I will never forget the day I submitted my PhD for examination. Having spent most of the previous night proofreading the final draft, I rose early that morning, excited by the prospect of moving on from four long and often frustrating years as a postgraduate.

I had envisaged being overcome by a mixture of joy and relief when the dissertation was finally handed over, but as I carried my PhD in loose leaves to the university's bindery, it slowly dawned on me that submission was not the end of my postgraduate story. It was just the beginning of a new and particularly anxious chapter: the viva.

Viva (or viva voce), from the Latin "in living voice", is the title given to the oral examination of certain advanced third-level degrees. The term can be used for the assessment of some masters, but generally in the UK it refers to a PhD's oral assessment. A viva has two examiners, one internal to the university and one external, except when the candidate is in the university's employ, in which case there are two external examiners. There are a range of outcomes, from a pass with major, minor or no corrections to an outright fail.

The length of time between submission and viva varies. Three months is the norm, although it can take anything from one month to a year. This interim period is a rich breeding ground for lurid and often apocryphal tales of viva disasters. Year-long waits for a viva date, examiners sending out for a Chinese takeaway during nine-hour ordeals, candidates given two years' worth of corrections to make: in the weeks leading up to my viva, I heard them all.

ÖйúAƬ

Thankfully, I did not become another folk tale. Within a month of submission I had my viva date. Six weeks later I breezed out of the examination room smiling broadly, thanking my examiners and with only a handful of typos to correct.

Involving two hours of friendly and enthusiastic questioning, my viva was a far more enjoyable experience than I ever imagined it could be. But because I had little idea of what awaited me, the weeks leading up to the big event were marked by uncertainty.

ÖйúAƬ

My experience is not atypical. For Malik, a PhD candidate in sociology, the entire process was shrouded in mystery. "I was given precious little guidance before the exam. Nobody ever really said 'this is how a viva works' or 'this is what to expect'," he says.

"The only clue you have is talking to other people, but their experiences seemed so varied that I really didn't know what would happen."

That the viva appears almost inscrutable to most PhD students is hardly surprising: in many UK universities there is a marked lack of reliable information about it and what it entails.

So what can a candidate do to counter the viva's seemingly inherent unpredictability? It appears that most prepare in similar ways. The dissertation is dusted off, new summary notes are made to guide revision, and websites such as Andrew Broad's Nasty PhD Viva Questions () are consulted to identify common lines of questioning and formulate answers in advance.

Malik, who earned a book contract with a leading academic publisher on the strength of his PhD, recalls making copious notes before his examination. "I arrived with my copy of the dissertation covered in Post-it notes. Whether I'd read it that thoroughly my examiners weren't to know, but when I laid it down on the table they both took a deep breath. I thought, 'first battle won'."

Preparation is definitely crucial to a successful viva, but being too organised has its perils, as Malik soon found out. "From looking on the internet, it seemed that 'what is your thesis?' is a really common question, so I decided to learn off my abstract. But when they asked the question, I stumbled. My voice was two semi-tones higher than it should have been and I tripped over all the words. The external examiner just looked at me and asked: 'Do you want to try that again?'"

Choosing the right examiners is crucial for a successful viva. The decision to award the degree rests entirely with the internal and external examiners, although the latter often drives the viva and has the deciding say in whether a candidate passes or fails.

Ensuring a good match between student and external examiner, known in PhD circles as "externals", is of paramount importance. The selection process normally involves both the candidate and the supervisor, and takes place about six months before submission. Selecting an expert in the field often means an alliance of opinion, increasing the likelihood of a shared intellectual outlook and a less stressful viva. Thus, it is probably best not to ask a Marxian scholar to mark a post-structuralist dissertation.

ÖйúAƬ

Examiners may be chosen for less overtly academic reasons, too. For Simon, a PhD candidate in ecology, a personal connection, however tenuous, was the deciding factor. "I knew nothing about my external. My supervisors had met him at a conference and said he was all right, so I went with him," he says.

A well-suited external can give the candidate a viva that is intellectually and personally stimulating and an ally in the post-PhD academic world. Simon, who is looking for work, now rather regrets his choice. "I'm still in touch with my external and he writes references for me. He's been very helpful, but with the benefit of hindsight I might have chosen someone more eminent with a few more connections."

Generally speaking, externals do not look to fail PhD students, but an injudicious choice can have a disastrous impact. Although the bulk of PhDs are awarded with relatively minor corrections, not all vivas go so smoothly. Some students are asked to undertake rewrites that can take up to a year - or may even fail outright.

