongoing suffering reveals an equal albeit less attractive capacity for indifference. Can we make sense of this duality? Why do chronic conditions fail to move us even though they do so much harm?
Sudhir, Roy, and Cherian (2016) and Small (2010) have both suggested that reference dependence could provide an explanation for this issue, drawing on the work of Kahneman and Tversky (1979). Reference dependence is the notion that people don’t value absolute amounts, but instead view gains and losses relative to a reference point (Kahneman and Tversky 1979; A. Tversky and Kahneman 1991).
Studies from Small (2010) analyze this effect beyond the utility of the individual, evaluating reference dependence in the context of others’ utility. She predicts that because sympathy could be based on change, rather than on a state, then people who have experienced a loss will induce greater levels of sympathy than those who have chronic issues. Small (2010) summarizes the move from analyzing reference dependence for one’s own self, to analyzing reference dependence when viewing the circumstances of others, as follows:
Although the reference dependence modeled in prospect theory was conceived to explain the value of outcomes for the self, this research suggests that feelings and behavior with respect to the outcomes of others similarly responds to changes in welfare, not just absolute states. The studies provide insight into how the hardships of others appeal to our emotions. Bearing reference dependence in mind, the government, media and humanitarian agencies can better appeal to sympathy by shifting the focus of attention away from stats of need and instead to losses relative to a reference point.
She also suggests that the high levels of sympathy and donations seen with natural disasters like Hurricane Katrina, and the relative lack of sympathy for AIDS and malaria, are consistent with this idea.
The first study from Small (2010) examined whether people feel more sympathy for those who have had a negative change, by asking participants to rate their sympathy for people with health conditions that were described as either chronic or recent onset. She found that sympathy was significantly greater in the loss victims than the constant state victims, which she argued was the first evidence that sympathy may be greater in cases that involve a clear loss than for equivalent cases that involve a chronic misfortune.
Her second study examined how this idea translated from emotion into actual decision making, using a dictator game. It found that allocators in the dictator game gave more money, and felt greater sympathy, to recipients who had recently lost than to those recipients who stayed constant. This led her to conclude that “prosocial behaviour is not simply a function of others’ states, but additionally responds to whether that state is chronic or changed”.
In a study from Sudhir, Roy, and Cherian (2016), where they randomized charitable marketing content (described above), they also found evidence for sympathy being driven by reference dependency. They found that when victims were labelled as currently destitute, having been well-off before, this generated 50% more donors, and average donations per mailing increased by 33%. This is relative to advertisements depicting someone also described as destitute, but with an undescribed past. They concluded that this supported their hypothesis that those in chronic poverty are likely to get less sympathy than those who have suffered a change.
These reference dependence effects may only be significant when the target is an identified victim. In another study from Small (2010), she examined whether the affective quality of the victim moderates the reference dependence effect. She found, consistent with her other studies, that identified loss victims elicited more sympathy than identified constant-state victims. However, when this effect was examined with statistical victims, the effect was reversed, with the constant state victims getting more sympathy than the loss victims. Small suggests that this could be because when judgements are less emotion-based, as in the evaluation of statistical victims, people are more likely to account for the duration of the hardship. She evaluated the significance of this result as follows:
“This finding lends support to the key theoretical claim that reference-dependent aid judgements are driven by emotion-based thought; when the victim description is unemotional, the effect does not persist”9
6.9 Distance - Temporal (future problems/individuals) and Hypothetical (probability of occurance)
Overview
When considering how we can reduce suffering through effective altruism, immediate problems such as hunger and poverty are clearly visible. However, as well as the sentient beings alive today, altruists may consider challenges that can impact the welfare of future individuals as well. Benjamin Todd outlines this in his article on ‘longtermism’ (worth reading for a more in-depth discussion of the issue) as follows:
Since the future is big, there could be far more people in the future than in the present generation. This means that if you want to help people in general, your key concern shouldn’t be to help the present generation, but to ensure that the future goes well in the long-term… This thesis is often confused with the claim that we shouldn’t do anything to help people in the present generation. But the long-term value thesis is about what most matters - what should do about it is a further question. It might turn out that the best way to help those in the future is to improve the lives of people in the present, such as through providing health and education. The difference is that the major reason to help those in the present is to improve the long-term”
This consideration of future people and future problems includes well-known issues such as climate change. It also includes less-discussed) lower-probability or more distant events such as nuclear war, bioterrorism and AI related issues (Macaskill 2015). Individuals such as Ord (2020) and institutions like the Future of Humanity Institute have argued that these problems are important and neglected.
