What do we want graduate school to be?

Figure 1: Pie chart of respondent demographics.

Recently, an email that was sent to graduate students at the astronomy department of a US university caught the attention of faculty and students in departments across the country. Perhaps the most important result of this letter is a serious discussion about what the graduate student experience should be like.

A few weeks ago, we asked you — the readers of Astrobites — to share your thoughts about this issue. We extend our thanks to everyone who completed the survey, a remarkable number — more than 400 readers. While the survey was unscientific and open-response (not randomized), and therefore cannot represent exactly the views of the astronomy community or academia generally, the large response rate allows us to make some statistically interesting observations and emphasizes the interest in this subject among students, faculty, and post-docs.

Below we report some of the results from the survey. We intend for the survey (and this post) to motivate and provide relevant information for a conversation in the community. We encourage you to continue the discussion by leaving a comment below.

You can click on all the figures in this post to enlarge them. Note that the total number of respondents to each question varies throughout the survey, as responding to individual questions was not mandatory.

Reactions to the letter

Figure 2: Graduate students were asked if the letter would make them feel more or less favorably about the faculty who sent it.

First, it needs to be stipulated that the faculty who sent the email presumably had good intentions, as many of our respondents commented. However, the survey results clearly show that the the letter itself is unpopular, and several respondents wrote that it “undermined” its intended message.

Among graduate students, 158 of 284 respondents said the letter made them less interested in pursuing a career in academia, 10 said it made them more interested, and the rest indicated it would have no effect. Asked directly whether the email would make them feel more or less favorably towards the faculty that sent it, responses were strongly skewed toward “Less favorable” (see Figure 2). As one respondent wrote, “I was appalled by that email… It would not have motivated me at all and I would have struggled under the weight of such pressure.”

Only 3 of 59 respondents who identified themselves as faculty members said they would have been in favor of sending the letter as written, and 42 of 59 indicated they would be strongly against sending it.

Figure 3: Respondents were asked to select any terms which they would use to describe the content and/or tone of the email.

Asked to pick from 12 positive and negative terms describing the tone and/or content of the email (Figure 3), the most commonly picked term among all groups was “Condescending” (~70% of respondents). The least-commonly selected terms were “Helpful,” “Motivating,” and “Professional” (~10%). Undergraduates were about twice as likely as graduate students or faculty to select the term “Professional.”

About 40% of respondents labeled the letter “honest.” One respondent who identified herself as a female Ph.D. who previously held a tenure track position wrote, “If you really want a job in academia bear in mind that you’ll have to work your *** off to the exclusion of much else (family, hobbies, income, down time) to outcompete your peers for a declining number of tenure-track jobs that require intense dedication and don’t pay particularly well. If you don’t like that assessment, get out of grad school now.” This perception is echoed in several comments from respondents, although it conflicts somewhat with the ‘working hours’ responses we present below.

One respondent who self-identified as a graduate student not in physics/astronomy wrote, “That letter was much nicer than those that my research adviser writes. It is honest, shares the truth about research while still being respectful of the reader. Sometimes my adviser treats us like disobedient children.”

Working hours

Figure 4: Working hours per week reported by respondents in various occupations (see legend) and at institutions of different types (different panels).

Arguably the most provocative premise of the letter was that it is necessary to work “80-100” hours per week to achieve success. Many respondents commented that it is more appropriate to evaluate students based on accomplishments rather than hours clocked, but as the letter rightly states, graduate students often want to know what is “normal” or expected of them. To assess norms, we asked respondents to estimate their working hours (Figure 4). The mode among respondents was 40-50 hours per week, and only 10 of 414 respondents to this question (~2%) said that they work 80-100 hours per week. No faculty or post-doc respondents reported working that much.

Another controversial implication from the email is that students who intend to pursue a career focused on research, as opposed to education or non-academic pursuits, need to work longer hours. The survey data addresses this only indirectly. In Figure 4, we show the hours worked by students and professors as a function of institutional focus. The differences in working hours between academics at research and teaching-focused institutions are small (and not statistically significant according to the KS test). The mean working time is slightly higher at research-focused institutions, but is 40-50 hours per week at all institutions.