Most viva horror stories stem from a mismatch between the external and the dissertation. Lisa, a cultural geographer who now lectures in New Zealand, found this out to her cost. "It really knocked my academic confidence," she says.

Her problems began when the academic originally earmarked for the external role suddenly became unavailable. "The person my supervisors and I had wanted pulled out due to work commitments. A replacement was hastily arranged, but less thought had gone into his selection," she recalls.

The replacement external's academic background was markedly different from Lisa's, which caused major problems. "The external railed against seemingly minor issues pertaining to the description of my field site in the introduction, occasionally brandishing books he had brought with him in order to query why I had not cited such purportedly vital texts. I was not questioned about my thesis, but only on minor details that were superfluous to the overall body of work.

"I was left with no idea if my thesis was of publishable quality, if the arguments contained within it were sound, only a realisation that it was inadequate to speak of rocks as rocks: I should refer to 'limestone aggregate'."

The external's opening remarks can set the tone for the entire viva. Positive, non-adversarial observations generally signal a smooth examination. On the other hand, testy, confrontational comments can be harbingers of a difficult defence, as Nigel found out at the beginning of his viva for a PhD in sociology.

"The first question they asked was: 'What is a thesis?' The external clearly wasn't happy with the dissertation, so with that question he was telling me where he stood."

Although the examination lasted more than three hours, a recalcitrant examiner gave little opportunity for constructive discussion. "The external had already made his decision. I wasn't really given a chance to defend my work at all. Every explanation I offered he disagreed with.

"It was as though the result was predetermined before I walked into the room. They might as well have written it down on a piece of paper, given it to me and dispensed with the whole discussion element," says Nigel, who is still correcting his PhD. He will have to resubmit and sit another viva.

By their very nature, vivas will always be stressful and challenging, but the presence of a supervisor in the examination can often be useful. Although some students fear that being accompanied could be seen as a sign of weakness, increasing numbers are exercising this right. As well as protecting against an unfair examiner, the supervisor - who is a silent observer and sits out of the candidate's eyeline - can act as a chronicler. They are free to note everything that is said, which is particularly useful for recording key points and arguments that can be developed into published papers later.

C.P. Snow's well-worn distinction between "two cultures" - the sciences and the humanities - seems to be borne out in the viva. According to Simon, "science vivas are generally factual, full of questions with right or wrong answers. In the social sciences, there seems to be more potential to have a serious disagreement."

In the humanities viva, where the potential for examiners to hold wildly divergent interpretations of a dissertation is arguably greater, hermeneutics often plays a major role. In his examination, Leslie, who holds a PhD in historical geography, found himself navigating between contradictory and conflicting readings of his work.

"My examiners regarded my thesis in entirely different ways - one historical and detailed, the other socially focused and interested in generalities.

"You can never quite prepare for how someone else will interpret your work. That day there were really three theses being discussed: my own and the examiners' interpretations of it," he says.

Science vivas generally place more emphasis on elucidating facts and figures than on defending dense, complex arguments. Not that this makes them any less problematic: seemingly minor omissions or typos can fundamentally alter the meaning of crucial figures and results, although attentive proofreading can go some way to alleviating this concern.

Across all disciplines, thorough preparation is the key to a successful viva. But is there anything that universities should do to improve the process? Leslie believes ÖйúAƬ institutions need to offer more support and information in this area.

ÖйúAƬ

"Given the number of transferable-skills courses universities run, surely they could offer one on viva preparation?" he says.

Such courses could help to clear the air of mystery that has obfuscated the viva. It is also vital that, as far as possible, the viva horror stories circulating through postgraduate offices up and down the country are consigned to legend.

On the back of her particularly negative experience, Lisa feels that the presence of a non-aligned third party should be made mandatory. "At present, most institutions do not allow for the presence of a mediator. I think this should be considered, so that the internal examiner is not forced to adopt a mediating role to counterbalance the external's excesses," she argues.

Regardless of what universities do to improve the information on offer and protect students from potentially unfair assessments, there is plenty that PhD candidates can do to help themselves.

As in so many other facets of life, a positive outlook is indispensable. Walking into the exam room with confidence can make all the difference. And, as Malik notes, candidates must always remember that when it comes to their work, they are the experts.

"You have one really crucial advantage over your examiners - you know your PhD better than they ever will. Never forget that."

