Episode 162: How the “Data-Driven” Label Sanitizes Cruel Austerity Politics

Citations Needed | June 1, 2022 | Transcript

Barack Obama and Melinda Gates visit the Gates-financed TechBoston Academy in Dorchester, Massachusetts, in 2011. (Doug Mills / The New York Times)


Intro: This is Citations Needed with Nima Shirazi and Adam Johnson.

Nima Shirazi: Welcome to Citations Needed, a podcast on the media, power, PR and the history of bullshit. I am Nima Shirazi.

Adam Johnson: I’m Adam Johnson.

Nima: You can follow the show on Twitter @CitationsPod, Facebook Citations Needed, and become a supporter of our work through Patreon.com/CitationsNeededPodcast. All your support through Patreon is so incredibly appreciated as we are 100 percent listener funded.

Adam: Yes, if you can, please give to our Patreon, it helps us out a lot. We have special little goodies there, over 110 or 120 or so little mini episodes, it helps keep the episodes themselves free, and helps keep the show sustainable.

Nima: “Follow The Data” is the name of a Bloomberg Philanthropies podcast that debuted in 2016. “How Data Analysis Is Driving Policing,” a 2018 NPR headline read. “Data suggests that schools might be one of the least risky kinds of institutions to reopen,” an opinion piece in The Washington Post told us in the early days of the COVID-19 pandemic.

Adam: Over the last 20 years or so, a trend of labeling concepts as quote “data-driven” emerged. It applied, and continues to apply, to policies affecting everything from education to public health to policing to journalism. Decisions affecting these areas will be more thoughtful, the idea goes, when informed and supported by data. In many ways, this is a welcome development: The idea that a rigorously scientific collection of information via surveys, observation, and other methods would make policies and media stronger seems unimpeachable.

Nima: But this isn’t always the case. While gathering “data” is a potentially beneficial process, the process alone isn’t inherently good, and is too often used to obscure important and requisite value-based or moral questions, assert contested ideological priors and traffic in right-wing austerity premises backed by monied interests. When our media tell us a largely unpopular, billionaire-backed idea like school privatization or “targeted” policing or tax incentive handouts to corporations all have merit they’re backed by “the data,” what purpose does this framing actually serve? Where does the data come from? Who is funding this gathering of data? What data are we choosing to care about and, perhaps most important of all, what data are we choosing to ignore?

Adam: On today’s episode, we’ll look at the development of the push to make everything data-driven, examine who defines what counts as “data,” which forces shape its sourcing and collection, and how the fetishization of quote-unquote “data” as something that exists outside and separate from politics is more often than not, less a methodology for determining truth, and more a branding exercise for neoliberal ideological production and reproduction.

Nima: Later on the show, we’ll be speaking with two guests: Abigail Cartus, an epidemiologist at Brown University where she focuses on perinatal health and overdose prevention in her work at The People, Place & Health Collective, a Brown School of Public Health research laboratory.

[Begin Clip]

Abigail Cartus: This push towards policy-based evidence-making just conducting analyses that justify a political goal that has already been decided with basically no opportunity for public input or public deliberation, and then forcing the debate over that to take place on the terrain of statistical methods, data quality, you know, things that laypeople and the general public really can’t always weigh in on.

[End Clip]

Nima: And also with Justin Feldman, an epidemiologist and a Health and Human Rights Fellow at the Harvard FXB Center for Health and Human Rights.

[Begin Clip]

Justin Feldman: I think there is this liberal value of trusting science, trusting expertise, and some of what we’re seeing in Emily Oster’s work, in the writing by David Leonhardt in New York Times Daily Newsletter, these are people who are basically laundering right-wing arguments and trying to make them appealing to liberals.

[End Clip]

Adam: So to start off with a typical caveat we do in the show, we of course, are not against data collection. What we’re examining here is capital “D” Data journalism as a kind of branding exercise. This is similar to our episode we did on investigative journalism, similar to the episode we did on capital “S” Science as a tool of white supremacy and colonialism. Where we’re not opposed to the idea as such, obviously data can be good. We use data on the show, I’d like to think.

Nima: Yeah, that’s a lot of our citations.

Adam: What we’re criticizing is this idea of data journalism as something that exists outside the realm of politics or ideology, and that people promote their particular political agenda, whatever it may be, and are coy and cheeky about the fact that they have one and insist that they simply believe in X because that’s where the data led them. So the data is this kind of abstract thing that exists in nature and that a person sat down and followed the data to their conclusion instead of having a sort of preconceived notion of what they believed before, it’s sort of this, it’s kind of a quasi-scientific method. The data told them what to believe, they didn’t have something to believe, and what we’re going to argue is that’s kind of mostly bullshit. There are instances where I do think that around the margins, data can maybe push someone one way or the other on a particular given issue, but mostly as a kind of branding exercise if somebody claims that they’re post ideological, and that they’re just following the data, you probably need to check for your wallet.

Nima: Yeah, it’s true. I mean, similar to the ideas of, you know, tech, and AI, these aren’t just naturally occurring organic things in nature, they are created by people and therefore, are flawed the way people are flawed. They have the biases the way people have biases, it’s sometimes garbage in, garbage out in terms of how these things are produced, created, what methodologies are used, what the data is actually seeking to find, and therefore how it is analyzed and used, and as you said Adam, you know, we’re not against data, we’re not against science, do all the stuff, do all the studying, do all the research, do all the surveys. Cool. But what we are investigating on this show, is how this trend of journalism tells us more about what the journalists or the data journalism verticals, what is ideologically motivating them, what are we seeing when we’re just supposed to quote-unquote “follow the data.”

Now, there’s no singular definition of quote-unquote “data journalism.” When ostensible journalistic authorities do try to define this term, those definitions usually are not terribly specific or particularly meaningful. For example, Alex Howard of the Columbia Journalism School’s Tow Center describes data journalism as, quote, “gathering, cleaning, organizing, analyzing, visualizing and publishing data to support the creation of acts of journalism,” end quote. The American Press Institute states that those who practice data journalism, quote, “tend to agree on one principle: data journalism is, first and foremost, journalism. It simply uses data as a source in addition to humans,” end quote.

Adam: So on the origins of data journalism, data and journalism have, of course, of course, have been inseparable since journalism’s inception. It would seem impossible to think of an example of journalism that isn’t dependent on some form of data. W.E.B. DuBois and Florence Nightingale famously designed imagery meant to convey information that we might now, perhaps somewhat insultingly, call “infographics” or “data visualizations.” But by many accounts, “data journalism” started to be considered a formal discipline in the mid-20th century, as computing technology became more accessible and large numbers were able to be quantified and represented.

One of the most commonly cited early examples of “data journalism” comes from the Detroit Free Press’s reporting on the Detroit uprisings of July 1967. Journalist Philip Meyer conducted surveys of political stances among 437 Black people in the vicinity of the uprisings in East and West Detroit. The surveys were sponsored by multiple major Detroit-based businesses and their philanthropic arms, including the Campbell-Ewald Foundation and Henry Ford II, and would garner a Pulitzer Prize for Detroit Free Press.

The survey itself was in theory a useful form of journalism, as it gave much attention to the voices of an oppressed and aggrieved group. But the piece, again, funded by corporate philanthropy, still editorialized against the quote-unquote “rioters.”

Nima: Yeah, so “rioters” is something that we’ll hear again and again, that apparently, you know, a term that the data reinforce. So here’s an excerpt from one of the articles by Philip Meyer, again of the Detroit Free Press. This is from August 20, 1967, quote:

These, then are the rioters: Young people, raised in the North, with little concern for their fellowmen and a frustration in meeting near-term goals — people susceptible to the black nationalist philosophy that the law and order of a white-built society is not worth preserving.

In contrast, look at Detroit’s Negro community as a whole:

Only 10 percent of the Negroes in the survey believe their situation is getting worse. Fifty-one percent say they are better off than they were three years ago.

End quote.

Now, Meyer would go on to call this style of reporting, quote, “precision journalism,” end quote, and would author a book on this topic in the next decade.

Adam: Inspired by Meyer’s reporting on the uprisings, or as they call them, “riots,” the Guardian introduced a “data-driven” series in 2011 on the quote-unquote “riots” in Tottenham, North London, after uprisings began in response to the 2011 police killing of Black London resident Mark Duggan. One of the headlines from this data-driven project read, quote, “Riots report shows London needs to maintain police numbers, says mayor” — that mayor was Boris Johnson. The paper also surveyed residents in an initiative called “Reading the riots — community conversations,” yet the majority of the interviews — at least two-thirds, probably more — led with condemnation of or condescension to the so-called “rioters,” with headlines reading: “Peckham riots: ‘We need to help young people make better decisions.’” “Liverpool riots: ‘They wrecked what little they have here.’” “Croydon riots: ‘People were like animals that night.’” “Birmingham riots: ‘It was a bandwagon thing.’”

The, quote, “Reading the riots” project happened just a few years after “data journalism” had started to become a full-fledged genre, spearheaded by technocratic aspiring media personalities. In 2008, Nate Silver, whose dubious poll-based work we spent much time dissecting in Episode 87, launched the quote, “data blog” FiveThirtyEight. Prior to launching FiveThirtyEight, Silver had been pseudonymously blogging the 2008 election outcome predictions at the Daily Kos. This was the time that so-called data journalism began to explode, but more on that later.