Figure 5: Do professors and post-docs work as much now as they did when they were graduate students? On average (red line), they say they worked 4 hrs/wk more as graduate students. The area of the points indicates the relative number of respondents in each bin.

Some commenters have speculated that the working hours suggested in the letter are inflated (as our survey results show), and reflect a “selective memory” or “back in my day” effect where senior researchers seem to remember having worked more as graduate students than they really did. We asked senior researchers responding to the survey how much they worked as graduate students. While we have no way of knowing if their responses are accurate, in Figure 5 we compare their responses to the number of hours they report working in their current positions. More often than not, current faculty and post-docs say they worked more as graduate students than they do now.

Finally, we asked those 77 faculty respondents who mentor graduate students to state how many hours per week they expect their students to work. About half of them responded “40-50” hours per week, 11 said less than that, and only one indicated more than 70 hours per week.

However, if these are the true expectations of advisors, this is not always communicated to students. One graduate student respondent wrote, “We have discussed this letter at my department and though many professors side with grad students that the letter crosses a line demanding unrealistic things, when pushed they all still say that 80-100 hours isn’t unreasonable. They are hesitant to say it because they know how bad it sounds but deep down a lot actually agree with the contents of the email.” Another graduate student wrote in with an alternate view, “We discussed this in my astro department and while almost all of the professors agreed that astronomy is not a 9-5 weekday job, intelligent and efficient students should be working AT MOST 60 hour weeks… Their resounding opinion was that if you have to work 80 hour weeks to get a PhD in astronomy then you probably don’t have what it takes in the first place.”

Job prospects

Figure 6: How likely is it that a graduate student can achieve a faculty position?

Another controversial excerpt from the letter to the graduate students is as follows, “We know that you are concerned about the market for post-docs and faculty positions. Yet the market is no worse or better than it is has been for at least a decade or two.” This gets to a point of sincere anxiety for many students, the prospect that there may not be a job waiting for them no matter how hard they work.

Our survey results suggest that opinions on this subject are diverse (Figure 6). We asked “How likely do you think it is that an intelligent and productive graduate student in astronomy can become a university professor, if they so desire?” Responses spanned the gamut from very unlikely (1 on our scale) to very likely (5). The differences between groups (professors, graduate students, etc.) were not large enough to be statistically significant given the sample size, but there are hints of trends. 25% of graduate students and 30% of undergraduates said it was likely (responded 4 or 5), compared to only 9% of post-docs. Among 115 post-doc and faculty respondents, only 1 said it was “very likely” (responded 5).

The faculty and post-doc responses seem closer to the situations outlined in a recent white paper. The number of post-doc positions has grown rapidly in the past few decades, suggesting a rise in the number of students interested in permanent research positions in astronomy, but the number of faculty positions is increasing much more slowly. As one respondent summarized the issue, “Our current graduate school system is akin to a Ponzi scheme. Faculty are encouraged and expected (demanded?) to have more grad students than there are professional positions (either in the public or private sector) available.”

Figure 7: Respondents were asked about the representation of women and minorities in their department, their field, and (for non-academic respondents) science generaly.

Some respondents felt that the letter reflected a widespread sentiment in academia that tenure-track positions at research-focused universities are the only desirable outcome for a Ph.D. student in astronomy. About 50% of physics and astronomy graduate student respondents stated that their career goal is to become a professor. One post-doc respondent expressed his concern, “I know a lot (more than 20) people who have extreme talent in astronomy research. They have all left (or very nearly left) the field because of this idea we have in our community that research comes first. How are we supposed to have lives (for example, finding a spouse and having children) if our community demands 60-80 hours a week? It’s bad enough graduate school and post-docs make us 35 [years old] before we are even able to really settle down. This is to speak nothing of the slave wages we are giving people for this kind of work. What possible motivation should a graduate student have to work themselves to death when paid ~22K a year? Especially when people MUCH LESS TALENTED than them get paid 3-4 times that in private sector fields that are similar.”

