Take the two most rigorous, nationally representative health surveys the United States government conducts, isolate the lawyers in each, and ask a simple question: are attorneys in psychological distress? You will get two different answers. According to a 2026 study in the Journal of Empirical Legal Studies, the National Health Interview Survey finds that lawyers experience psychological distress at rates lower than the general public and only modestly higher, at most, than their similarly educated peers. The National Survey on Drug Use and Health, drawing on the same population of practicing attorneys, finds that more than forty percent report moderate or serious psychological distress in the past year, a rate well above the general public and well above what the other survey reports. Two gold-standard datasets, one profession, opposite conclusions.
This is not a minor technical footnote. For nearly a decade, the legal profession has organized its understanding of itself around the idea of a well-being crisis. The purpose of this piece is to take that idea seriously enough to examine the evidence behind it, and to argue that the honest reading of the data is more complicated, and more useful, than the headline figures suggest. The numbers that frighten attorneys most are also the least stable. The number that should concern them is the one that survives every method of measurement, and it is not the one the profession spends most of its time discussing.
How the crisis narrative was built
The modern conversation about attorney mental health has a clear origin point. In 2016, Patrick Krill, Ryan Johnson, and Linda Albert published The Prevalence of Substance Use and Other Mental Health Concerns Among American Attorneys in the Journal of Addiction Medicine. Funded by the American Bar Association and the Hazelden Betty Ford Foundation, the study surveyed 12,825 licensed, employed attorneys and produced figures that have been cited ever since: roughly twenty-one percent screened positive for problematic drinking, twenty-eight percent reported some level of depression, and nineteen percent reported symptoms of anxiety. Younger attorneys and those earlier in their careers carried the highest risk. The lead author called the findings a call to action, and the profession responded. Well-being committees, bar association task forces, and law firm wellness programs all trace some lineage to that paper.
The study deserves credit for putting empirical weight behind concerns that had circulated for decades as anecdote. But its design carried a limitation that shapes everything downstream. The respondents were volunteers. Attorneys received an invitation to participate in a survey about substance use and mental health, and those who chose to answer were, by definition, a self-selected group. That single feature is the hinge on which the entire debate now turns.
Why the numbers diverge
A survey that people opt into is vulnerable to response bias. The attorneys most motivated to spend time answering questions about drinking and depression are often those for whom the subject is most salient, which is to say those who are struggling or who care about the issue. The result is a sample that may not resemble the profession as a whole. This is not a flaw unique to the 2016 study; it is a structural feature of nearly all the influential survey work on lawyer well-being, and it means the resulting prevalence figures should be read as upper bounds rather than population truths.
In 2021, Yair Listokin and Raymond Noonan made this argument directly in Measuring Lawyer Well-Being Systematically, also in the Journal of Empirical Legal Studies. Rather than rely on volunteers, they turned to the National Health Interview Survey, a large randomized dataset administered by the Centers for Disease Control and Prevention that happens to capture a steady sample of lawyers each year. Their conclusion unsettled the conventional wisdom. On that data, lawyers were not particularly unhappy. Their rates of mental illness were lower than the general population and not meaningfully different from those of similarly educated professionals such as physicians and dentists. The unhappy-lawyer narrative, they suggested, rested on a weak empirical foundation.
The 2026 study by Benjamin Pyle and Clifford Rosky extends this work and complicates it further. One might expect that simply switching from volunteer surveys to rigorous random ones would settle the question. It does not. When Pyle and Rosky compared two high-quality national surveys, both randomized, both methodologically sound, they found the surveys disagreed with each other. The National Health Interview Survey continued to show lawyers at or below general-population distress levels. The National Survey on Drug Use and Health showed distress in more than forty percent of attorneys. The divergence does not appear to be a story of careless data; it reflects genuine differences in how each instrument asks about distress and how it is administered. The lesson is sobering for anyone who wants a clean number: psychological distress, as a measured quantity, is unstable enough that the choice of instrument can move the headline by a factor of several.
The finding that survives every method
Amid that instability, one result holds steady. Across the volunteer surveys and the randomized ones, across the optimistic readings and the alarming ones, lawyers consistently show elevated rates of alcohol misuse relative to the general public and relative to their similarly educated peers. Listokin and Noonan found it in the National Health Interview Survey even as they found low rates of mental illness, and they noted that problematic drinking among lawyers had grown more common over the preceding fifteen years. Pyle and Rosky found it in both surveys they examined. The 2016 study found it as well, and identified it as most pronounced among the youngest attorneys.