Nima: Yeah, so, while data journalism didn’t really begin in earnest, or at least, you know, being called that as a new genre of the industry until really the Obama era, the first instances of the term quote-unquote “data-driven” were appearing in news media around the early 1980s. The term was primarily used by industry, showing up in computer science job ads and business management news articles. But eventually, the term’s application broadened. Around the late 1990s and early 2000s, it appeared more frequently in local news reports about schools, which increasingly touted their use of — what else? — quiet-unquote “data” like standardized test scores in order to preserve funding. As we’ve discussed on Citations Needed before, schools are one of the, if not the most common targets of “data”-related sanitization of right-wing agendas.

One of the earlier education-based data driven articles appeared in an Indiana newspaper, The Palladium-Item from Richmond, Indiana, at the beginning of 1999. In an article, “Testing Education,” the sub headline reads, “Richmond schools become data-driven,” and the article starts like this, quote:

By requiring administrators to systematically collect hard numbers about student performance for publication, ‘I’m trying to immerse data into the culture,’ said Associate Superintendent Sharon Studebaker-Puckett. ‘I want this to be the most important document we have.’

The article continues this way:

The goal is to use the data to identify trends district-wide and in individual schools and use that information to improve curriculum and instruction, school officials said.

‘In the past, we’ve tended to add programs, but we didn’t have a good way to evaluate whether one program was more successful than another, and in this time of increasingly scarce financial resources, we have to be smarter about how we use those resources,’ school board President Robert Green said.

In a similar article, also from an Indiana newspaper, about a year later, this is from February 2000, there’s this headline, “School scores reflect work,” and the article talks about school rankings in the state of Indiana quote:

Smoky Row Elementary third-graders jumped from number nine last year to number three this year.

Smoky Row Elementary School Principal Rhonda Buzan said the school scored well for several reasons.

‘We are a data-driven school,’ Buzan said. ‘Our testing and assessing that we do promotes common instructional focus.’

End quote.

Adam: This “data-driven” model culminated in the 2002 No Child Left Behind Act, legislation enacted by George W. Bush that required all states to test every child annually in grades 3 through 8 in reading and math and report test scores by race, ethnicity, low-income status, disability status, and English proficiency. The NCLB threatened to slash schools’ Title I funds if their test scores didn’t meet certain benchmarks in a stated effort to improve US education.

The No Child Left Behind Act has been roundly criticized for replacing curriculum in subjects like arts, history, civics, physical education, science, and foreign language with standardized test training. Many have noted that it was responsible for driving teachers and students out of public schools, unfairly testing special-ed and English language learner students, and further entrenching funding disparities in schools in wealthy vs. poor neighborhoods by increasing punishments for quote-unquote “underperforming” schools. Education historian Diane Ravitch in 2013 would state, when commenting on the No Child Left Behind Act, quote, “The thirst for data became unquenchable.” Of course, following this, this very specific, very ideologically driven definition of quote-unquote “data schools” were then deemed failures and were at risk of being closed or turned into charter schools, which, of course, was the goal.

George W. Bush waves during a bill-signing ceremony for the No Child Left Behind Act in January 2002. (Tim Sloan / AFP / Getty Images

The Obama administration doubled down on this policy. In 2009, the administration implemented its Race to the Top program, which was functionally identical to the No Child Left Behind Act, with some minor differences. As implied by its title, the initiative encouraged states to “compete,” mostly via standardized testing, in order to qualify for federal funding. Race to the Top continued to punish districts, schools, and teachers who failed to produce consistently rising test scores and forced states to adopt uniform curricular standards in order to compete for federal funding. This was an agenda pushed by Bill Gates, helped subsidize it and basically, therefore, became what the Washington Post called the default secretary of education under Obama. And to be eligible for the program, states had to increase the number of charter schools and, for quote-unquote “underperforming” schools take steps like firing staff and closures. Despite the longstanding criticism of the similarly structured No Child Left Behind Act, legacy media like the New York Times lavished praise upon Race to the Top:

Nima: Now, of course, this was an Obama administration policy and here is a clip of President Obama himself explaining the goal of new quote-unquote “data systems” to track student performance as part of the program Race to the Top’s alleged reforms.

[Begin Clip]

Barack Obama: We urge states to use cutting edge data systems to track a child’s progress throughout their academic career, and to link that child’s progress to their teachers so we know what’s working, and what’s not working in the classroom.

[End Clip]

Nima: Much of the data, as one might by this point guess, was standardized test scores. Curiously, data not mentioned by Obama offered some unflattering insight into the program. A 2009 National Research Council report found that the competitive structure engineered by Race to the Top had no empirical evidence to suggest it would be effective.

The incorporation of what Obama called “data systems” was also a massive opportunity for the private sector. The director of Race to the Top, Joanne Weiss, was previously COO of NewSchools Venture Fund, who received millions of dollars from Eli Broad and Bill Gates and their respective foundations to assist charter management organizations. By 2010, the Gates Foundation had given at least $650 million to projects that advanced charter schools, testing, and what it called, quote, “teacher effectiveness.” This “data systems” push would open the door for the Michael and Susan Dell Foundation and the Chan Zuckerberg Initiative as well.

Relatedly, the same year, in 2010, the Gates Foundation and Scholastic Corporation released a survey claiming to reflect the opinions of 40,000 public school teachers on how to “improve” US schools — a cause the foundation called “education reform.” The survey found that, quote, “Teachers aren’t opposed to standardized tests as one way to measure student performance,” end quote, and proposed that one way to improve schools would be to, quote, “Innovate to Reach Today’s Students,” end quote. The Washington Post eagerly reported on the study in December 2010 under the headline, “Gates foundation research finds test score growth is sign of a good teacher.” And here’s an excerpt, quote:

While debate rages in the education world about how to measure effective teaching — or whether it is even possible to do so — research funded by a prominent advocate of data-driven analysis has found that growth in annual student test scores is a reliable sign of a good teacher.

End quote.

Adam: But this study, as is often the case with polls, was full of misleading questions which were clearly set out to achieve certain outcomes. For example, one of the questions from the poll read, quote, “How much of an impact do you believe the following efforts would have on improving student academic achievement?” And the options were:

“Clearer academic standards for students.”

“The establishment of common standards across all states.”

“Tougher academic standards for students.”

“Fewer academic standards for students.”

“Fewer academic standards for students” doesn’t sound very good, right? “Clearer academic standards for students,” on the other hand does, which is technically a category under which standardized testing and common core standards which Gates helped pay for, fall under. The answer choices are designed to steer respondents in a particular direction, specifically that which favors a testing heavy charterized, privatized, and union free education model.

Here’s another example, question: “How important do you think each of these items is in retaining good teachers?” The options were:

“Supportive leadership.”

“Time for teachers to collaborate.”

“Access to high-quality curriculum and teaching resources.”

“Clean and safe building conditions.”

“Professional development that is relevant to personal and school goals.”

“Higher salaries.”

“Collegial work environment.”

“Opportunities for alternate careers in the classroom, like mentor teaching.”

“Pay tied to teachers’ performance.”

Now, one might be tempted to ask why teachers should have to rate things like decent pay, supportive leadership, clean and safe buildings, and high-quality curriculum. It should be a given that all of these things would be provided and necessary in public education, but they wanted to reinforce the Gates’ agenda. The Gates Foundation would publish subsequent studies, which would reliably receive uncritical coverage in mainline press. The Washington Post in January of 2013, “Gates Foundation study: We’ve figured out what makes a good teacher.” Denver Post: “Denver schools, Gates foundation identify what makes effective teacher.”

Now for years, up until probably 2015, 2016, Bill Gates’ sudden and mysterious interest in education was seen as largely this altruistic, philanthropic effort. It was not seen as someone who had an ideological and financial agenda into privatizing education. So there wasn’t a lot of skepticism on reporting what his foundation found. The Gates/Obama charterization model was widely influential. Over a period of just under 20 years, there was a significant increase in privately managed charter schools. During the 2000–2001 school year, there were 1,993 charter schools in the United States. By the 2019–2020 school year, there were 7,547 charter schools throughout the United States. So clearly, the obsession over data and standardization in education is not necessarily by definition a right-wing agenda. In theory, these things are kind of politically neutral. But how they manifested was to set up a situation whereby the teachers unions and school boards, which are democratically elected, are viewed as being in the way of accountability, and the way we get accountability is to measure things, and how we measure those things is informed by political preferences and ideological priors. But it sounds bad to say we’re doing a privatized takeover of your education system, because people genuinely like their school, you know, somewhere between 75 and 80 percent of people if you poll them like their school. Now, if you ask them if they think public education is good, in general, the numbers are 28 to 30 percent. But if you ask them if their particular school is good, people actually really, really liked their public schools.

Nima: Yeah, I like my Congress member, but Congress in general sucks.

