The 100/nonfiction
So, I basically took 2020-2021 off in almost every way and am slowly getting back to a new normal. And one thing I finally did was finish The 100. And now I am mad--mad enough that I've lain in bed fretting for a couple of nights! Which may be displacement but is still annoying. I haven't been this mad at an ending since BSG. It was all pointless! The solution to humanity was to get rid of it! This was not a happy ending but the show seemed to want it to be! Killing Bellamy was in fact pointless! The show seemed to want Clarke's last killing to be badass ("pencils down" is indeed badass) but also wrong? Even though the dude was literally running Omelas? I cannot accept that Bellamy was right to be into this guy, but I can accept that he would be, though Octavia and Echo also got over that pretty damn quick. Urgh. I don't want the show to have ended in S2, despite that season's perfection, but strange women lurking in stars handing out ascension is no way to run a system of government.
Sheera Frenkel & Cecilia Kang, An Ugly Truth: Inside Facebook’s Battle for Domination: Yes, FB is as bad as you think it is. From the lack of internal safeguards that let male engineers look up the information of women they were interested in-one of whom tracked down a woman with whom he’d had a fight after she left the hotel where they’d been staying--to crushing competition, it’s an organization out of control. The worst stories are about FB’s indifference to non-English speakers, where an inability to even understand hundreds of languages left FB ignorant of the genocidal campaigns being conducted on the site, particularly in Myanmar. The one person in charge of monitoring Myanmar spoke only Burmese, out of the about 100 languages present-as if a German speaker were responsible for moderating all Europe. Then FB refused to share information about deleted posts to prove the military’s responsibility for the genocide. “Facebook would cooperate … only if the United Nations created a mechanism to investigate human rights crimes. When [the relevant person] pointed out that the United Nations had such a system in place, known as the Independent Investigative Mechanism for Myanmar, the Facebook representative looked at him with surprise and asked him to explain.” With great power does not, in fact, come great responsibility-not unless either law or an ethical compass provide it.
Eyal Press, Dirty Work: Essential Jobs and the Hidden Toll of Inequality in America: Really good, horrifying book about the way that we make low-paid workers responsible for doing the worst things in our society, then blame them for having done those things within structures that make their bad behavior almost inevitable. This “dirty work” requires the tacit consent of “good people,” maybe even more in a democracy than in an autocracy. After all, how are we going to deal with all the mentally ill people we have decided not to care for outside of prison, or fight the wars we’ve decided to fight without lots of troops on the ground? Covers prison health care, military drone operators, slaughterhouse workers, and other jobs that most people don’t aspire to or pretend are noble (not, for example, police officers and teachers, who are also given individualized blame for structural failings including the structural failings that socialize them into behaving badly, but are also lionized in the abstract). For example: “Nobody told Curtis and his fellow guards to get brutal. But no one really needed to tell them this. It was enough to pay them modest salaries to enforce order in overcrowded, understaffed prisons that were neither equipped nor expected to do much else.”
These jobs tend to evoke disgust and shame, affecting both how others see the workers and often how the workers see themselves-Press discusses the idea of “moral injury,” especially in the context of drone operators. Although you might think they’d treat death like a video game, many of them instead react negatively-and they end up seeing more death and destruction than most Special Forces on the ground. Moral injury is a useful concept, Press argues, because PTSD, while also descriptive, can depoliticize and individualize what is a problem of what the system asked the individual to do. Meanwhile, drone operators aren’t seen as “real” soldiers, a status deriving “from the very thing that made drone warfare appealing to politicians and the public”-it saved money and lives on our side.
Press emphasizes that many of the workers he talks to are not the primary victims of the systems they work in-prisoners, foreigners subject to drone strikes, and maybe animals are--but they are also suffering as they cause suffering, and we should not let individualized blame obscure that they are doing what we as a society want them to do. This is particularly true because these are jobs disproportionately filled by poor people without other opportunities and people of color, walled off from others by geography, fences, and other barriers so we don’t have to think about them. Hedge fund guys, disproportionately white, don’t face the same stigma even as they do lots of damage, and they are rewarded with money and prestige for doing so. So, when the BP oil rig exploded, even the workers’ families understood that images of oil-covered birds would generate more public outrage than pictures of the loved ones they’d lost. But when these workers try to challenge unsafe conditions, they find they’re easily replaced, unlike high-tech workers whose protests are often heeded. (Interesting contrast to Tyler Schultz’s narrative of whistleblowing about Theranos-he definitely suffered, but his suffering had a point, which most of these workers can’t say.) “What do we owe these workers? At a minimum, it seems to me, we owe them the willingness to see them as our agents, doing work that is not disconnected from our own daily lives, and to listen to their stories, however unsettling what they tell us may be.” (Of course, this framing accepts that they aren’t likely to be reading the same books as “we” are.)
