Deciding to do a PhD is a huge commitment as you are just dedicating a couple of years of your life to a single cause. When I decided to do a PhD in 2006 I was not completely sure what would expect me. I just finished my master of science (MSc) at ETH Zurich and I knew that I really liked working on the research project that I chose. On the other hand I worked in programming and web development for the last couple of years to finance my studies and I discovered that only doing project/programming work is too boring for me. I had no clear picture how the PhD would be like, but I assumed that it would be a combination of a continuation of the research I did during my master thesis, writing papers, going to conferences to present your research, and in the end writing a thesis that proofs your thesis statement.
Many others before me have already blogged or written books about the life as a PhD student, how to carry out a PhD, defined what research, yet I think that I can add my two cents to the existing stories. My experiences are still fresh (I started my PhD in 2006, defended beginning of May 2012, and started my PostDoc here at UC Berkeley in September 2012) and I hope that I can give a good description about the qualitative aspects of a PhD.
The three phases of a research PhD
In my time as a PhD student at ETH Zurich I had uncounted discussions with peers, assistant professors, professors, researchers, and people related to academia. During these discussions we discovered that a PhD can be segmented into three basic phases: the first phase identifies the research topic and bootstraps the student; the second phase is the productive phase where the research bears fruits and is turned into papers (or at least communicated to other people); and the last phase consists in wrapping up, writing the thesis, and finally the defense. Every successful PhD student goes through three phases during his PhD. Each phase has different requirements and relies on different personality aspects. The phases follow after each other but there is no clear transition from one phase to the other yet one knows if the transition completed.
The search for the holy grail: ramping up and finding your topic
The first phase of a PhD is like the 'get to know each other' phase at the beginning of a relationship. You don't really know what to expect but you are interested and you want to explore. Yes, there are some peculiarities but you are not too concerned about that right now. In this first phase you dig into the field that is interesting to you and you start reading up about prior research approaches about the problem you want to tackle. Some ideas start to form in your head and you decide on a topic for your thesis. In more practical research areas you start building prototypes and you execute small tests; in more theoretical areas you try to come up with some theorems and you have a rough idea about the general topic.
Depending on your adviser he or she will restrict your search in one way or another. The most restrictive adviser tells you exactly what you should do (sometimes they have a grant proposal for a specific project and/or only need a code monkey to implement their grand idea) other advisers let you run free and let you come up with your own idea in a (sub-)field. Both extremes have their advantages and the kind of adviser that is best for you is basically an optimization problem that you need to solve for yourself.
On one end of the scale you are given a topic by your adviser. This approach ensures that your adviser cares/should care deeply about your project and it gives you a head start, on the other hand you'll have to make up for this pre-selection and you'll have a harder time to bring in your own (research) ideas into the already existing project. On the other end you can choose your topic in an open void. It is up to you to make your adviser care about the project and you need to come up with the core idea. This great opportunity usually results in additional time needed to complete this phase. In the real world your adviser will usually set the limits anywhere between these two extremes.
From my experience I would say that the average student stays in this phase from one to three years. I was lucky enough to have an adviser that allowed me to choose freely in any field that he was comfortable with. During my first three years I discussed a huge set of possible research ideas with him and even after we had settled on a specific topic the details kept changing as I passed through the later phases. Two key aspects in this phase are: (i) creativity, you need to come up with a good research idea and research plan. Nobody will tell you what to do, there is neither a customer who demands a specific feature, nor is there a pre-set plan like in a play. You need to be focused and keep yourself together to pass through this phase. (ii) Due to the fact that you need an idea and you are responsible for the schedule you build up a huge amount of psychological pressure. Most people that drop out during this phase fail due to either a lack of ideas (and creativity) or because they are unable to self-organize themselves in an unstructured work environment.
The paper mill
After you have decided on the (main) topic of your thesis and built the first prototype (or come up with the first layout of the system) you gradually transition from the first phase to the second phase. You are now trapped in the paper mill. This is the most productive phase as you try to publish your work in as many (good) conferences as possible. The quality of your work is (somewhat) measured in the number of papers that you publish at top tier conferences. You are evaluated based on the quality of your work combined with the human factor how you present yourself.
In this phase you start to become an expert in your field: you have claimed a little spot in a bigger field that you extend with your research. Your adviser will try to keep you in this phase as long as you are able to produce more good papers (given that your adviser has enough funding and that you do not run into any hard time limits given by your university).