Adam: Right. So what you have to do is you have to start measuring things based on your own preferences and then when you measure those things, again, with a kind of loaded framework, then you start to create systems and models whereby if they don’t meet a certain threshold, they’re deemed failing, and then once things are deemed failing, they therefore are susceptible to privatization, which, of course, was the goal. The more cartoonish version of this is what happened in New Orleans, which we talked about in great deal on Episode 1, where they arbitrarily changed what was a failing school from 60 to 86, just so they could justify privatizing the entire New Orleans school system, which now to this day is almost entirely a charter. Because you can kind of just make the data whatever you want to make it. We don’t say that to be obviously nihilistic or anti-intellectual, that is not to say that data can’t be a somewhat objective metric for certain things if given certain context, but as a political tool, this idea that, ‘Oh, we’re just following the data,’ it’s just an easier pill to swallow than, ‘Oh, yeah, we totally want to fire the school board and privatize education and bust all the unions,’ because that doesn’t really sound good.

Nima: Well, also, because, you know, standardized testing doesn’t fall from the heavens, right? It’s created by people. There is context.

Adam: And a lot of it was created by the Gates Foundation, because they’re the ones who were underwriting much of the Race to the Top programs for the states. They were giving grants to states based on their participation in Race to the Top, and, again, Bill Gates is just some guy, he’s not someone anyone elected. He wasn’t appointed by the president. He has no democratic input into his authority. He just, for about 10, 15 years from, you know, the mid-2000s to the end of the Obama era, he kind of just ran education because he had so much money, and, again, he’s just some guy. He has no experience in education, no masters in pedagogy. He’s just a guy.

Nima: But he’s a guy, Adam, who really liked the data.

Adam: Right. It’s not ideological, he just likes data, Nima.

Nima: Exactly. He just, you know, Bill Gates likes data. So, in this climate of Obama-era neoliberal tropes like this Privatization for the Good of Education, data journalism as a practice continued to accelerate. The New York Times introduced its, quote, “politics, policy, and economics” site The Upshot in 2014. The Upshot claims its aim is, quote, “to help readers better navigate the news using data, graphics and technology,” end quote. In 2016, Bloomberg launched its podcast “Follow the data,” which is effectively a PR series for Michael Bloomberg’s own foundation, Bloomberg Philanthropies. Interestingly, while the series focuses heavily on “public health,” the data it cites never seems to include the benefits and popularity of — of, I don’t know — a single-payer healthcare system.

Bloomberg Philanthropies CEO Patti Harris and Bloomberg Associates’ Katherine Oliver record an episode of “Follow the Data.”

“Data-driven” policing began to surface around the late 1990s and early 2000s, roughly the same time public schools were rolling out “data-driven” test-based instruction. The term was often meant to convey a successful tool for crime reduction, but it denoted “predictive policing” — policing that further targeted and surveilled poor Black and brown communities under the guise of identifying, quote, “crime hot spots,” end quote. The term “predictive policing” was often used to describe, and in fact sanitize, the policies of Bill Bratton, the former NYPD commissioner and LAPD chief infamous for expanding the use of broken windows policing and stop-and-frisk.

Adam: The New York Times Editorial Board in August of 2016 wrote, quote, “Moving Past ‘Broken Windows’ Policing.” The piece makes a few reformist overtures, gestures but ultimately defends Bratton taking his stated goals at face value, quote:

“Mr. Bratton is known for pioneering aggressive, data-driven policing that targets hot spots of crime, seeks to contain mayhem before it spreads, to catch dangerous criminals and scofflaws for minor violations, to get guns off the street. Few would argue that those goals, in the abstract, are anything but good.”

Nima: I mean, hey, especially when you’re tracking down scofflaws Adam in 2016.

Adam: So again, you take this idea of like, ‘Oh, well, there’s a lot of liberal backlash to stop-and-frisk and the unconstitutional detaining of Black and brown people to frisk them. So around this time, you started to see post Black Lives Matter and increasing obsession with this idea of precision policing, data driven policing, targeted. There’s a reason why they use the same language when they do cruise missile strikes. It’s targeted, it’s smart, it’s sort of very sanitized. NPR continued this trend, publishing a June 2018 an article headlined, quote, “How Data Analysis Is Driving Policing,” that celebrated the LAPD’s sleek new data search system, sold by the quote-unquote “somewhat controversial” Palantir. With a few obligatory quotes from activists opposing the system buried halfway down the article, it mostly relied on the Deputy Chief’s word that data-driven policing had brought crime down.

Nima: Now, police departments and media manipulate data in order to make reactionary points that they want to make in the first place, right? They just need the data to back themselves up so that liberals can be like, ‘Oh, it’s the data.’ One of the more recent examples of this is the narrative about how crime was surging after the onset of the pandemic. Founder of Civil Rights’ Corpse, and former guest of the show, Alec Karakatsanis, has detailed how the bulk of reporting on so-called “crime surges” uses low raw numbers so that increases can appear artificially high. As Karakatsanis has noted, an increase of 10 shootings to 12 shootings over a certain time period is then reported as a 20 percent increase which sounds incredibly ominous even though the raw number is two. Now, furthermore, the overall crime rate in 2020 was down, however media emphasized rising rates of killings, conveying a sense of a “surge” or “wave” of crime sweeping the nation. This framing continued throughout last year, 2021, and into this year.

Adam: Yeah, so I wrote about this for The Appeal in 2019 and 2020 several times, specifically New York Times writer Ali Watkins was obsessed with creating these, clearly spoon fed to her by the police or public relations to the police, these totally arbitrary, you know — like ESPN does those factoids, where they’re like, you know, he’s only the third pitcher since 1970 to have 11 strikeouts, less than two walks, on a Tuesday.

Nima: When the temperature was beneath 42 degrees.

Adam: Yeah, and then you look at that you’re like, I guess on the surface that sounds impressive but clearly, they just had the data and then they went back and reverse engineered something, some meaningful stat, which is what people at ESPN do, they do a quite a good job at it, because you know, you have to make this shit seem interesting, right? I mean, otherwise, it’s just another fucking meaningless baseball game in July. Who cares? Ali Watkins was obsessed with doing this. In 2019, in early April, she wrote a report about the quote-unquote “murder spike” in Brooklyn, telling the reader that, “As of March 24, the borough had recorded 28 homicides so far this year, compared with 17 in the same period last year, a 64 percent increase.” A 64 percent increase, which is in the social media context that was reported by other media, is kind of this big number. But then the piece hedges on whether this is even meaningful with their To Be Sure section, acknowledging that, quote:

“Crime rates rise and fall periodically and it is too early to tell if the increase in killings foretells a new crime wave that would challenge the sense of security that has become part of the city’s identity. Over time the increase may be offset by quieter periods, and trends that appear worrisome in March often level off by August.”

So by August, five months later, the aggregate murder rate in Brooklyn it was actually down 12 percent, and then cut to August, Watkins writes another piece, she moves the goalposts. She now says that quote-unquote “shootings” are up 10 percent across northern Brooklyn for that year in some neighborhoods, like East New York and Crown Heights, they have doubled. All of these stats are true enough, but they omit important contexts such as that shootings did indeed double in these neighborhoods but the total murder rate in North Brooklyn was virtually unchanged from 39 to 38. Shootings in South Brooklyn were down 26 percent and murders in South Brooklyn have been down 32 percent since 2018. So this is kind of a good example, and she did this all throughout 2019, 2020 — this is before murder rates sort of actually increased during COVID — where you have this, I can look at a very detailed map of New York in the CompStat data that NYPD releases, and basically you just go where the little red line goes up, and then you go interview three people in the neighborhood and you say crime has increased in northern Brooklyn. Well, if crime is down in 10 other precincts, where’s that story? So this is another sort of example of how the sort of crime surges, crime spikes stories are almost all bullshit, because you can pretty much take anything you want and make it look like a crime spike, and now again, I know that murder rates, crime is actually down, but murder rates are actually up quite significantly in most major metropolitan areas, there are lots of explanations for that, and that is objectively true. This was not true, however, when this story was written, and it was clear for many months that there was an effort to basically scare rich liberals who read The New York Times, and if you’re going to sort of keep moving the goalposts, again, her article goes from Brooklyn to North Brooklyn to two precincts in Brooklyn. I mean, that’s not a very meaningful stat, and of course, it was only over a couple of weeks, and The New York Daily News and The New York Times, and of course, The New York Post, are obsessed with this. They do this all the time, and they especially did this when there was a token handful of police departments that quote-unquote, “defunded” the police, which is to say they didn’t increase the funding. In Minneapolis, especially, they would go in and target every little thing, this crime, that crime, to create the illusion that there was this warrior-like despotism of crime. And then, of course, in 2022, when all the police departments in New York, Chicago and LA saw their police budgets balloon by five to fifteen percent, and then crime increased, still, for, again, reasons that have very little to do with the actual budgets of police, with the discretion of prosecutors, we didn’t really get all the panic, selective crime stories, because it didn’t fit the narrative, and so, again, it’s not as if the data is false, it’s just that if you have a lot of data, you can sort of paint whatever narrative you want to paint. Because there’s a political context here, which is at the time Brooklyn had someone who was quote-unquote a “progressive prosecutor,” which is clearly why they were focusing on Brooklyn, they were trying to demagogue Eric Gonzalez, who’s not even a progressive, but they were trying to make them look bad, and clearly, this was an NYPD curated story. So again, this kind of idea of data as this value neutral thing flies out the window when you spend 10 minutes interrogating this kind of reporting, because what data are we going to highlight and why are we zooming in on this fucking obscure precinct?