You should read it; it’s mostly about the US though there are a few fascinating comparisons, such as to research on the prison system in France, which also found that guards were ashamed of what they did for a living. In Norway, where the prison system is much more rehabilitative, the staff seemed much prouder (though he doesn’t have the same depth of ethnographic data).
Behavioral Science in the Wild, Nina Mažar & Dilip Soman, eds.: Really interesting overview of difficulties translating behavioral science/nudges into working policy. Consistently, policies have much smaller effects than they do in the lab-a “voltage drop.” That doesn’t mean we should stop-the average effect of a government nudge is significant and positive, and even a 8% change in behavior can mean a lot of people helped, but it is worth thinking about how we go from 30% in the lab to 8% in practice. A lot of this seems to be explained by publication bias: studies that don’t show a (big) result don’t get published.
Sometimes the nudge was the wrong lever. For example, whether organ donation is the default or opt-in varies across countries; when Wales flipped the default, the organ supply didn’t seem to increase, perhaps because a default opt-in can be overridden by family members. It’s possible that, pre-intervention, a country’s default simply matched its population’s overall willingness to donate organs.
Sometimes the studied nudges systematically vary from implemented ones. Academic nudges are more likely to be delivered in person and to be about asking people to choose between 2 options, both of which are more effective than other types. The government may face barriers changing a service from opt in to opt out despite the effectiveness of that kind of nudge. Still, nudges are generally quite cheap for government, with often no marginal cost to changing how they communicate (though if we wanted them more effective, we’d need more people to do things in person).
But backfires are possible. A South Korean test of messages to deter excess credit card spending worked for 12% who were the heaviest spenders but backfired for 88%; maybe better targeting would have helped. A Mexican pension experiment showing people how much they’d have for retirement with the goal of encouraging savings showed a similar backfire effect (though that may have shown that they were targeting the wrong problem for that group, which was invested in a low performing pension fund).
The contributors constantly emphasize that context is everything. “Every detail of how the intervention is implemented matters. The fundamental behavioral insight, say peer comparison, may transfer as is, but how it is packaged and delivered often does not.” So, for example, different groups may need different tones, different timing/frequencies of messages, and different graphics. It was fascinating to learn that, at ASU, “a four-year school with a traditional student body,” an intervention to help students get more financial aid needed to email both students and their parents. “But at CUNY, students tend to be older, commonly with children of their own, and are often the first in their families” to go to college; they didn’t need emails to their parents. ASU emails written in a friendly, casual tone were “approachable and unintimidating,” but they seemed “unprofessional and untrustworthy” to CUNY students, “who see college as providing a service more than an experience.” For recorded messages, having the research assistants record them leads to hangups, but voice actors and clear scripts get listened to more.
There’s an interesting chapter on debiasing interventions and how we shouldn’t try to change people’s minds, but rather “redesign their systems and environments, so that biases have no place to hide.” You may have heard of blind auditions and their effects on gender bias, but there’s also guidance on job listings, replacing words like “entrepreneurial” and “strong” with neutral synonyms like “creative” and “dedicated”-the latter attracted more women, who represented 4% more of the resulting pool, but also more applicants in general, especially “men who were more weakly identified with their gender.” In evaluating job candidates, it helps to compare individual responses “horizontally,” by “looking at one question or criterion for all applicants and then moving on to the next question or criterion,” instead of trying to assess each candidate separately. The chapter didn’t talk about this, but I expect it may fight back against the known problem that position requirements get shaped to the strengths of the most attractive male applicant. Likewise, it may help to change numerical evaluations, as dumb as that sounds. It turns out that “[e]valuators are actually less likely to give 10/10 - an indicator of perfection and brilliant performance - to high-performing women,” but a 1-6 scale instead “closed the gender gap on perfect scores.” And within organizations, an opt-out scheme for consideration for promotion helps women by eliminating the gender gap in who puts themselves forward.
