Jonathan Haidt’s Substack After Babel explores the effect of social media on teen mental health. The site is rich with data and analysis, containing 300+ pages of meta-analyses and nuanced discussions on the required burden of proof. Haidt’s critics responded in kind, with thoughtful articles of their own. In the world Gen Zers like myself have to navigate — defined by charged tweets, cancellation campaigns, and extreme polarization — Haidt’s Substack serves as a rare forum for healthy discussion of contentious political and cultural issues. It was possible for me to be a part of it too, by contributing to the collaborative Google docs and adding my thoughts and questions to the comments section.

Unfortunately, this is not the norm. Most discussions across political and ideological lines falter at the first step: establishing and recognizing a common set of facts. Instead people tend to talk past one another, focusing on completely different data points to advance their preferred narrative without deeply (let alone charitably) engaging with the facts marshaled by people on the other side of the issue.

Take immigration, for instance. Liberal-leaning media outlets, such as the New York Times and CNN, cite many statistics on the positive aspects of immigration: Children of immigrants tend to do better than natives from a similar background, less skilled immigration doesn’t displace natives’ jobs or increase crime, and immigrants boost the GDP. But no articles mention the significant debt that results from supporting immigrants without a high school education. According to table 8-12 in the 2017 National Academies report, Economic and Fiscal Impact of Immigration, the present value of the net lifetime cost for an immigrant without a high school degree ranges from $117,000 to $301,000. Given that the per-capita income in the Northern Triangle is around $4,000, it’s unclear whether foreign aid may be more cost-effective.

Similarly, conservative-leaning media outlets, consisting of the Wall Street Journal and Fox News, frequently criticize DEI initiatives. But their reporting on racial discrimination is more scarce. A recent nationwide study — the largest ever conducted — found that African Americans are 2.1% less likely to be called for an interview, with 5% of firms responsible for more than 50% of the discrimination. Another study found that racial discrimination accounted for a difference of 4.2% in New York City’s pretrial release rates. These significant findings went unmentioned.

And it’s the same story with every single policy topic. What impact do “no excuses” charter schools have on students’ noncognitive skills, college completion, and future income? How cost-effective are social services compared with police in terms of reducing homicides? What impact does imposing more regulation on hospitals (e.g., capping fees, global budgets, etc.) have on health care costs and quality? These are central empirical questions where the academic literature has plenty to say, but individuals are so divided across political and ideological lines that they tend to selectively mobilize apparent facts that flatter their preexisting views and broadly ignore data that undermines their preferred narrative.

It’s not just individuals that operate this way. Part of the reason we are unable to achieve clarity on the simplest of ground-level facts is that many institutions of knowledge production — from media outlets to think tanks to colleges and universities — have themselves become partisans in political struggles and culture wars. People have few sources that are trusted across the political spectrum. For those who want a more comprehensive understanding of where the evidence points on various issues, there aren’t many places to turn.

Americans seem to be deeply dissatisfied with this status quo. Trust in institutions related to knowledge and information production has consistently declined over the past 50 years. This is likely related to the growing politicization within and about these institutions. A recent study reveals that highly politicized institutions are less trusted, even by those sharing the institution’s ideological slant.

There’s a strong appetite among Americans for more balanced, objective analysis. This presents a strong moment of opportunity for those committed to addressing persistent failures in institutions tied to knowledge production.

One promising approach that can actually harness people’s divergent commitments in the service of truth-seeking is adversarial collaboration between academics who disagree. These collaborations compel participants to reckon more fully with the arguments and inconvenient data of the “other side” and build out a zone of common reality. For instance, a recent paper produced via adversarial collaboration on gender bias in academia found that there has been genuine and significant progress in many dimensions of gender discrimination. Along some dimensions, gaps between men and women have more or less completely closed. However, along other dimensions, evidence of bias does persist. That is, both sides of the discussion were importantly correct in different respects, and as a result of the adversarial collaboration we now have a fuller understanding of the specific ways in which the different narratives are true, and more important, how they might cohere or be reconciled into a more complete and nuanced perspective.

University of Pennsylvania’s Adversarial Collaboration Project aims to fund and help organize more original research of this nature that harnesses competing viewpoints. For those interested in addressing polarization or eroding faith in the press or higher ed, this could be an important model to draw from.

What could the simplest instantiation of this look like? For each major policy issue, have a couple of analysts from different sides of an issue collaborate to produce three-to-five-page briefs that accurately capture the diversity of opinion and try to establish some common ground or shared facts. Post the document online and have the authors respond to the highest-quality comments either directly or in a follow-up brief. Extremely basic? Yes. However, it could provide a desperately needed empirical common ground, plugging a gaping hole that is tearing apart our nation of more than 330 million.

By simply ensuring that all perspectives are represented fairly, it would likely build a large audience very fast, much as Hadit’s Substack has. And something as rudimentary as having participating scholars engage with high-quality comments could help give everyday Americans more of an opportunity to have a voice in science, its findings, and their implications and applications.