Nima: Well, yeah, and so this very thing was actually replicated across the country throughout 2021. Not only Fox 5 NY saying, “NYPD top cop blames bail reform for surge in crime,” but you also have articles like this from USA Today in April of 2021, “Why violent crime surged after police across America retreated.” You have May 2021 in The New Republic, “Can the Politics of Police Reform Survive the Crime Rates of Our Pandemic Year?” The Denver Gazette said, “Denver police chief says early release has contributed to crime surge.” PBS Channel 13 said, “Is criminal justice reform to blame for the rise of crime in NYC?” The Atlantic said, “Why America’s Great Crime Decline Is Over.” The Wall Street Journal’s opinion page said, “Crime Is Up and Democrats Are Scrambling.” The Heritage Foundation, unsurprisingly, said, “FBI Statistics Show a 30% Increase in Murder in 2020. More Evidence That Defunding Police Wasn’t a Good Idea.” While the MacIver Institute in September of 2021 said, “New FBI Data Proves ‘Defund the Police’ was a Deadly Mistake.” You had CNN saying, “How US cities are preparing for a potentially bloody summer of gun violence,” and also saying, “Defund the police encounters resistance as violent crime spikes.” The Economist says, “As violent crime leaps, liberal cities rethink cutting police budgets.” The Wall Street Journal, back again, said, “Police Wrestle With Surge in Crime in U.S. Cities Amid Defunding Efforts.” NPR said, “Rising Violent Crime Is Likely To Present A Political Challenge For Democrats In 2022.” Bloomberg, “How Bail Reform, Crime Surge Mix in an Angry Debate,” and Newsweek saying, “New York City Is Cutting Police Budgets As Murders, Shootings Surge.”

Adam: To be clear, they never actually cut the budget, they moved the money around to a different department, the budget never decreased under de Blasio. Now, we saw these headlines over and over and over again in 2021, they would hone in on the handful of obscure cities that may have cut it a little bit, talked about how they’re going to rethink it. Now, we’re now almost six months into 2022 and I wrote about this, on my Substack, we have not seen any similar pattern of articles at all about whether or not we need to rethink refunding the police or funding the police in lieu of the fact that crime in most of these major cities is up even more than it was in 2021, and their reason is because there are many other factors which are contributing to this with zero to do with the actual police budgets, and so this is again, you have this torrent of stories about defund the police in 2020 and 2021, second half of 2020 and 2021, nonstop defund the police, never fucking shut up about it, and yet we get no such corresponding stories when crime increases in 2022.

Nima: No, of course not. Police funded for 100 years yet crime still exists. Maybe it’s time to rethink funding police.

Adam: I mean, this is why the whole data driven policing is so silly, because if you looked, if you were data driven, you would say that the US has the highest murder rate of developed nations quote-unquote, “developed nations,” “wealthy nations,” it’s not even close. Data from 2010 showed that the average American was 25 times more likely to be shot by a handgun than any other country. Consistently the highest murder rate even after the murder rate fell precipitously since the 1990s and yet we also have 25 percent of the world’s incarcerated population, despite having only 5 percent of the world’s population, we have by far five times the mean prison population, we have the most incarcerated, we’re the most incarcerated country, and it’s not even close. Certain states have numbers that are higher than that, that are 8, 9 times the national average, states like Mississippi, Oklahoma. So you would look at that if you were data driven, and you’d say, ‘Man, this doesn’t seem to be working,’ we still have an extremely high murder rate compared to other countries, and in some cases we have a higher violent crime rate too than comparable countries, and yet, we also lock up the most people. So the data would suggest that maybe what we’re doing isn’t working, but it never works that way. Carceralism is an ideology that can never fail, it can only be failed.

Nima: That can never be investigated by the numbers.

Adam: Yeah, and so the whole numbers thing is like, okay, clearly you just want more police funding, because that’s the cheapest, easiest way to deal with surplus populations and real estate people want them and rich liberals want them and Republicans of course want them so we’re just going to have more police, and the reporters just goes out and finds data that conforms to that. So stop acting like at some fucking, like you’re some bespectacled sage-like monk who’s sitting on a mountain seeking wisdom. Clearly you have a fucking charge. You have a top down editorial decision to go find scary fucking crime numbers. So let’s just say that’s what you’re doing.

Nima: “Data-driven” policing was not the only media story during the pandemic. Around the same time, charter school donor-backed “data-driven” surfaced in attacks on public education. One of their chief ambassadors has been Emily Oster, a Brown University economist who’s been on a crusade to get children back into classrooms regardless of what the public health data says, yet, using data to make her point. Oster has been platformed heavily. She’s been cited and interviewed in hundreds of articles, she’s appeared on TV and radio and leading media about schools and the COVID-19 pandemic and has written over a dozen pieces for the Atlantic, Washington Post, and elsewhere. Among them headlines like these, quote, “Opening schools might be safer than you think,” end quote. That was written in May of 2020. The same month, she wrote this, “The ‘Just Stay Home’ Message Will Backfire.” She also wrote in August of 2020, “How the media has us thinking all wrong about the coronavirus.”

Emily Oster (Cody O’Loughlin / The Guardian)

Adam: Oster’s work has influenced US public health policy quite a bit, or has provided the pretextual narrative for pre-existing needs of capital and our political class. It’s been cited favorably by Florida’s ultra-right-wing Governor Ron DeSantis, Rhode Island state officials, and institutional scientific authorities like the CDC and the EU’s European Center for Disease Control. In 2020, Oster created a “data dashboard” to track K-12 schools’ responses to the pandemic. The project attracted funding from mysteriously a lot of the pro-charter education privatization organizations. The Chan-Zuckerberg Initiative, the Walton Family Foundation, Arnold Ventures, and Emergent Ventures, a project funded by Peter Thiel and the libertarian Koch funded Mercatus Center.

Unsurprisingly, Oster’s data is somewhat spurious. Oster authored an October 2020 piece for the Atlantic, quite brazenly headlined, “Schools Aren’t Super-Spreaders,” that argued COVID transmission rates in schools were too low to warrant school closures and remote classes. So, you know, we couldn’t do a quote-unquote “data driven” episode without talking about Emily Oster. In many ways, she’s the, which we’ll talk about with our guests, she’s the successor of John Stossel and has many similar funders, which is to say, the basic premise of most of her work is that liberals are not rational, they don’t understand the real concept of risk trade offs, and that we need to think like economists and make rational decisions in that a bunch of bedwetting liberals and tort lawyers need to look at the data and numbers and we’re not being rational. We’re being too scared. We don’t manage our fears appropriately like a true economist does.

Nima: Right. David Leonhardt of The New York Times does a very similar thing.

Adam: He does a very similar shticks. So we’re going to talk about that with our guests who wrote probably the definitive takedown of Emily Oster and the whole genre of economic thinking and data driven analysis.

Nima: We are now going to be joined by our guests Abigail Cartus and Justin Feldman, co authors of the Protean Magazine article, “Motivated Reasoning: Emily Oster’s COVID Narratives and the Attack on Public Education,” which was published in March of 2022. Abigail Cartus is an epidemiologist at Brown University where she focuses on perinatal health and overdose prevention in her work at The People, Place & Health Collective, a Brown School of Public Health research laboratory. And Justin Feldman, also an epidemiologist, and a Health and Human Rights Fellow at the Harvard FXB Center for Health and Human Rights. Abby and Justin will join us in just a moment. Stay with us.


Nima: We are joined now by Abby Cartus and Justin Feldman. We are thrilled to have you with us today on Citations Needed. Welcome.

Abigail Cartus: Thank you.

Justin Feldman: Thanks so much for having me.

Adam: Your March 22 article, “Motivated Reasoning: Emily Oster’s COVID Narratives and the Attack on Public Education,” places Oster squarely in a kind of Freakonomics current of pop media. You argue that she, along with others, is the successor to not only a particular brand of economic style of reasoning, as you put it, but also a libertarian or neoliberal ideology dressed up for the NPR set — for liberals basically as her primary market. Since the style of writing is something we’ve spent a great deal of in this, I guess, media posture for want of a better term, we spent a great deal in the intro talking about, I kind of want to begin here. What is the brand of economic style of reasoning? Why do you think it’s been embraced by both the far right, think tank world and the sort of more centrist media and think tank world?

Justin Feldman: I think we got very lucky in the timing of when we wrote this piece for a few reasons, but the reason I’ll highlight here is that we knew there was something about Emily Oster that reminded us of people like Nate Silver, Matt Yglesias, people who come up often on your show, I imagine, Twitter, people on the left get pretty annoyed by them. So we wanted to frame theoretically what’s going on there with the through line between these people and through Freakonomics? And what was lucky was that this sociologist named Elizabeth Popp Berman came out with a book a week after we wrote this, but essays beforehand, laying out her argument, and she described something called the “economic style of reasoning,” which is not identical with economics as a field. It’s a way of seeing the world making decisions, making arguments about policy, that prioritizes efficiency, and I would say, shores up the power of the ruling class in most cases, and kind of waves away concerns about social injustice, and Popp Berman argues that this style of reasoning actually predated neoliberalism and came up through the Great Society programs and sort of liberals and Democrats imposing order on this newly expanding set of government programs. It then took on a life of its own and spread, as you said, to centrists and right-wingers, and it’s very powerful, because it allows the person wielding it or the institution wielding it to obscure their interests, and not make debates about questions of values or ethics, but of basically what they’re describing as objective constraints on what we can do given the resources they claim we have or given the particular costs that a policy move will incur society as a whole, never really understood as particular groups in that society incurring costs.