What about environmental acts? Can we avoid moral licensing/backlash? How do we get people to think of themselves as the kind of people who are environmentally responsible? Turns out that people who take “small, ‘token’ public actions such as a social media post supporting environmentalism” can behave less responsibly afterwards, whereas if you can get them to take “larger, effortful actions” or private actions like a home energy audit, that can have positive spillover effects on their other actions.
But again, everything has to be contextualized. One great chapter discussed localization in Kenya: The researchers started with “a well-cited intervention from psychology to induce mild stress …, in which those assigned to the treatment group were given a mock job interview in front of a panel of ‘experts’ while those assigned to the control group were simultaneously to give a speech about a friend.” It didn’t work in Kenya, because public speaking was not culturally understood as stressful and also, as one respondent asked, “Why are we being interviewed by butchers?” The experiment had followed the Western convention of “white coats = scientists.”
Alex Wellerstein, Restricted Data: Fascinating history of US nuclear secrecy. Big takeaways: after the initial burst of post-Hiroshima/Nagasaki fights about secrecy, the US declassified substantial amounts in order to promote the business of nuclear power; capitalism opposed national security (if you assume, which was hotly contested, that secrecy promoted national security). It’s also a story about the importance of know-how and material access versus just abstract scientific knowledge, which is apparently not as much help in making a bomb as you might have thought-Wellerstein refers to this as the idea that there was “a secret” or “nuclear secrets” as opposed to a thick web of know-how.
Tyler Schultz, Thicker than Water: Audible original in the voice of a Theranos whistleblower (and not accidentally, grandson of George Schultz). Really powerful illustration of how even someone raised to value truth and public service, with economic security and cultural privilege, can be intimidated to the point of near-breakdown by legal threats-even those from a fraudster. He speaks affectingly about how he felt betrayed by his beloved grandfather and about how the pressure of the threat of a lawsuit against him, and the associated surveillance, made him so distressed that the reason that he didn’t buy a gun was that he was too worried he might use it on himself.
David M. Perry & Matthew Gabriele, The Bright Ages: During the Dark Ages (as you may know them), Europeans were actually doing all kinds of things, including interacting with Africa and Asia, developing religion and culture, and both fighting and trading amongst themselves, not just going on Crusades and killing Jews (though they did a fair amount of those things too). Leans too hard on repeating “the Bright Ages” for my taste, and focuses on kings and queens more than ordinary people, but I thought it fit interestingly with David Graeber’s last book about how the quiet/meaningless/unrecorded periods of history might well have been better, and more meaningful, for the average person than the nearby exciting/war-filled times.
comments on DW | reply there. I have invites or you can use OpenID.
Sheera Frenkel & Cecilia Kang, An Ugly Truth: Inside Facebook’s Battle for Domination: Yes, FB is as bad as you think it is. From the lack of internal safeguards that let male engineers look up the information of women they were interested in-one of whom tracked down a woman with whom he’d had a fight after she left the hotel where they’d been staying--to crushing competition, it’s an organization out of control. The worst stories are about FB’s indifference to non-English speakers, where an inability to even understand hundreds of languages left FB ignorant of the genocidal campaigns being conducted on the site, particularly in Myanmar. The one person in charge of monitoring Myanmar spoke only Burmese, out of the about 100 languages present-as if a German speaker were responsible for moderating all Europe. Then FB refused to share information about deleted posts to prove the military’s responsibility for the genocide. “Facebook would cooperate … only if the United Nations created a mechanism to investigate human rights crimes. When [the relevant person] pointed out that the United Nations had such a system in place, known as the Independent Investigative Mechanism for Myanmar, the Facebook representative looked at him with surprise and asked him to explain.” With great power does not, in fact, come great responsibility-not unless either law or an ethical compass provide it.
Eyal Press, Dirty Work: Essential Jobs and the Hidden Toll of Inequality in America: Really good, horrifying book about the way that we make low-paid workers responsible for doing the worst things in our society, then blame them for having done those things within structures that make their bad behavior almost inevitable. This “dirty work” requires the tacit consent of “good people,” maybe even more in a democracy than in an autocracy. After all, how are we going to deal with all the mentally ill people we have decided not to care for outside of prison, or fight the wars we’ve decided to fight without lots of troops on the ground? Covers prison health care, military drone operators, slaughterhouse workers, and other jobs that most people don’t aspire to or pretend are noble (not, for example, police officers and teachers, who are also given individualized blame for structural failings including the structural failings that socialize them into behaving badly, but are also lionized in the abstract). For example: “Nobody told Curtis and his fellow guards to get brutal. But no one really needed to tell them this. It was enough to pay them modest salaries to enforce order in overcrowded, understaffed prisons that were neither equipped nor expected to do much else.”
