A recent Harvard study concluded that graduate students are over three times more likely than the average American to experience mental health disorders and depression. The study, which surveyed over 500 economics students from eight elite universities, also concluded that one in 10 students experienced suicidal thoughts over a two-week period, a result consistent with other recent reports. While these findings are alarming to some, as a current graduate student myself, I regard them as hardly surprising. But to understand the struggles graduate students face, you have to understand the structure of graduate school itself.

Most people probably lump doctoral students into the same category as undergrads or students in professional schools such as law or medicine. The reality is their lifestyle and the nature of their work are fundamentally different. In the STEM fields where I have personal experience, as well as many other fields, graduate students are really hardly students at all. For most of their programs, which last over six years on average, they aren’t preparing for written exams, taking courses or doing any of the tasks usually associated with student life. Instead they are dedicating often over 60 hours a week towards performing cutting edge research and writing journal articles that will be used to garner millions of dollars in university research funding.

While graduate students are compensated for their work by a supervising professor, their salaries substantially lag what the open job market would offer to people with their qualifications, which often include both master’s and bachelor’s degrees. For example, graduate student salaries are typically around $30,000 a year for those in STEM—and can be substantially lower for those in other fields.

Further, unlike many professional school students, doctoral students do not leave their program with job security or even optimistic financial prospects. In fact, according to a study in 2016, nearly 40 percent of doctoral students do not have a job lined up at the time of graduation. Even for those who do snag a job, mid-career salaries can be significantly less than those for individuals who graduated from other professional programs.

So if doctoral students are underpaid and overworked, why do over 100,000 students—more than the number for dentistry, medical and law schools combined—complete these programs every year?

There are many answers to this question, and they vary from department to department, individual to individual. For some, graduate school is a convenient next step, a way to inch towards adulthood while keeping your career options open and remaining in a familiar university environment. For others, graduate school offers something they simply cannot get elsewhere. These students enter graduate school because they are extremely passionate about their field—passionate enough that they are willing to dedicate over six years to studying off-the-wall research ideas in excruciating detail.

Universities, with a commitment to intellectual freedom, are one of the few environments capable of providing the funding and resources necessary for this type of work. So, we put up with the hours, put up with the pay, and put up with the dwindling career prospects in the hope that we can pursue research we are passionate about—and then we cross our fingers and hope the rest will work out.

Unfortunately, as the study pointed out, it often does not work out. Mistaking casual interest for passion, many students realize halfway through their degree that they aren’t as enthusiastic as they thought about their research. Still several years away from graduating, they have to deliberate between grinding through the remainder of their program or exiting early and entering the job market in an awkward position: underqualified compared to other doctoral graduates and inexperienced compared to others who joined the workforce directly after college.

Even those who are interested in their work have to grapple with seemingly infinitely postponed graduation dates. Unlike other programs, there is no “units threshold” you have to meet in order to graduate—instead your graduation date is overwhelmingly determined by the amount of novel research you perform. No matter how hard you may work, no results will likely mean no degree. Even the best researchers can see years slip by without any significant results as a result of factors completely out of their hands such as faulty equipment, dwindling research budgets or pursuing research ideas that simply just don’t work.

Even for students who are lucky enough to produce results, frustratingly, individual professors have their own standards for what constitutes “enough research” to graduate. Is it four first-author research articles? What about one review paper and a few conference presentations? The answers you hear will vary widely, and ultimately, a student’s supervising professor usually has sole power in determining when a student graduates. At best, this creates a confusing system where students perform substantially different amounts of work for the same degree. At worst, it fosters a perverse power dynamic where students feel powerless to speak out against professors who create toxic working conditions, even resulting in cases of sexual exploitation.

Then there’s always the existential, “what even is my purpose?” mental black hole that many graduate students fall into. Yes, research has historically produced innovations that have revolutionized society. But for every breakthrough there are many other results without any clear social application, and given the slow, painstaking process of research, you may not be able to tell which is which for decades. As a student, it’s can be easy to doubt whether you’re pursuing work that will ever be useful, producing a sense of meaninglessness for some that can facilitate depression.

Clearly, if nearly 10 percent of the graduate population is experiencing suicidal thoughts, something is not working right in the system. Still, progress on these issues has been slow, largely because the people who are most affected—graduate students– are often the ones with the least agency to spur change. As a student, by the time you’ve seen the cracks in the academic infrastructure, you’ll likely only have a few more years until graduation. Do you really want to dedicate time towards fixing a system you’re leaving soon when you could be performing career-vaulting research instead? Are you willing to risk upsetting professors whose recommendation letters will dictate your employment prospects? For many, the answer is no.

Granted, the issues surrounding graduate student mental health are much easier to describe than to solve. But if academia is good at anything, it’s tackling complex, multifaceted problems exactly like these, and there are number of starting points for both students and administrators to push forward. For example, universities could require multiple advisors within a student’s field to evaluate degree timelines, preventing labor exploitation by a single professor with vested interests in prolonging graduation dates.

Departments could also streamline their graduation criteria to reduce disparities in student workload amongst different research groups and to increase transparency of degree requirements. Further, administrators could increase funding for popular student mental health services and subsidized housing that help graduate students offset cost-of-living expenses. Some universities have already adopted these policies in earnest and others only in name, but the point is academic institutions need to be making a concerted effort to improve the graduate student experience. For all the research they have done, graduate students deserve to start seeing some results.