Justin Feldman

Adam: Yeah, because it’s a superficially attractive posture. I think it’s very attractive to a lot of people, specifically people in technocratic fields, people who maybe are software engineers, or economists or corporate types. It’s something we’ve criticized a lot on the show. But we haven’t actually focused on this kind of data driven punditry aspect of it, which I think is quite pernicious, this idea that somehow these people approach these problems as a blank slate, and they just call balls and strikes, and they go where the data takes them. Obviously, for a sort of jaded idealogue like us, this obviously reeks of bullshit. This is a branding exercise more than anything, because it’s a way of obscuring what is more often than not very glib or very cruel policy or policy that talks about trade offs or tough decisions or limited resources, it evokes very often a sense of false austerity, that’s kind of just asserted. From your experience, it is kind of a romantic vision of the world, this post ideological vision of the world, one that I think many people thought had fallen out of favor, but seems to be making a bit of a comeback of late.

Abigail Cartus: Yeah, to me, the really salient aspect of the economic style of reasoning, especially as it has sort of pervaded the field of public health during the pandemic, specifically, is that it allows for the replacement of values and value based arguments with highly technical methods and arguments over, you know, statistical methods, and I think that part of that also is the replacement of the work of politics with, you know, the work of policy or policy making, which is, I think, to most people’s minds, you know, policymaking is a step removed from the democratic process. We’ve been reading throughout the pandemic about all of these private consulting companies that are, you know, like McKinsey is like fucking up vaccine distribution and stuff like that. I think that the economic style of reasoning is definitely a value orientation but the sheen of mathematical rigor and the sheen of kind of numbers-based objectivity allows for some kind of mystification about that. I’ve been reading a lot about strategic governance initiatives and the role of private consultants in public health, which is why I’m kind of thinking about this, but it’s interesting that you brought up this austerity, this is always a cover for deregulatory moves, and it’s always justified by crisis rhetoric, ‘We have to do this because we have no time to do anything else, we just have to take a look at the data and do what the data are telling us,’ and you know, we can get into this more later on, but data don’t speak for themselves, and that kind of represents a fundamental misunderstanding about what data are and the role of uncertainty in scientific knowledge generation. But I think that is the key to the whole thing, and I think that this speaks to why it’s been embraced by the right-wing, centrists and liberals is that it’s this collective fiction that we can achieve optimal outcomes or reach optimal decisions through a process that doesn’t involve conflict, and through a process that doesn’t involve explicit articulation of values.

Abigail Cartus

Nima: Let’s talk about this broader media trope of focusing on quote-unquote “the data.” There’s just steely-eyed facts here, we’re going to put it in a dashboard, we’re going to ban it, we’re going to have a data dashboard that’s constantly updated, we have all the data analysts on it, we have scientists checking this. Let’s talk about this, not only dashboardization, but also this data-driven kind of reporting and punditry, that always, shockingly, winds up aligning with the interests of those pushing more economic modes of reasoning, right, rich people, those who benefit from current economic setups. In your excellent piece, Justin and Abby, you write this quote:

This brings us to the cornerstone of the economic style of reasoning: the data. The evangelists for the economic style of reasoning exhort their audience to “follow the data” — the alternative being, it is implied, to capitulate to the irrational demons of fear and anxiety.

End quote.

So as we’ve been discussing, far right think tanks love to name things, “reason,” “rationality,” “rational discourse,” right? The rational reason institute of sane conversation funded by the Koch brothers, sober economic choices fellow at the reasoned and rational discourse society of America. So the implication as you’ve laid out, being that leftist socialists, even liberals, are just a bunch of super emotional, and there are all kinds of gendered issues here, feminizing is too emotional, you just have to look at the hard data, it’s not based on empiricism. Whereas economists are post ideology, they just look at the data. Leftists and others are just, you know, running around chickens with their heads cut off. Talk to us about this kind of framing, why it is so widespread, why it is so appealing, and why it is appealing not only to rich people and right-wing ideologues but why does it capture this individualistic liberal imagination?

Justin Feldman: Yeah, I think there is this liberal value of trusting science, trusting expertise, and some of what we’re seeing in Emily Oster’s work, in the writing by David Leonhardt in New York Times Daily Newsletter, these are people who are basically laundering right-wing arguments and trying to make them appealing to liberals, I imagine mostly professional class liberals, but they are not going to be approaching it from an explicit ideological standpoint that makes it clear like, ‘Oh, we value small government, we don’t think workers should have power.’ Instead, it’s going to be, ‘Well, the risks posed by this pandemic are actually small or the costs imposed by the social welfare programs are too big, we simply can’t do it, the balance doesn’t doesn’t make sense when you add the numbers together.’ They’re often not actually doing any kind of data analysis. Certainly that is also a thing, there is data journalism of varying quality, but often it’s just a rhetorical allusion to the existence of different quantities of risk, and that’s something, you know, as epidemiologists, we are seeing again and again, about every activity supposedly being of minimal risk, and it makes you wonder how we got 2 million deaths in the US.

Abigail Cartus: Yeah, and if I can just add an example from this, our piece was kind of about Emily Oster’s COVID advocacy, which mostly has to do with schools and learning, and the major lightning rod question was, you know, should school buildings be open during times of very high COVID transmission? To Justin’s point, I think it’s a salient example, you know, a lot of the actual epidemiologic research on school reopening or transmission of COVID in schools was really just designed around the kinds of questions that economists or epidemiologists or statisticians could answer with available information. So I think what we saw was a lot of studies that were correlating the case rate in schools with the case rate in the community and, you know, looking at the lower rate in schools to say, ‘Well, look, it’s a little lower so it must be, you know, the unspoken,’ well, sometimes it was spoken, the implication being that transmission isn’t happening in schools, schools are safe. But I think, really an analysis that just correlates case rates in versus out of school is not the same thing as a study that actually tries to answer the question that we’re really interested in answering, which is are kids getting COVID in school? Are they spreading it to other kids or to teachers? Is transmission happening in schools? The answer to that question is obviously, yes, but, you know, a scientific study that shows that is not really very convenient. So the use of data and data analysis to steer the discourse towards asking more convenient questions is something that we really noticed a lot with the schools debate, and, you know, there are studies that show that transmission is happening in schools, but especially in the US discourse, those studies, they have to meet a higher evidentiary threshold, in my experience, because they’re formulated in a very inconvenient way.

Nima: Right, then just some random assertion in the pages of The Atlantic magazine.

Adam: To be clear, for those who are listening may be confused, y’all are both epidemiologists, it’s not as if empiricism and data aren’t important, I want to be very clear here, the issue is that the institutional biases and the sort of pop media function of quote-unquote “data” is logically going to serve a political ends, and it’s incredibly insulting to everyone’s intelligence to act like that isn’t what’s going on here. It’s sort of like, very often in a nonprofit or an educational establishment or even in corporations, if you want to push out a CEO, nobody really wants to do it so what they oftentimes do is they’ll sort of, ‘We’re going to hire a third party consulting firm to tell us how we can be more efficient,’ and they come back with an answer that’s like, ‘Oh, you got to fire the CEO,’ because nobody sort of wants to be the one who’s the executioner, and it’s the way you launder responsibility and off put moral responsibility, and so much of this data driven framing is basically like, ‘We need to do this really cruel and mean thing, but nobody really wants to take credit so we’re going to take this moral burden off of policymakers and off of electeds, and mayors and school officials and governors, and we’re going to put it on the data.’ It’s very similar to the function of the crusades in the Middle Ages that said, ‘If you do this, you’ll sort of be free of sin.’ ‘It’s the data, it’s no one’s fault, it’s kind of the way it is,’ and one of the things when you talk about Oster, and this is what you guys did, is you you sort of broke the taboo about talking about funding sources, and how those can lead to conflicts, if not corruption, in terms of charter school, billionaire Arnold Foundation, Koch brothers, well, Koch brother, and this caused a lot of outrage from that circle, because I really think it’s capital “T,” capital “T,” The Taboo, you’re sort of really not supposed to talk about that. People kind of broadly understand that money has influence until you get into the think tank academic world, and then it’s nonstop pearl clutching, and I think that was a huge point of criticism of your piece, I thought unfairly, but there was so much institutional pressure to basically say COVID is over. People sometimes will claim there was an opposite incentive to be more panicked, but that was never real. That was true for maybe like the beginning for five minutes. But mostly, it was like, holy shit. It’s austerity time. It’s inflation time. It’s time for Grandma to die and for Junior to go back to school, and, you know, what happens happens, basically, especially with Omicron. And then we’re all supposed to sort of assume that oh, actually, like, what are the odds, the data went to the part that the Chamber of Commerce and the Republican Party were pushing six months prior and the Democrats now due to inflation have absolutely no choice but to embrace. The capital classes very clearly have this agenda? What are the, out of all the gin joints in all the towns in all the world, what are the odds that the data happens to align with that? I mean, you have to be a toddler to not see there being biases built into the system. So if you can comment on those biases, you talked about the conflicts, and how that poked the hornet’s nest of that world a little bit.