The core qualifications for this phase are that you can (i) work hard and (ii) present yourself at conferences. You have to write all these papers and at the same time you should refill your pipeline of ideas to get the next publications in order. In addition you have to express all these ideas to others and you have to play the social game at conferences. Building your social network during conferences is actually hard work and you need to try to make friends with some of the bigger shots in your field as this can be helpful for future collaborations, additional papers, or even reviews. If other people understand your ideas from beginning to end and you can convince them (in person) that the ideas are great then they are more likely to accept your paper that they'll review later on. This (productive) phase usually lasts about two years and is your opportunity to build your social network for possible future positions. A side quest during this time is to select your thesis committee. You already know in what area you will write your thesis about, you should have an idea about your thesis statement, and you should know who the experts are in this area. Hint: conferences are a good place to chat up possible committee members!
The exit strategy: wrapping up and writing your thesis
After you churned out a bunch of papers during your time on the paper mill you are finally ready to graduate. At one point in time you start to feel ready, you have published a couple of papers (or at least written a couple of technical reports if you weren't able to publish the papers) and you developed a good understanding of what research in your area is about. In your (regular or not so regular) meetings with your adviser you should also get a feeling if he or she thinks that you are ready.
At ETH you write your thesis from scratch. I started with reading (or at least skimming) my prior papers before I wrote my thesis. After that I came up with the basic outline and the vision that my thesis should get across. The goal of the thesis is to write down your thesis statement and to prove said statement. It is important that your thesis is a self-contained and self-consistent document.
In this phase you should have settled on your thesis committee. Your committee consists of a couple of professors (or people with a PhD depending on your university) that are experts in your area of research. These experts will read your thesis and grill you during your defense. Some people quit at this late stage of the PhD due to timing constraints or due to burn out. After a couple of years they are worn out and run into deadlines imposed by their adviser or by the university. It makes me sad when I see colleagues that come up to this point in their career and then quit so close to the finish line.
The defense is the end of this last phase and after you have passed this final test you have almost completed your PhD. At most universities there are some bureaucratic hurdles that are left for you to take and there might be some minor (or major) revisions to your written thesis that you have to carry out. Other than that you are done and at some delta t after the defense you are allowed to call yourself PhD (or Dr. Sc in my case).
The teaching aspect in a PhD
Teaching is an orthogonal experience to the research experience. At ETH you are supposed to be the teaching assistant (TA) for one lecture per semester. At the beginning you will start off as a regular TA but as you progress in your PhD you will be trusted with more and more responsibilities, e.g. you will teach lectures if the professor is sick, or you will be head TA for a complete lecture. Head TAs are responsible to coordinate the lecture, to supervise the other TAs, and to organize the individual exercises.
I really enjoyed teaching during my almost 5 1/2 years at ETH and I think that it is a great experience that every PhD should have. During the exercise hours that you have to supervise you learn how to speak in front of a group (a skill that is not that common in Switzerland and that is not part of the regular curriculum at high school or college), you learn how to prepare the material so that you can present it to a group of students, and most importantly you learn how to interact with all kinds of students that have different problems with the material, the course, or a specific exercise. All in all you learn a new skill: how to teach some specific material to students. If you plan an academic career then this might come in handy at one point in time.
The PhD grind
Right when I was finishing my thesis Philip Guo published his memoirs about his own PhD experience in an e-book called "The PhD Grind". In the book Philip tells us about his PhD experience year for year in chronological order. Each year of his PhD is covered in a chapter and he writes about all the ups and downs during that time. The book is a great read and I warmly recommend that you read it too if you are interested in research and academia.
I agree with many of his observations, e.g., that it is important that you focus on conferences that your adviser already has published in recent years. Your adviser will know the right lingo and the right buzzwords for the conference and the members of the committee will know your adviser which in turn will help as well.
On the other hand there are a couple of things where I don't agree with Philip. Most importantly Philip mentioned that being a TA only delays the PhD (page 70, of the e-book). As I explained above I really valued the teaching experience - from a personal as well as from a professional point of view. During the discussions with the students I learned several new and interesting details in areas that I actually assumed to know in depth. Another area I don't agree with Philip is about how to select your thesis committee. Philip tells us that usually the adviser selects the thesis committee (page 56, of the e-book), I on the other side had the pleasure to select my own committee. I discussed possible members of the committee with my adviser and then approached each member myself. For me this was an interesting experience - first coming up with possible and plausible fits and then approaching these professors myself.
A PhD is certainly not the right career path for everyone. Doing a PhD is challenging, requires a lot of self-control and self-organization. The PhD is not (pre-)structured and you will need a lot of creativity. In addition you get less pay than an industry position with a comparable education. On the other hand you learn a huge amount of new skills during your PhD. This is your opportunity to do academic research, you can publish at conferences, you'll meet a huge amount of new people, and you learn to network. I enjoyed my time and I recommend doing a PhD to all the curious and interested people out there that are interested in academia and who want to go that extra mile.