However, even motivating ‘giving for current issues’ presents many challenges; encouraging giving for (e.g.) ‘individuals who are yet to be’ born could prove even more difficult.11
See 80000 hours podcast with Toby Ord and the articles and books linked (especially ‘The Precipice’) for a more comprehensive introduction to the long-term future.
See also these resources from effectivealtruism.org, see here for a talk on the importance of future generations, and here for an overview of the long-run future as an issue.
Discounting future payoffs/assets/welfare
Considering how heavily we should weigh the future is a subject that has been raised frequently in the field of Economics (perhaps most notably climate economics in recent times), leading us to the idea (and various conceptions) of ‘discount rates’ (Ackerman 2008). Greaves (2017) outlines how discount rates can work as follows (unfold)
Suppose we have an opportunity to undertake an investment project, sacrificing k < 1 units of consumption today in order to secure an increase of 1 unit in consumption a time interval t later. Our basic question is: what is the threshold value of k at which the status quo becomes socially preferable to such a sacrifice? The answer to this question is the social discount factor for consumption at time t, R(t). (One also has a private discount factor, corresponding to private as opposed to social preferences; in the remainder of this article, the focus is on the social version.) One generally expects R to decline with time — we are willing to sacrifice more today to gain an increase in one unit of consumption tomorrow than to gain an increase of one unit of consumption next year. The decline, however, may be faster or slower, and for current purposes the rate of the decline is crucial.
The choice of discount rate is crucial in the evaluation of projects some of whose important effects are long-term. Analyses that use a higher discount rate will tend to favour the short term: projects requiring sacrifices in the short term for the sake of benefits in the further future will be more likely to fail cost-benefit tests.
An informal discussion from David Reinstein (unfold)
In considering whether there is a bias against helping people in the future we need to establish a baseline: At what rate (if at all) should an effective altruist weigh (far) future lives/pleasure/suffering as less-valuable as current or near-future lives? At what rate should we ‘discount’, if at all?
I think it’s actually not clear that we, as a society, should be discounting at all, either for consumption itself, or for some conception of ‘happiness’ or ‘utility’ (where here I’m thinking of ‘utility’ as something that does express well-being, not the very abstract economic concept of ‘the thing that we maximize’).
Individuals might discount future “pleasure”:
because they might not survive to the future (although this should enter in as a ‘survival discount’ tied to the probability of surviving)
because they are impatient (but this is potentially not in their own interest)
Individuals may discount future ‘income’ (in currency, inflation-adjusted) or ‘physical consumption’
If we think the ‘future me’ will be wealthier than the present me (consumption smoothing)
Or if we think I will be less sensitive to pleasure/pain
Businesses/investors may discount future earnings (again, inflation adjusted)
- If they think that there are other opportunities for ‘returns above inflation’ … because ‘returns to capital exceeds the rental cost of capital’
This should be true for growing societies as well, i.e., where there is real-GDP/capita growth (I’m being a bit loose here)
This is not the same as discounting future utility or welfare. There’s no clear reason we should do this.
Reasons the distant future may be neglected
Beyond the technical analysis of “at what rates people might discount the future”, Schubert (2020) outlines a few reasons why the distant future may receive less attention and action.13
He argues:
Present issues are likely more salient than future issues, and because these issues are more salient, they will be discussed more.
Uncertainty is likely an important factor. Even if you are concerned with future well-being, if it is unclear how you can impact the distant future, you may be more likely to focus on other issues.
Lack-of-polarisation: “When there is no political conflict, a question might get less discussed, even if it’s seen as morally important”.
While uncertainty could pose a problem for encouraging giving for distant future causes in particular, factoring in uncertainty is not necessarily a bias. It may be a reasonable choice, and there is further discussion of uncertainty and how this may link to a form of risk-aversion (see here).
The other barriers in this section also link to a potential lack of future-focused giving. For example, when it comes to the availability bias, we suggested that vividness and salience of events may influence how people estimate probabilities. Future events may be hard to picture, particularly more abstract risks like the dangers of artificial intelligence
Jason Schukraft:
Whenever longtermism is the subject, I always prefer to be specific about times. Some people reasonably consider 2070 to be “the distant future.” Other people would reserve the term for dates 10,000 years in the future or later. The former is obviously much less neglected than the latter.