Justin Feldman: I think one of the reasons we wrote this was because each of us have gotten pushback in the past for identifying and criticizing funding sources in the context of the pandemic, and what happens when you raise that point, at least within academia, is that people accuse you of accusing the researcher of essentially making up, fabricating data at the behest of their funder, and that very well may happen. Certainly there are historical examples where that did happen. But that’s not what we are necessarily arguing. So what we wrote the piece in part to do was to answer the question, why does it matter that Emily Oster is taking money from people like John Arnold and Peter Thiel and the Walton Family Foundation?

Adam: Right.

Justin Feldman: Why does that matter, even if she is doing what she believes to be rigorous and honest research? And we still got some pushback about that, I mean, the answer to why it matters is because there is a convergence of interests. These organizations have sort of this ecosystem around pushing what they call education reform, which is school privatization, and charterization, breaking the backs of teachers unions, and they, particularly the Walton Family Foundation, cited Emily Oster’s work, included her at their events, including an event by their other grantee called Bellwether Partners during the Chicago teachers’ strike, where she was there with the person who was then the Chicago Public School CEO, talking essentially in favor of the pro-management, anti-union position. So Oster served as a cog in this machine that existed before the pandemic, and also produced research that we show, based on the work of others who have examined it and our own critiques, is quite flawed, despite some of it being published in peer reviewed journals, though not necessarily through the standard process, because one of these papers only took two weeks to be accepted, which is not normal for most journals or for this journal.

Nima: So I’d actually love to shift gears a tiny bit and talk about some larger threads that are woven through your entire piece, and through this study of, you know, Oster’s work, but also this type of reliance on data to do a thing. It’s not reliance on data for data’s sake, to report on research, it’s to do a thing and it is ideologically driven, as we’ve been discussing. I’d love to hear your thoughts on how you kind of discuss both the precautionary principle versus this gospel of personal choice, how those two modes of thinking are inherently in conflict. Let’s discuss what they are, and how they inform this kind of, ‘Hey, we’re just looking at the facts,’ but actually, it’s doing this other thing, which is, you know, all about personal responsibility, and this veneer of control, which is so attractive to not only an American audience, but as you’ve really identified, Oster and others audiences as predominantly white, affluent, cisgendered, et cetera.

Adam: Well, people who read The Atlantic.

Nima: Yeah, exactly. Let’s unpack why people who listen to The Daily really love this gospel of personal choice, and reject, or are open to rejecting the precautionary principle, but I’d love for you to talk that out a bit.

Abigail Cartus: The precautionary principle is kind of like a heuristic for making decisions under conditions of great uncertainty and where there is potential for harm. So the precautionary principle is used, traditionally, quite a bit in public health, and it would argue in favor of, for example, closing school buildings in March 2020, when almost nothing is known about COVID, and how it transmits, because COVID could be very dangerous, there could be significant substantial harms associated with allowing COVID transmission to happen unchecked, and so the precautionary principle militates in favor of exercising, or erring on the side of caution, if there’s potential for harm, but you don’t really know what it is. On the other hand, this kind of personal risk, personal choice thinking is extremely common, I think, in American culture, generally, but it’s also extremely common in, I would say, the biomedical fields and biomedical research, and particularly in the practice of medicine, where I feel like the last 10, 15 years has been this push towards personalized medicine like where I’ve even heard the term precision medicine, precision public health, which doesn’t make any sense.

Nima: That sounds like, you know, surgical bombing strike?

Abigail Cartus: Yeah, yeah, exactly. There’s something interesting to me about this risk orientation, because there are two distinct senses of the word “risk” that I think get conflated, especially in the discourse around this. One sense of the word “risk,” or, you know, the idea of risk, is what I would call maybe, like, empirical risk, right? So you can look at a data set of 100,000 people, and see how many of them got COVID, you can come up with a proportion, and that is like, essentially, a risk estimate. There’s another sense of the word risk, which is like the probability, the stochastic probability of an unknown future event, and I feel like David Leonhardt really does this all the time, right, just collapsing these two things together, and I think that’s really not appropriate to do. But this idea of being able to calculate your personal risk, I think, appeals to the American psyche, maybe for obvious reasons, but it also has a very real basis in the very statistical methods that we use to understand things like the pandemic. So a lot of the statistical methods that we use require that you treat individuals as independent from one another, right, that you break things down to the smallest possible unit, it’s almost I don’t, you know, I’m not a philosopher of science, but it’s almost a little bit positivist, and I think that that really appeals to people, you know, there have been all sorts of risk calculators, I think Emily Oster herself came up with a risk calculator where you can type some inputs into an excel sheet and come up with, you know, a number that’s supposed to represent your risk of either contracting COVID or having a severe outcome from COVID, and that just makes not a lot of sense because, again, your risk of contracting COVID is like either zero or one.

Justin Feldman: For people who are in power and don’t want new policies in place, for instance, like paid sick leave that would keep workers home rather than in the workplace infecting people, they want to do a couple of different things, they want to downplay the risk of COVID, and they want to shift all the risk on to individuals. If you only wear a mask, if you only got vaccinated, if you only got boosted, then everything would be okay. Which I think Americans being as individualistic, and vindictive, as they often can be, often are on board with because they’ve been hearing arguments like this, we’ve been hearing it our whole lives.

Adam: Yeah, another Koch funded pundit. This was a stick that John Stossel did for years on 20/20 and elsewhere, it’s sort of America’s too risk averse, here’s why risk is off, you know, again he’s a fellow at the Cato Institute and sort of talks about how everyone’s kind of an irrational, highly feminized alternative. And I want to dig in, if you don’t mind, a little bit to the Oster data, because I think some people listening would say, ‘Oh, you’re just a bunch of a kid Ivy League liberals who want us all to wrap ourselves in bubble wrap forever,’ the extreme COVID scold end of the spectrum, and obviously, there are going to be trade offs, right, there are social and deleterious effects to not going to school. I mean, I think everyone sort of broadly agrees with that. Not socializing, not going to football games, like these things have social value. It is true that on a basic level, what they say is superficially true, you do have to make trade offs, you have to sort of do a cost benefit analysis, and come up with that. I think where we call bullshit is that they’ve gone to this extreme of every individual atomized, Randian hero, this sort of protagonist thrown narrative has to make these highly individualized choice and basically forfeited on any concept of public health, that there’s no public health anymore, it’s basically the individual’s job to go get an N95 mask and to do their own thing, and I think a lot of public health experts have been, forgive me if I’m not summarizing this correctly, but I’ve been very disillusioned by that, that we’ve sort of given up on any concept because this pandemic doesn’t really seem to be going anywhere, there’ll be new variations. We now kind of foreclosed on any sense of public health, and I think the sophistry of the data and the risk aversion, everybody’s too risk averse element is, is appealing to a certain kind of person, because everyone’s frustrated, right? Universally, people are frustrated, and so there’s a market for someone to come along and say, look at this graph, and look at this graph. ‘This is The Atlantic, you’re not crazy. Public health officials are a bunch of fucking weenies and we can go back to normal.’ That’s very attractive, right? Nobody wants to be the one who has to tell you to eat your vegetables. If you can, can you please kind of go into the actual, and your article does a really good job doing this, can you talk about the actual empirical or data or science criticism of the Oster narrative, what your objections are to it and where there’s flaws in this narrative, especially in the sort of last winter and last spring?

Justin Feldman: Sure. What’s interesting here is that there was never a debate over whether or not to open the indoor seating area at Arby’s. That was kind of taken as a given that it was going to be open and there was not going to be any public debate or scientific debate about what kind of risk that posed because there’s no children involved, necessarily, you can have your children there or not, and probably more importantly, their workers have no power and are low wage, whereas teachers are unionized, work for public institutions that are supposed to be publicly accountable. So we were in a unique place with schools, where there was a lot of debate that took place in the usual spheres of punditry, but also on the pages of scientific journals, and inspired many studies, and this is where we find Emily Oster square in the middle, and she’s put out a couple of studies. So there were arguments over time, under what conditions is it safe for children to be in school, safe for children, safe for their caretakers, save for teachers, and each step of the way, where we saw shifts from remote learning to hybrid, hybrid to in person, in person with testing and quarantine and masks, to gradually peeling away each of those layers, making each of them individual choices or even stigmatized, you know, if you’re one of the few people wearing masks, there’s social pressure against it. So one key moment we highlight in Emily Oster’s role in these debates was back in early 2021, when schools were largely back in person, but in hybrid mode, where students would go every other day, and every other day, they would be at home doing remote instruction on their computer. So there were debates then about how safe it was to have all of these children in the classrooms together, and these debates were phrased as three feet of distancing in classrooms versus six feet of distancing, where three feet of distancing would allow students to come back five days a week. So her paper that she wrote with several other people looked at schools in Massachusetts, at COVID case rates for schools that had policies requiring six feet of distancing versus those allowing three feet of distancing, and presumably being fully in person. The study was quite weak, I would say, it committed some basic statistical errors, such as I don’t want to get too technical, but there is a scenario where you interpret what we call wide confidence intervals, or a high amount of uncertainty, that is not supposed to be interpreted as no difference. So she took that scenario where there was wide uncertainty between the case rates in three foot distanced schools versus six foot distanced schools, and said, ‘Look, there’s no difference,’ instead of saying, ‘Look, it’s inconclusive.’ But there are other ways of criticizing these studies too. Actually a postdoctoral fellow and a graduate student obtained the data, tried to reproduce the study, found basic flaws in the data, and also found that changing an arbitrary threshold that the Oster study set, even the slightest bit, this was a threshold saying what school districts are we going to consider fully remote in a given week, if you changed it from 5 percent to 5.1 percent, excluded a slightly different set of schools from the analysis, you showed a clear benefit of further distancing. This isn’t to say that schools should have been hybrid or remote at the time or fully in person, that is kind of a tangential concern. The concern here is that Oster, and her colleagues had a clear policy preference and produced very low quality research in line with their policy preferences, in line with the funders policy preferences, and it did make a difference, at very least, in serving as a piece of evidence that people like Rochelle Walensky, Tony Fauci could point to, in changing CDC guidance to no longer recommend six feet of distancing.