So what advice do examiners themselves have to offer? "The most obvious thing is to read and think about the thesis, especially its broad aims, subject matter, arguments and achievements," says Catherine Nash, reader in geography at Queen Mary, University of London, and an experienced viva examiner.

Examiners want evidence of critical thinking. What areas would the candidate revise or develop if they were to write the PhD up for publication? What are its strengths and weaknesses? How could the dissertation be improved? Candidates must be able to give coherent answers to these questions.

It is easy for nervous and overwhelmed candidates to become defensive when asked about their work's weaknesses. They should, Nash cautions, avoid aggressively "defending" their dissertation: probing questions more often signal interest than admonishment.

"Many of the issues examiners raise will not be problematic ones, but ones they feel are very interesting, that they want to hear more about."

Playing an active role in an open, constructive discussion is the best route to viva success. Candidates should be honest about the limits and failings of their work, she advises, although obviously without denigrating their efforts too much.

"Be critical and engaged rather than simply negative about your work. You don't want to change the examiners' minds if they are planning on passing you."

John A. Finn, author of Getting a PhD: An Action Plan to Help You Manage Your Research, Your Supervisor and Your Project, believes that candidates must recognise that the viva is markedly distinct from submission. "The face-to-face and immediate nature of the viva makes it a more intense emotional experience," he says.

A properly structured and presented dissertation is crucial. Walking into the examination with examples of published work helps, too. "While many universities do not formally require that parts of the PhD thesis be published, it is expected that at least some of the research be publishable. If you have had a paper or two accepted by an international peer-reviewed journal, it makes an external examiner's decision much easier," Finn says.

Even with a raft of publications, however, a candidate may still feel unable to answer a question in the viva. What then? Nash suggests first pausing and reflecting on what the examiner is asking. If that fails, request clarification.

Still stumped? "Then it is best to say something like: 'That is a very useful question/comment and one that I have not thought about before, but will certainly reflect on in the future.'"

In the UK, Ireland and Hong Kong, PhDs are examined in private by examiners with no substantial prior contact with candidates' work. But these protocols are by no means universal: elsewhere, defences can be carried out in public, involve lectures or may not even take place at all.

For example, North American universities use a committee system: each PhD candidate assembles a panel, typically between three and five people, who both supervise and examine the research. Committee members are chosen by the student in conjunction with the primary adviser, and one member must be drawn from a different department.

Conducted by the candidate's committee, the PhD oral examination is open to the public and typically involves a short presentation outlining the aims, methodology and results of the project, followed by one to two hours of questioning.

Jane, a PhD graduate in anthropology from the City University of New York, remembers her viva well. "I had four people on my committee: my adviser, an associate professor from our department, a sociologist and an environmental psychologist.

"I gave a PowerPoint presentation about the main points of my PhD and afterwards they took it in turns to ask questions. I was out in just under two hours, which is probably a bit shorter than most."

The Danish model incorporates North American elements, including the public defence and the committee. But instead of a presentation, candidates deliver a 45-minute lecture on their PhD to a much larger group than is common in the US or Canada.

"My entire family was there: parents, in-laws, husband, kids, staff and students from the department. There must have been 200 people," says Trine, who defended her PhD at Aalborg University.

As in North America, failures or serious revisions at the defence stage are extremely rare in Denmark. Although it is not encouraged, candidates can submit without the approval of their supervisors in the UK; under a committee system, adviser sanction is the sine qua non of a PhD's progression, increasing the likelihood that the work is of sufficient quality.

The momentum built up behind the candidate by the time the viva arrives seems to be another factor behind the comparatively low failure rates in committee defences. As Trine remarks: "I had already rented out the hall for the after-party. All my committee members were coming, along with my family, colleagues and friends. The defence was long and gruelling, but deep down I knew I was going to pass. How could they fail me and then come out with my friends and family after?"

In most established Australian universities, the situation is different again - there is no PhD oral defence at all. Instead, dissertations are assigned to two external examiners, generally one national and one international, who read and mark the work. The candidate responds with any revisions requested without conversing directly with the examiners.

With experience of both the Australian and British doctoral examination systems, Jane Jacobs, professor of cultural geography at the University of Edinburgh, is well placed to judge the relative merits of different approaches to the viva.

"The viva in the UK is truly a defence: the examiner can go in sceptical about the quality of the research in the dissertation, but in the course of the viva can be convinced about the veracity of the work," she says.

ÖйúAƬ

"In the British system, examiners can be genuinely swayed by the intellectual and research capabilities the student displays in the viva conversation."