Another aspect of the academic job market addressed by the survey is the representation of women and minorities in science (Figure 7). Asked whether or not they agreed with the statement that “women and minorities are represented appropriately in my field,” 303 of 425 respondents disagreed. However, this ratio was lower (221 of 400) among students and faculty asked about their own departments. While we don’t know the actual demographics of researchers at the universities of the respondents, this seems to suggest that respondents are reluctant to acknowledge when the problem of underrepresentation exists at their own workplace.

One graduate student respondent linked the problem of underrepresentation to the culture reflected in the letter, writing “We’ve received pretty similar emails from our grad student advisor which have stated amongst other things that everyone who has left the department is much better off for it/implying no one got pushed out which is very untrue. Our attrition rate for females is quite high and the department atmosphere is certainly a reason for that.” Several respondents pointed out that we should have separated women and minorities in this questioning.

Expectations for the graduate program

It deserves to be emphasized that the working conditions for graduate students and related effects on their mental health is a serious issue with significant consequences for the lives of real people. The everyday working experience of graduate students in science range from healthy and productive, fulfilling the ideal that motivates students to pursue training in research, to frankly horrific. Illustrating the latter case, one physics/astronomy graduate student respondent wrote, “It was almost a point of pride for one professor to say that his fellow graduate students were ALL single/divorced by the end of his graduate program because there was no time for family.”

The survey data is inadequate to address the many dimensions of this issue. One respondent wrote: “It seems to me that the most useful discussion that will come from this is talk about graduate student mental health. We all have to work long hours and work hard. That shouldn’t be a problem. It is a problem when the first reaction to poor performance is saying ‘you’re not working hard enough’ rather than having a discussion about what the source of the difficulty might be and how to address the problem.”

One quantitative aspect of graduate student life, which the survey addressed, is compensation. Some respondents suggested that this provides some context to the discussion of working hours. An 80 hour work week for graduate students who typically receive a stipend of $20-30k (in physics and astronomy) corresponds to a pay rate of $5.00-7.50 per hour, comparable to or below the federal minimum wage of $7.25 per hour. The national average salary for young adults with bachelors degrees is $40-50k. Among 119 faculty and post-doc respondents, 40% indicated that graduate students are paid adequately, while 60% said they should be paid more. None responded that they should be paid less.

What aspects of graduate student life do you see that can be improved? What changes would you recommend to the faculty at your institution? Do you think this change is likely to come? Please leave your responses as comments below.

About Astrobites

This post was written collectively by multiple members of the Astrobites team. Meet the authors of Astrobites.

Discover more from astrobites

Subscribe to get the latest posts to your email.


  1. The results of your survey are really interesting. It is so horrifying that people have advisers that treat them almost like slaves. This is a a large problem in my department. A recent example comes to mind – during the hurricane our university shut down due to the severe weather emergency and told all nonessential personnel to stay home. Apparently, some faculty thought it was appropriate to send emails to their students demanding they come in. The graduate student committee had to send out an email telling grad students that faculty couldn’t make them come in. They also had to send a reminder to the faculty. It is ridiculous! Why risk your safety for one day at work?

    • We had a similar situation regarding a weather emergency when a large snow storm blew through. We got a nasty e-mail from our dept chair criticizing our lack of attendance at the day’s colloquium. Ridiculous….

  2. I don’t think you realize just how incredibly helpful this was for me. The original letter has sent me into a panic over the past week or so, and when I received an end-of-semester self-evaluation from my advisor this morning I went into a panic all over again. I work a steady 70 hours a week and am not sure how much longer I can keep up this pace–so hearing an estimate of 80 to 100 gave me a feeling of “Well screw it all, I’m going into the private sector.”

    Again, thank you from the bottom of my heart.



  1. The survey results are in… « chembites - [...] Astrobites graduate student life survey fascinating.  Please take some time to read through the Astrobites article summarizing the [...]
  2. is a Ph.D. worth it? | Antipodal Points - [...] it’s making the rounds of the blogosphere; see responses on AstroBetter, Cosmic Variance, and Astrobites, focusing on different aspects…
  3. AJC Topic 12 Dec 2012 | AJC - [...] Recently, astrobites performed a survey about people’s impressions about the email and raised the question “What do we want…

Leave a Reply