Adam: I wanted to get in the weeds a little bit because, I mean, I do think the details matter here and if there are —

Nima: We do like citations.

Adam: If there’s sloppy work going on, it is valuable, and the whole time I see this because, you know, I looked at The Atlantic’s output, and it’s very clear, okay, there’s clearly an editorial line here. Whether or not it comes from the woman who owns it or whether or not it comes from Jeffrey Goldberg, there’s clearly a political directive to convince liberals that COVID is over, and then you sort of torture the data, you know, they get the data in a room, they put a light bulb over its head, and they interrogate it until the data tells them what they want to hear. I’m just thinking the whole time this is so stupid, why are we insulting everyone’s intelligence by acting like, I mean, what was Emily Oster going to do? Take money from the Waltons, the Koch’s and the Arnold Foundation and come back and be like, ‘Oh, sorry, actually, we got to close schools for another six months.’

Nima: We got to listen to teachers’ unions.

Adam: I mean, was that ever going to happen? I mean, money doesn’t corrupt, can you even envision that world ever? ‘Oh, sorry. That’s where the data took me, incidentally.’ It’s like no, come on. That’s bullshit. The decisions have already been made before we sit down and crunch the numbers, because that’s what her ideological position is. So why don’t we just skip all the bullshit and say that’s what she’s doing? And the reason is, of course, is that because liberals don’t want to hear that, they want to believe in this kind of capital “S” Sciencism to justify why they think teachers unions are a bunch of malingering, greedy, layabouts, because otherwise it’s a tacky position to have.

Abigail Cartus: I think that there are real dividends for these institutional funders, as well as this emerging public health policy elite, I think there are real dividends to forcing these discussions to take place on the terrain of statistical evidence, and I think that that also kind of serves the function of removing that debate from any real public deliberation, and so, you know, as Justin pointed out that 2020, 2021, you know, that winter when we were getting ready to come back from winter break, it was a political priority for the Democrats at that time to get schools open because, you know, as we noted in the piece, I think we, you know, we had more about this in earlier drafts of the piece, but, as we noted in the piece, another reason why we think that this debate has exploded so much over schools is that schools serve crucial social reproductive function, which is, you know, it’s obviously not in the interests of Koch industries, or whatever, to have schools closed, because if schools are closed, then people can’t go to work, and I think the Democrats really took that up as a political problem, and we’ve seen it ever since Biden took office, I mean, even before that, but like this push towards policy-based evidence-making, you know what I mean, just conducting analyses that justify a political goal that has already been decided with basically no opportunity for public input or public deliberation, and then forcing the debate over that to take place on the terrain of statistical methods, data quality, you know, things that lay people and the general public really can’t always weigh in on.

Adam: And I think in fairness, I do think also, it’s important to note, there was a gendered aspect to the labor of childrearing at home, that I did think animated a lot of the liberal 180 on this, especially from professional women, what you heard a time and time again, not necessarily from those who are sort of conflicted or partisan, but genuinely struggling with this, that this was falling upon women more and that that was a key demographic and from a purely cynical political perspective, Democrats were going to lose that demographic. I do think that is something that was real. What are your thoughts on that argument? Obviously, the teachers unions of Chicago aren’t a bunch of dudebros so it’s either way, right?

Justin Feldman: Yeah, we get to some of this in the piece. But there are some complicated gender, race, class dynamics going on in combination with one another. For professional class women, and particularly white women, we saw in the opinion polls that they were most in favor of fully returning to in person instruction. This even included time periods. I think we have a tendency to view things from the perspective of the present, to remind everyone, this was before vaccines were available, or were just starting to become available so the risk was much higher, and at that time, there were expanded unemployment benefits that were primarily going towards working class people, they were at points quite generous, at points these expanded unemployment benefits brought people’s income to higher than what they were earning on the job. You could also take advantage of these benefits for childcare. But if you were someone who was in a professional-class job and a woman and tasked with childcare, you were also tasked with continuing to remain on your job, because you could do the job technically, from your home. So you have these two impossible competing tasks of childcare and remote work at the same time, and that is, you know, an untenable situation, I do sympathize, and I also think it’s important to understand like, what’s going on sociologically, also taking time off from work, extended unemployment, gaps in your resume, maybe a little more important to be able to explain if you are in more of a professional class career trajectory, despite being more economically privileged and having higher levels of education, and all of that. So that’s all very real. The problem was that by downplaying the harms of COVID transmission in school, and at points to harms of the virus itself, it made it so that there were no minimum standards set for schools. It was not as if there were groups of organized scientists out there, or politicians out there saying, we are not going to open schools in our country or our state unless we have surveillance testing of everyone, rapid tests in some, you know, regular basis, certain standards of ventilation, certain standards of masks, they’re basically saying, just open the school under whatever circumstances you can muster, and whatever that is, the benefits outweigh the costs regardless of the specific conditions.

Abigail Cartus: The Democrats, I think, just shot themselves in the foot with this, because, and this is something that we touch on in the piece, and we actually, we got a lot of criticism, I think, mostly from one specific person about this, but we sort of note that this whole episode, everything that’s happened since Biden took office, all of this, well, I mean, and even before that, all of this advocacy, all of this explosive fighting, you know, all this summer over the school board meetings and everything, we really see Emily Oster’s, the effect, I don’t know about the intent of her project with her COVID advocacy, but I think that the effect of it has been to sort of make a connection between these extreme far right and libertarian think tanks and foundations, and otherwise sort of progressive, liberal, upper class, white mothers, and I think that the Democrats just weren’t on the ball with that.

Nima: Shocking.

Abigail Cartus: Right, it should have been framed like, well, your government is failing you, we need to get COVID under control, you need to pressure not the teachers’ union at your kids’ school, but you need to pressure your local government to close bars and indoor dining and stuff so that it could be safer for kids to go to school. But the Democrats couldn’t figure that out.

Adam: Well, because of inflation, too.

Abigail Cartus: Yeah, well, yeah.

Adam: To close bars, you’d have to pay people to stay home and that was completely off the table. I mean, this is like with a lot of the big-city mayors, like, you know, I put myself in Lightfoot’s position when she just forces open bars and restaurants and, you know, without any deficit spending available it’s true, the whole thing was just extortion, you really had no choice, because there was no money to give people to stay home.

Abigail Cartus: Yeah.

Adam: You know what I mean? Inflation, you know, had basically foreclosed on that.

Abigail Cartus: Yeah.

Justin Feldman: I think they took it a step further than that, there was a White House speech back in, I think it was the fall of 2021, where Biden himself explains that, at least in their White House theory of why inflation was happening, it was that there was not enough spending in the service sector and too much spending on products that were affected by supply chain shortages, and if only they could convince people to spend more money on services that were in person, that were potentially dangerous to their health, could they get the economy back on track and inflation under control. So we’ve not just seen, we’ve not solely seen a shift from collective responsibility to individual responsibility in the context of the pandemic, but we’ve also seen a convincing of people to actually take riskier behaviors, because they are good for both inflation and for profit of certain industries.

Abigail Cartus: Yeah, well think about what happened just this past winter with the Omicron surge. You know, we found out about Omicron right around Thanksgiving, and the Biden administration’s response was, ‘No, it’s fine. Like, travel. If you’re unvaccinated, you’re going to die, but that’s not a reason to cancel your flights, definitely still take your flights.’ That was a completely disastrous response. But yeah, I don’t know, I don’t know to what extent, you know, the Democrats are constrained by economic realities but that was not good.