When a finding appears regardless of who is asked and how, it earns a different kind of confidence than a finding that swings with methodology. The elevated drinking is real in a way that the contested distress figures are not. And yet it tends to receive less sustained attention than the broader and shakier claims about a mental health crisis, in part because the profession has a complicated relationship with alcohol. Drinking is woven into how legal work is conducted and rewarded, from the client dinner to the firm reception to the networking event organized around a bar. A behavior that is normalized is a behavior that is easy to overlook, even when the data keep pointing at it.
What a clinician hears in the disagreement
The temptation, when two distress surveys disagree, is to conclude that the truth lies somewhere in the middle and that lawyers are therefore doing roughly fine. From a clinical vantage point, that is the wrong inference, and the steady alcohol finding is the reason why.
Consider what these instruments are actually measuring. A psychological distress scale asks a person to report on internal states over a recent window of time, often the past thirty days. Accurate reporting requires that the respondent notice the state, interpret it correctly, and feel willing to disclose it. Each of those steps is fragile, and each is arguably more fragile for attorneys than for the average respondent. The professional identity of a lawyer is built around composure, control, and the capacity to absorb pressure without visible strain. The same disposition that makes someone effective in an adversarial setting, the ability to power through, can interfere with the honest self-report that a distress questionnaire depends on. A profession trained to project that nothing is wrong is not a profession well suited to telling a survey that something is.
Alcohol use is a different kind of measurement. The screening instruments do not ask how a person feels; they ask what a person does. How often, how much, whether others have expressed concern, whether the person has tried and failed to cut back. These are behavioral facts, harder to reframe and harder to hide from oneself than an interior mood. This is why the drinking signal is so durable across studies: it is anchored to conduct rather than to introspection.
Read together, the two patterns suggest a coherent story rather than a contradiction. The clinical literature on self-medication holds that people frequently use alcohol to regulate states they are not fully processing in any other way, and an elevated rate of drinking can therefore be the visible, behaviorally measurable trace of distress that the feeling-state questionnaires are failing to capture. The instability of the distress numbers and the stability of the alcohol numbers may be two views of the same underlying reality: a population under genuine strain that is more reliably detected by what it does than by what it reports feeling. That interpretation also fits the third strand of this research. In Stressed, Lonely, and Overcommitted, published in Healthcare, Krill and Anker found that among lawyers, perceived stress, loneliness, and overcommitment to work were powerful predictors of suicidal ideation, with the most stressed attorneys far more likely than their calmer colleagues to contemplate suicide. Distress, when it is present, has serious consequences. The difficulty is simply that distress is hard to measure cleanly, while its downstream behaviors are not.
What this means for the attorney reading the headlines
The practical value of getting the measurement right is that it changes what an attorney should do with the next alarming statistic that crosses the desk. When a figure appears claiming that some large percentage of lawyers are depressed or anxious, the useful first question is where the number came from. A figure drawn from a survey that attorneys volunteered to complete describes the people who chose to answer, not the profession. That does not make it worthless, but it does make it the wrong tool for self-diagnosis. No individual should conclude that they are or are not struggling on the basis of a contested population estimate.
The more reliable move is to watch behavior rather than to interrogate mood. A person can fail to feel depressed in any way a questionnaire would capture and still be drinking in a manner that signals an unmet need. The robust finding in all of this research points to alcohol as the place where strain in this profession most consistently becomes visible, which makes it the most honest thing to take personally. The relevant questions are concrete. Has consumption increased. Is it serving to manage stress at the end of a difficult day. Has anyone expressed concern. Has an attempt to cut back proven harder than expected. Those questions do not depend on resolving an academic dispute about prevalence; they depend only on a willingness to look at conduct without the reflex to explain it away.
Honest measurement is not a detour from helping attorneys. It is the precondition for it. A profession that misreads its own data will build the wrong interventions, aim resources at the wrong targets, and reassure or frighten itself for the wrong reasons. The American Bar Association and Krill Strategies currently have a new national survey in the field, with results expected this year that will, for the first time, attempt a national baseline for burnout and examine how artificial intelligence and the digitization of practice are affecting attorneys. That work will generate another round of headline figures. The attorneys who benefit most from it will be the ones who read it the way the data deserve to be read, with attention to method, skepticism toward any single dramatic percentage, and respect for the one finding that has never wavered.
AttorneyTherapists.com maintains a directory of licensed therapists and professional coaches who work specifically with attorneys, for readers who recognize something of themselves in the patterns described here and want to speak with someone who understands the profession.
