One lab at a time Is not enough: What ten years of toxic lab reporting actually tells us

In the fall of 2015, journalist Michael Balter began reporting on what he would later describe as the dark side of science: the abuses that powerful academics visit upon students and young researchers. He has been doing it ever since.

Over the past decade, his investigations have exposed dozens of labs across the United States and Europe. Scientists at Yale, Adelaide, Harvard, and institutions across the world. Cases involving bullying, data theft, sexual harassment, gaslighting, financial exploitation, and the systematic destruction of young researchers' careers and mental health. His reporting has contributed directly to the termination or forced resignation of multiple abusers.

It is important, rigorous, and necessary work. And it is not enough.

Not because the journalism is inadequate. Because investigative journalism, by its nature, operates case by case, lab by lab, story by story. And the problem it is documenting is not a collection of individual cases. It is a system.

The pattern that runs through every case

Read enough of Balter's reporting and a pattern emerges that transcends any individual lab or individual supervisor.

In the Jetz lab at Yale, seven former members described nearly identical experiences: belittling, micromanagement, publicly humiliating meetings, data appropriated without credit, careers subordinated to the supervisor's agenda. Yale knew about the problems from at least 2011. A distinguished professor documented the abuses in writing, brought them to the administration repeatedly, and watched as the institution promoted Jetz anyway because, as Yale's own faculty member later wrote, the university valued "the money he raises in grants and the fame generated by his publications in high impact journals more than the emotional and professional damage done to the lives of graduate students and postdocs."

This is not a Yale problem. It is the same calculation made at institutions around the world, every time a tenured professor with a strong grant record faces accountability for how they treat the people working under them.

The costs of abusive supervision fall on individuals who pass through quickly. The benefits of tolerating it accrue to institutions that stay permanent. The math always comes out the same.

"Walter is a symptom of a broken system"

One former postdoc in the Jetz lab captured it precisely in a statement shared with Balter:

"The academic model as it stands prioritizes and rewards exceptional output at any cost. It rewards those who are willing and able to claw their way over the bodies of colleagues and employees to achieve results. As long as the machine continues to produce, there are too few consequences and too great a reward to mistreating others. In fact the system selects for this."

This is the diagnosis that Balter's decade of reporting has produced, not from his own commentary, but from the researchers living inside it. The system does not merely tolerate abusive supervisors. It selects for them. The behaviors that produce the highest research output, relentless pressure, exploitation of early career dependency, willingness to claim credit and deflect blame, are the same behaviors that produce the most toxic lab environments. The system cannot distinguish between the two because it is not designed to.

Institutional accountability mechanisms, grievance procedures, climate reviews, HR processes, operate within this same system. They are managed by institutions with the same incentive structure that produced the problem. The Yale timeline from Stearns is a ten-year documented record of exactly what happens when institutional accountability confronts a well-funded, highly cited, tenured professor: the institution finds reasons to continue.

What investigative journalism can and cannot fix

Balter's reporting has had real consequences. Abusers have lost positions. Institutions have been forced to respond publicly when they would have preferred to manage problems quietly. Young researchers who were otherwise silenced have had their experiences documented and taken seriously.

But each investigation begins after the damage is already done. The seven former Jetz lab members who spoke to Balter had already left the lab. Many had already left science entirely. The investigation documented what happened to them. It did not, could not, help the researchers who were considering the Jetz lab during the years the problems were known but not yet public.

This is the structural gap that no amount of investigative journalism can close. The information that would protect the next researcher has to exist before they make the decision, not after they survive the experience.

The information that travels in whisper networks

Balter notes in his Jetz reporting that the lab's bad reputation had spread far and wide through the research community. Former lab members were actively warning potential recruits to stay away. The information existed. It just existed in informal networks that only reached you if you happened to know the right people.

This is how researchers have always tried to protect each other. The hallway conversation at a conference. The DM from a former lab member. The quiet warning from someone who knew someone who had been through that lab. It is the original peer review of labs, running decades before any formal mechanism existed, and it still operates primarily through access and luck.

If you knew people who knew people who had been in the Jetz lab, you might have heard the warnings before you applied. If you were new to the field, if you were an international student without an established network, if you were the first person from your institution to consider that lab, you probably did not.

The whisper network is not a system. It is the absence of one.

What changes before the decision is made

RateYourPI (rateyourpi.com) is built around the same motivation that drives Balter's reporting: the researchers who have direct experience of a lab should be able to share that experience with the researchers who are about to make a decision about it.

The difference is architecture. A journalistic investigation documents what happened in a specific lab at a specific time, requiring that a researcher find it and read it before making their decision. A permanent, searchable review platform means that every researcher who submits an honest account of their experience contributes to a record that is available the next time someone searches that supervisor's name, whether that is one year later or ten.

Reviews on RateYourPI are verified through institutional affiliation. The email used for verification is deleted the moment a review goes live, making it technically impossible to identify who wrote what after the fact. A researcher who experienced what the Jetz lab sources described does not need a journalist to be willing to investigate and an editor to be willing to publish. They need five minutes and an institutional email address.

This does not replace investigative journalism. Balter's reporting can force institutional accountability in ways that anonymous reviews cannot. A published investigation with named sources and documented evidence creates public pressure that a review platform cannot replicate.

But the information problem that produces new victims every year while investigations are ongoing is a different problem, and it requires a different solution. The seven researchers who eventually spoke to Balter had all already suffered. The question is how many researchers entered the Jetz lab during the years the institution knew and said nothing, and whether any of them could have been reached by an honest account of the lab's climate left by the people who had already been through it.

Ten years of documentation points to one conclusion

What Balter's decade of reporting has established, case by case across dozens of labs and institutions, is that abusive supervision in academia is not an aberration. It is a predictable output of a system with specific structural features: extreme power concentration in the supervisor-trainee relationship, institutional incentives that protect grant revenue over researcher welfare, and an almost complete absence of information flow from people who have direct experience to people who are about to enter the same situation.

The investigation of individual cases exposes the symptom. Changing what information researchers have access to before they commit addresses the condition that keeps producing them.

The former Jetz lab postdoc was right. Walter is a symptom. So is every other case in Balter's archive. The system selects for this. The question is what changes the selection pressure, and whether any of that can happen before the next researcher walks into the next lab without knowing what they are walking into.