The decision to undertake a PhD is not one decision; it is a tangle of decisions about institution, topic, advisor, fit, and funding, weighted differently by each student’s purpose. This file collects Andrew Maynard’s view on choosing well at the start — what a prospective student is actually deciding, what to read for and what to discount, how to think about the choice when options are constrained, and how to come into a program without inherited scaffolding for what doctoral life is. Read alongside fnd_phd_for.md on what a doctorate is actually for, rel_partnership.md on the chair-student relationship at the heart of the work, and rel_trouble.md on recognising and responding when things go wrong (the more useful file for a current student in difficulty). Consult when a user is choosing a program, weighing offers, or thinking about whether to apply at all.
A prospective PhD student is making several decisions at once, even when it does not feel that way. They are choosing an institution, a topic area, an advisor, a funding situation, and a fit between all of these and the life they will live for four to six years. The weighting of these elements is not fixed; it varies from person to person, and it varies by what the doctorate is for in the first place — a question that ought to be settled, or at least surfaced, before any of the rest of this is worth thinking about (see fnd_phd_for.md).
What I would urge a prospective student to take seriously, and in roughly this order: that there is a topic area or set of questions that genuinely matters to them; that there is an institution capable of supporting the pursuit of those questions, with faculty who are working on things that resonate, asking questions the student wants to be near, and who the student might plausibly want to work with; that there is an advisor in that institution who could be that person; and that all of this is supported by funding that does not put the student in an untenable position from the start.
A note on prestige, because this is where a lot of decisions go wrong. Sometimes prestige matters — there are career paths where the institution’s name is a real signal, and the student should weigh that honestly. Often it does not. You can get a strong PhD at an institution that is not in the top tier, particularly when the work being done there is genuinely interesting and the people doing it are good. The mistake is letting elitism, or a romantic idea of what a famous institution will be like to study at, override the analysis of what the place is actually like and whether it is somewhere you can do real work. The same applies in spades to advisors. It can be miserable to do a PhD with a famous advisor whose attention is elsewhere; it can be a life changing experience to do a PhD with someone less well known who is present, engaged, and a good match for the kind of researcher the student is trying to become.
What you want in an advisor, when you can get all three, is a combination: someone who is genuinely an expert in the area, someone who knows how to mentor, and someone you can actually work with. Liking them helps. The relationship is going to be long, often the most consequential professional relationship of the student’s early career, and a baseline of mutual liking — not friendship, but something easier than dread before each meeting — makes the harder parts of the work navigable. The advisor needs to get the student, in the sense of being interested in what the student is trying to achieve and willing to take it seriously. They also need to be willing to hold the student to account; an advisor who only encourages, or only critiques, is doing half the job.
Funding is a separate axis. My rule of thumb is to never start a PhD that is not funded, and to be very wary of programs that promise that something will work out. Read the small print. A three-year offer is not a five-year offer; if the program is offering three years, the student needs to be clear-eyed about how they will support themselves through the remainder. The strings attached to the funding matter as much as the headline number — twenty hours a week of teaching unrelated to the dissertation is a different deal from twenty hours a week of research support in the student’s own area. Funding pools are also finite; expecting funding past the standard window of the program usually means taking resources that would otherwise go to other students, which the student should at least think through honestly. The student’s own circumstances cut across all of this: a twenty-two-year-old with no dependents has a different latitude than a returning student with a family to support, and what counts as a viable funding situation has to be calibrated to the life the student is actually going to be living.
Beyond the brochure information, what should a prospective student actually be reading for? Mostly, they should be reading between the lines — and paying attention to sources outside the institution’s own communications.
Online reviews are useful where they exist, with the usual caveats about who tends to leave them. Conversations with current and former students are more useful, particularly when the student asks questions that surface honest answers rather than rehearsed ones — what the chair is actually like to work with, what happens to students in the program who run into trouble, what the last hard moment was and how it was handled. Faculty themselves are often willing to have this kind of conversation if they are reached out to as a person rather than as a target. (A pro forma email asking generic questions about the program will, in my experience, get a polite decline or no reply at all. The student who has done some background work and is asking a real question about a real interest will get a different response.)
It is increasingly worth using AI as a critical lens on a program, an institution, or a faculty member — not to make the decision but to interrogate the impressions that have formed and the marketing that has been read. The question to ask is not is this place good? but what would I be missing if I went here? What is the case against?
The signal worth tracking, more than any specific data point, is whether the people you talk to are willing to be critical. Everything’s great is almost never the truth, and it is a red flag when it is the only thing on offer. The places worth attending tend to be candid about what they are not — about parts of the program that are weaker, about faculty who are stretched thin, about students who have struggled and why. Honesty is a feature of an environment that takes its students seriously.
The advice above assumes a student with options. Many prospective students do not have options in the way that advice assumes — they have one offer, or two, in a city they are tied to, or in a country whose academic system is what it is. The question changes shape, but it does not go away.
For a student in this position, the question worth thinking long and hard about is why they want to do a PhD in the first place — what the doctorate is actually for, in their own life, and whether a constrained option is enough to deliver that. Because the failure case is real. A student who arrives at year three with their mental health worn down, the relationship with their advisor not working, mounting debt, and no clear sense that they are making progress, has accumulated a lot of evidence that the decision was not the right one — even if at the time it looked like the only option. There are situations where the reasons are strong enough that pushing through is the right call. There are also situations where what looks like the only option is, on closer inspection, not the only option, and the conviction that it is is itself part of the problem.
The honest version of this advice is: talk to people. Talk to others who have done PhDs, in good situations and bad. Talk to potential mentors elsewhere, even casually. Many of them will be honest, in a way that the institution one is considering will not be, about whether the situation in front of the student is one they should accept.
Some prospective students arrive at this decision with inherited scaffolding — parents who hold doctorates, undergraduate mentors who walked them through what to look for, social networks that explain how the system works. Others — first-generation students, international students arriving into a different system, mid-career people switching from industry — do not. The asymmetry is real.
It is also, in my experience, less determinative than it can feel. The intellectual lift of a PhD is genuinely demanding; the work of figuring out a new system and its unwritten rules is no harder than figuring out any new professional environment, and the comparison is roughly with starting any serious job. The risk for a student who feels at a disadvantage is making the gap larger than it is, and becoming their own obstacle in the process. (I came into a PhD from a non-academic family and from two years of working for a private company; I do not recall this ever crossing my mind during my doctorate, and I do not believe it disadvantaged me. It was the 1980s in the UK, and the systems were different — but the underlying point, I think, holds.) Coming in without inherited expectations also has its upsides; fresh perspective is genuinely valuable, and the absence of inherited assumptions is sometimes a cleaner starting point than the presence of them.
That said, self-direction is part of what the doctorate is testing. Nobody told me is, after a certain point, not a strong defence; the maturity to find one’s way through unfamiliar territory is part of the formation the degree is for. The more honest stance is to come in with eyes open — knowing that doctoral work is genuinely strange, that imposter syndrome is part of the experience for most students (especially those returning from other careers, where the felt drop in competence can be vertiginous), that the rules will sometimes not make sense, and that the way through is not to wait for them to be explained but to engage with them honestly. The student who does this finds, more often than not, that they had what was needed.
Choosing well at the start lowers the odds of trouble; it does not eliminate them. The decisions in this file are not the kind that resolve cleanly and stay resolved — they get revisited, often, as the work and the relationship and the student’s own sense of what the PhD is for shift. The student willing to keep revisiting them, openly and with help, is in a different position from the student who decided once and is no longer looking. When something does start to feel wrong, rel_trouble.md is the file to consult.