These jobs tend to evoke disgust and shame, affecting both how others see the workers and often how the workers see themselves-Press discusses the idea of “moral injury,” especially in the context of drone operators. Although you might think they’d treat death like a video game, many of them instead react negatively-and they end up seeing more death and destruction than most Special Forces on the ground. Moral injury is a useful concept, Press argues, because PTSD, while also descriptive, can depoliticize and individualize what is a problem of what the system asked the individual to do. Meanwhile, drone operators aren’t seen as “real” soldiers, a status deriving “from the very thing that made drone warfare appealing to politicians and the public”-it saved money and lives on our side.
Press emphasizes that many of the workers he talks to are not the primary victims of the systems they work in-prisoners, foreigners subject to drone strikes, and maybe animals are--but they are also suffering as they cause suffering, and we should not let individualized blame obscure that they are doing what we as a society want them to do. This is particularly true because these are jobs disproportionately filled by poor people without other opportunities and people of color, walled off from others by geography, fences, and other barriers so we don’t have to think about them. Hedge fund guys, disproportionately white, don’t face the same stigma even as they do lots of damage, and they are rewarded with money and prestige for doing so. So, when the BP oil rig exploded, even the workers’ families understood that images of oil-covered birds would generate more public outrage than pictures of the loved ones they’d lost. But when these workers try to challenge unsafe conditions, they find they’re easily replaced, unlike high-tech workers whose protests are often heeded. (Interesting contrast to Tyler Schultz’s narrative of whistleblowing about Theranos-he definitely suffered, but his suffering had a point, which most of these workers can’t say.) “What do we owe these workers? At a minimum, it seems to me, we owe them the willingness to see them as our agents, doing work that is not disconnected from our own daily lives, and to listen to their stories, however unsettling what they tell us may be.” (Of course, this framing accepts that they aren’t likely to be reading the same books as “we” are.)
You should read it; it’s mostly about the US though there are a few fascinating comparisons, such as to research on the prison system in France, which also found that guards were ashamed of what they did for a living. In Norway, where the prison system is much more rehabilitative, the staff seemed much prouder (though he doesn’t have the same depth of ethnographic data).
Behavioral Science in the Wild, Nina Mažar & Dilip Soman, eds.: Really interesting overview of difficulties translating behavioral science/nudges into working policy. Consistently, policies have much smaller effects than they do in the lab-a “voltage drop.” That doesn’t mean we should stop-the average effect of a government nudge is significant and positive, and even a 8% change in behavior can mean a lot of people helped, but it is worth thinking about how we go from 30% in the lab to 8% in practice. A lot of this seems to be explained by publication bias: studies that don’t show a (big) result don’t get published.
Sometimes the nudge was the wrong lever. For example, whether organ donation is the default or opt-in varies across countries; when Wales flipped the default, the organ supply didn’t seem to increase, perhaps because a default opt-in can be overridden by family members. It’s possible that, pre-intervention, a country’s default simply matched its population’s overall willingness to donate organs.
Sometimes the studied nudges systematically vary from implemented ones. Academic nudges are more likely to be delivered in person and to be about asking people to choose between 2 options, both of which are more effective than other types. The government may face barriers changing a service from opt in to opt out despite the effectiveness of that kind of nudge. Still, nudges are generally quite cheap for government, with often no marginal cost to changing how they communicate (though if we wanted them more effective, we’d need more people to do things in person).
But backfires are possible. A South Korean test of messages to deter excess credit card spending worked for 12% who were the heaviest spenders but backfired for 88%; maybe better targeting would have helped. A Mexican pension experiment showing people how much they’d have for retirement with the goal of encouraging savings showed a similar backfire effect (though that may have shown that they were targeting the wrong problem for that group, which was invested in a low performing pension fund).