We also discussed the identifiable victim effect. It could be significantly harder to create a story around a single victim if victims affected by that issue do not yet exist. Victims in the distant future may have to be statistical victims only, potentially eliciting less sympathy and consequently less donations. Charities that focus on the distant future may also have issues demonstrating evidence of their impact to donors, which can help to encourage donations (Karlan and Wood, 2017). However, the influence of demonstrating impact is mixed, as we further discuss here.
The impact of people being categorised into in-groups and out-groups could also be considered in analysing potential donations to distant-future causes. In their review of the literature, Sudhir, Roy, and Cherian (2016) suggest social identity works as follows:
Social identity has three major components: categorization, identification and comparison. Categorization is the process of putting people, including ourselves, into categories; for example, labelling a person as Chinese, black, female, or a lawyer are all different ways of categorization. Categorization also defines our self-image. Identification is the process by which we associate or identify ourselves with certain groups. We identify with in-groups, and we do not identify with out-groups. Finally, comparison is the process by which we compare in-groups with out-groups, creating a favourable bias toward the in-group. Overall, categorization of others as belonging to an in-group arouses feelings of greater closeness and responsibility and augments emotional response to their misfortune through greater sympathy (Brewer and Gardner (1996), Dovidio et al. (1997)) and willingness to help (Dovidio (1984), Dovidio et al. (1997)).
Distant-future generations might be mentally placed in their own category (considered separated from near-future and present generations). This categorical distance could mean that people identify with distant-future generations less. If they are seen as an out-group, this could pose a barrier for giving to distant future generations.14
(Evidence … work in progress)
What is the evidence that they are actually doing this (taking into account the ‘uncertainty’ issue (we know less about how to help the far-future).
What factors drive this bias or mitigate it?
This section was written by David Reinstein and Luke Arundel, with contributions from Annabel Rayner.↩︎
GiveWell focuses on charities that provide concrete interventions to help people alive today and soon-to be born, and focuses their evaluation on measurable outcomes such as [“QALY and DALY”](https://en.wikipedia.org/wiki/Disability-adjusted_life_year.↩︎
citation needed↩︎
Crimston’s (2016) ‘moral expansiveness scale’ aims to capture breadth (a ‘representative spread of human and nonhuman entities’) as well as depth (four defined boundaries expressing degree of obligation and moral standing). It expands earlier models and scales↩︎
Ref: Trope, Yaacov; Liberman, Nira (2010). “Construal-level theory of psychological distance” (PDF). Psychological Review. 117 (2): 440–463. doi:10.1037/a0018963 D. Crimston et al. (2018)↩︎
Loewenstein and Small (2007) propose a relationship between sympathy and helping behaviour, drawing on experimental findings from Psychology and Economics. This in turn relates to the dual system model (as discussed by Kahneman (2011)). This model distinguishes between System 1, which makes initial judgements (using ‘affective cognition’); these are monitored by a more ‘deliberative cognition’ System 2 (Morewedge and Kahneman 2010).(However, this model has been subjected to some criticism, Evans and Stanovich (2013) discuss the debate around dual system theories.)↩︎
Ref: small2007_friends small_loewenstein_07_scarecrow↩︎
This effect is discussed further under the ‘statistical/identifiable victims’ quantitative bias.↩︎
Here we should distinguish individual’s own response to shocks from their sympathetic/empathetic and other-regarding-response to learning about others’ shocks. I would also distinguish reference points a bit more carefully; there are all sorts of effects involving reference points.↩︎
This is particularly relevant to causes and charities dealing with the medium-term and long-term future.↩︎
However, Jason Schukraft notes: There may be cases where for some individuals longtermist causes are less “distant,” all-things-considered, than neartermist causes. If you’re a wealthy Silicon Valley tech entrepreneur, donating to an AI Safety group might be psychologically easier than donating to Give Directly.↩︎
For further reading on discount rates see Weitzman (1998), Ebert and Prelec (2007), Farmer and Geanakoplos (2009)↩︎
Schubert also discusses the psychology of long-termism and existential risk here↩︎
Whether those in the future are indeed commonly percieved or treated as an out-group may be an interesting research direction. This is something we need to look into more for this present survey.↩︎