Nima: It seems like so much of the function of this data forward, pseudoscience writing has to do with turning people’s very real personal frustration, and giving it license under the auspices or the or the guise of scientific, credible research, to do the thing that they just would rather do, right? And then ascribing a villain for the people who are like, maybe that’s not great because of all these reasons, and then they become the villain, and that villain is the villain that is identified by the funders. So it’s like, the individuals may not be like, ‘You know what’s really pissing me off about this whole COVID thing? Teachers unions.’ That’s probably not as much on the radar for Emily Oster readers who are just like, ‘Oh my god, it’s really frustrating that school is remote, and I can’t go out and we’re terrified about everything and there’s this plague and I believe that that is real, but I wish it weren’t.’ It just gives license, you know, but it’s who those people are, having the privilege to be frustrated in the ways that they are, you know, omits entire swaths of the population, I mean, therefore, we’re not thinking about immunocompromised people, we’re not thinking about people who don’t have the ability to work from home, who don’t have the ability to make that choice, and yet, it just seems like the focused audience for this kind of reporting is focused on purpose, right? And by speaking to what they most want to hear deeply, because they just wish it were true, you then get a new set of villains, which further the political project that maybe they’re not even thinking about.

Abigail Cartus: Yeah, well, this is David Leonhardt’s entire thing, right, is making just these asinine comparisons between the rate of the pediatric mortality from COVID versus car crashes, I mean, he loves to talk about car crashes, completely omitting the entire kind of regulatory apparatus around automobile safety that actually could be much better, automobile road safety in the US is not amazing.

Adam: Right. It’s a huge public policy failure.

Abigail Cartus: Yeah, and this is also part of, I think, Emily Oster’s larger intellectual project as kind of a public intellectual because she’s written these books on parenting, and the parenting books are very, very similar, you know, mostly they are arguing, or mostly she is arguing in these books, that universal guidelines, for example, health guidelines for pregnancy are overly cautious, Justin has identified this as a deregulatory attitude, and that the implicit reader of her books, which is an upper class, professional class, white woman, that those guidelines need not really apply to you, the reader, because, you know, you have all these unacknowledged privileges, obviously, and you have now with Emily Oster’s help the ability to evaluate this data for yourself and make your own risk assessment, and, you know, if you’re dying to eat salami while you’re pregnant, be my guest, you know what I mean? I’m not interested in necessarily being prescriptive about individual people’s behaviors, but that is a through line that we picked up on in her work on pregnancy and parenting that has carried through to her COVID advocacy, you know, maybe a federal program to curb COVID transmission would be great, but COVID is not really that big of a deal for children, your kid is so much more likely to die in a car crash, like all these really kind of ghoulish and heinous comparisons that rest on this idea that a certain type of professional-class person can exempt themselves from the social contract through personal risk assessment.

Adam: Leftism, or any ideology, socialism, whatever, progressive liberalism, et cetera, should be empirically driven. We don’t want to come off as anti-intellectual or anti-data, of course, which is sort of different than wielding the label of capital “D” Data to kind of inflate one’s ideological output. With an understanding that lowercase “s” science is very, very important as a foundational tool of any worldview, rather than, say, like a branding tool to wield to sort of shut down critics and call them all a bunch of irrational women. What is the role of data in progressive spaces? How can it be used for non-cynical reasons? And I guess, more important than anything, what are its limitations? What is the sort of limitations of data as above-the-fray, apolitical concept?

Justin Feldman: I’ll speak for Abby for a second, if I may, we’re scientists, we do believe not in science, but in the value of scientific inquiry, I would say. And that’s what I think science and data offer the left or any particular political movement, is a tool of inquiry. One thing we can do is help people identify problems that they may not know exists, because your lived experience is a very important source of information, but there are things that are beyond one’s lived experience. Like maybe there’s a chemical contaminant in your community’s drinking water that you would not have been aware of. Another thing we can do is give people tools to use the power they already have. I’m thinking in particular, Abby and I spoke with various groups of teachers, especially in the winter 2020 into 2021, about COVID safety, and some of these teachers were unionized, and some of them were not organized, and we helped them navigate questions about what they should be fighting for, what should be sticking points in negotiations around things like community case rates and masking and testing. That’s something we drew on our reading of what was out there in the science, including its limitations, and how to weigh the uncertainty that existed with that science. And there were groups, there was this group science for the people in the late 1960s and ’70s, that asked exactly this, like, ‘What can science and scientists and engineers offer the left?’ And that group has been restarted and some of us have been trying to do similar efforts during the pandemic.

Abigail Cartus: The only thing that I will add to that is that, I would just say that, you know, the biggest thing is being upfront about your value orientation. In science, we often pretend that a good scientific question or a meaningful scientific question is self-evident, but, you know, those questions are actually highly socially constructed on both sides, right? So I think it’s good just to be upfront about your value orientation, and upfront about how uncertainty is just a constitutive part of scientific knowledge and knowledge production, being explicit about what your values are in approaching that uncertainty, and whose values are being prioritized and what outcomes you would like to see.

Nima: Well, I think that’s a great place to leave it. We have been speaking with Abigail Cartus, epidemiologist at Brown University where she focuses on perinatal health and overdose prevention in her work at The People, Place & Health Collective, a Brown School of Public Health research laboratory, and also with Justin Feldman another epidemiologist and a Health and Human Rights Fellow at the Harvard FXB Center for Health and Human Rights. Abby and Justin, we really cannot thank you enough for joining us today on Citations Needed.

Abigail Cartus: Thanks so much. It was really fun.

Justin Feldman: Thanks so much. Great to be here.


Adam: Yeah, I think how the quote-unquote “left” or “progressives” or wherever you want to call them, or even liberals, like how they use data is an interesting one, because it’s not, my general thing is you can’t really out wonk yourself with most of these people. Like if someone comes into an argument, and their priors are hardwired, showing them a bunch of polls and data, it’s not going to do a ton without really going back and interrogating those priors, without getting to dorm room about it, because I do think you can become a little bit, ‘What’s the real crime, man?’ But I think that’s an interesting question, because one of the clever things Oster does when she talked about reopening schools for children, was effectively hand wave away the risk to the people who work in schools, which was the key element of the labor condition, again, we do not have robots that teach children and feed them and provide the medical care and provide security.

Nima: And they have lives outside schools. They don’t live in the school.

Adam: And so much of that was just completely ignored from this cost benefit analysis. ‘Well, they’re children, they’ll be safe.’ Well, okay, even granting that, which is a huge thing to grant, but even granting that, there’s people who work there, and this is why you had this kind of cohort of wealthier, most likely urban people, who basically, ‘But they’re the help, aren’t they basically just there to teach kids?’ It’s like, well, no, it’s a job. I know, we want to be romantic about these things but it’s a job and jobs require certain, unions have an obligation to maximize the employee’s health outcomes, right? I mean, that’s sort of a one-on-one union thing you would worry about.

Nima: Right. But because teachers’ unions are relatively strong in a country where unions are declining, where unions don’t have as much power as they did, but teachers’ unions do have some semblance of unity, power, influence, but because of that, they wind up being the villains of the story. But it’s also important to note that teachers, by and large, not every single individual teacher necessarily, I don’t have the data on it, but teachers want to be teaching and they have trained to be teachers in person with students, not to be online University of Phoenix professors. So the issue is not that teachers’ unions were backing something where teachers just wanted to be lazy and phone in or zoom in their fucking work, they wanted to be in person, teachers had an incredibly hard time and are still during the pandemic. It fucking sucks to have to do that job remote, but the issue is that if there were unsafe working conditions, those needed to be addressed either by not going in or, as we’ve said a million times, right, better ventilation, better PPE.

Adam: The main problem was that the teachers’ unions didn’t gather millions of dollars to create a school data tracker. See that was their problem. They had to rely on the Arnold Foundation and the Waltons and the Kochs.

Nima: And now that the data, Adam, shows that roughly three of every four kids in the country have gotten COVID, but you know, remember, kids don’t get COVID.

Adam: Right.

Nima: Right. Schools can’t possibly be places where people contract this disease that the entire globe has.

Adam: Yeah, liberals just love to close things and shut themselves inside, they love —

Nima: That’s the liberal — (laughs) — that’s what the data shows.

Adam: Liberals just have a pathological irrational fear of everything, they all just have brain disease, unlike the Koch billionaire-backed Center for Economic Reasoning and sanity, so you’re not just a bunch —

Nima: That’s right.

Adam: Of crazy women.

Nima: Where you just, you know, stare down a virus that cowers in fear. Exactly. That will do it for this episode of Citations Needed. Thank you, everyone, for joining us. Of course you can follow the show on Twitter @CitationsPod, Facebook Citations Needed, and you can support the show in a couple of different ways. If you’re so inclined, you can go to our new merch store, where you can pick up a Citations Needed t-shirt or a tote bag. You can find that at Bonfire.com/store/Citations-Needed. And of course you can become a supporter of the show at Patreon.com/CitationsNeededPodcast. All your support through Patreon is incredibly appreciated as we are 100 percent listener funded. And as always, a very special shout out goes to our critic level supporters through Patreon. I am Nima Shirazi.

Adam: I’m Adam Johnson.

Nima: Citations Needed is produced by Florence Barrau-Adams. Associate producer is Julianne Tveten. Production assistant is Trendel Lightburn. Newsletter by Marco Cartolano. Transcriptions are by Morgan McAslan. The music is by Grandaddy. Thanks again for listening, everyone, we’ll catch you next time.


This Citations Needed episode was released on Wednesday, June 1, 2022.

Transcription by Morgan McAslan.