The contributors constantly emphasize that context is everything. “Every detail of how the intervention is implemented matters. The fundamental behavioral insight, say peer comparison, may transfer as is, but how it is packaged and delivered often does not.” So, for example, different groups may need different tones, different timing/frequencies of messages, and different graphics. It was fascinating to learn that, at ASU, “a four-year school with a traditional student body,” an intervention to help students get more financial aid needed to email both students and their parents. “But at CUNY, students tend to be older, commonly with children of their own, and are often the first in their families” to go to college; they didn’t need emails to their parents. ASU emails written in a friendly, casual tone were “approachable and unintimidating,” but they seemed “unprofessional and untrustworthy” to CUNY students, “who see college as providing a service more than an experience.” For recorded messages, having the research assistants record them leads to hangups, but voice actors and clear scripts get listened to more.
There’s an interesting chapter on debiasing interventions and how we shouldn’t try to change people’s minds, but rather “redesign their systems and environments, so that biases have no place to hide.” You may have heard of blind auditions and their effects on gender bias, but there’s also guidance on job listings, replacing words like “entrepreneurial” and “strong” with neutral synonyms like “creative” and “dedicated”-the latter attracted more women, who represented 4% more of the resulting pool, but also more applicants in general, especially “men who were more weakly identified with their gender.” In evaluating job candidates, it helps to compare individual responses “horizontally,” by “looking at one question or criterion for all applicants and then moving on to the next question or criterion,” instead of trying to assess each candidate separately. The chapter didn’t talk about this, but I expect it may fight back against the known problem that position requirements get shaped to the strengths of the most attractive male applicant. Likewise, it may help to change numerical evaluations, as dumb as that sounds. It turns out that “[e]valuators are actually less likely to give 10/10 - an indicator of perfection and brilliant performance - to high-performing women,” but a 1-6 scale instead “closed the gender gap on perfect scores.” And within organizations, an opt-out scheme for consideration for promotion helps women by eliminating the gender gap in who puts themselves forward.
What about environmental acts? Can we avoid moral licensing/backlash? How do we get people to think of themselves as the kind of people who are environmentally responsible? Turns out that people who take “small, ‘token’ public actions such as a social media post supporting environmentalism” can behave less responsibly afterwards, whereas if you can get them to take “larger, effortful actions” or private actions like a home energy audit, that can have positive spillover effects on their other actions.
But again, everything has to be contextualized. One great chapter discussed localization in Kenya: The researchers started with “a well-cited intervention from psychology to induce mild stress …, in which those assigned to the treatment group were given a mock job interview in front of a panel of ‘experts’ while those assigned to the control group were simultaneously to give a speech about a friend.” It didn’t work in Kenya, because public speaking was not culturally understood as stressful and also, as one respondent asked, “Why are we being interviewed by butchers?” The experiment had followed the Western convention of “white coats = scientists.”
Alex Wellerstein, Restricted Data: Fascinating history of US nuclear secrecy. Big takeaways: after the initial burst of post-Hiroshima/Nagasaki fights about secrecy, the US declassified substantial amounts in order to promote the business of nuclear power; capitalism opposed national security (if you assume, which was hotly contested, that secrecy promoted national security). It’s also a story about the importance of know-how and material access versus just abstract scientific knowledge, which is apparently not as much help in making a bomb as you might have thought-Wellerstein refers to this as the idea that there was “a secret” or “nuclear secrets” as opposed to a thick web of know-how.
Tyler Schultz, Thicker than Water: Audible original in the voice of a Theranos whistleblower (and not accidentally, grandson of George Schultz). Really powerful illustration of how even someone raised to value truth and public service, with economic security and cultural privilege, can be intimidated to the point of near-breakdown by legal threats-even those from a fraudster. He speaks affectingly about how he felt betrayed by his beloved grandfather and about how the pressure of the threat of a lawsuit against him, and the associated surveillance, made him so distressed that the reason that he didn’t buy a gun was that he was too worried he might use it on himself.
David M. Perry & Matthew Gabriele, The Bright Ages: During the Dark Ages (as you may know them), Europeans were actually doing all kinds of things, including interacting with Africa and Asia, developing religion and culture, and both fighting and trading amongst themselves, not just going on Crusades and killing Jews (though they did a fair amount of those things too). Leans too hard on repeating “the Bright Ages” for my taste, and focuses on kings and queens more than ordinary people, but I thought it fit interestingly with David Graeber’s last book about how the quiet/meaningless/unrecorded periods of history might well have been better, and more meaningful, for the average person than the nearby exciting/war-filled times.
comments on DW | reply there. I have invites or you can use OpenID.