«The Unwritten Rules of PhD Research Gordon Rugg Marian Petre THE UNWRITTEN RULES OF PHD RESEARCH THE UNWRITTEN RULES OF PHD RESEARCH Gordon Rugg and ...»
There are lots of words and phrases which indicate weakness. Some of these are of a type usually known as ‘weasel words’, whose purpose is to help you wriggle out of committing yourself to an assertion and substantiating it.
Weasel words usually have no place in academic writing, and certainly not in a dissertation. It is not normally good enough to say that something ‘seems’ to be something else: is it or is it not? Similarly, ‘probably’ isn’t good enough.
Weasel words suggest that the author has not looked hard enough, or is making speculations which cannot be substantiated.
Whenever you justify something, you raise the question of why it needed justifying in the ﬁrst place, and whether your justiﬁcation of it was good enough. Anything which involves assumptions is also leaking blood (e.g.
‘presumably the respondents believed that...’) Words such as ‘probably,’ ‘presumably’ and ‘must have’ are another way of saying, ‘I have no ﬁrm evidence for this, and am guessing’. If you have evidence, present it; if you don’t have evidence, and the issue is important, then get evidence and ﬁnd out whether you are actually right in your guess. Speculation is something which should either be explicitly labelled as speculation (and therefore of only WRITING STYLE 113 tangential relevance) or saved for the closing stages of the discussion section where you are discussing future work, or preferably both.
Vagueness is not acceptable. From an actual examiner’s report:
In an academic argument the details should all be nailed down, as far as this is possible. Often it is best to omit things of which one is unsure.
If this is possible they should not be present anyway. If it is not possible, they must be established deﬁnitively. Otherwise the conclusion will inherit the lack of precision. Then the whole work may simply become a piece of unproven speculation, which is unacceptable for a doctoral thesis.
One thing worth watching out for in this context is the temptation to use public-domain principles as explanations. The reason for this advice is that these principles are often untrue or seriously misleading. This is one case where the internet is positively useful – the sci.skeptic and alt.folklore.urban newsgroups are rich sources of widespread beliefs which have no basis in truth. Make sure you have a proper academic source for any explanatory principles you want to use. If you can’t ﬁnd one, then it might be because the principle just isn’t true...
Don’t bluff If you don’t understand something, make time and work on it until you understand it. If you try to use technical concepts without understanding them properly, the critical reader will spot it instantly.
One apparent exception to this is when you are using advice about specialist tests (generally in the context of choice of statistical test). The normal convention in many disciplines is that in such cases, where the expertise is outside what a researcher in the domain could be fairly expected to know in detail, it is acceptable to take advice from two or more independent authorities in the relevant area and follow that advice. If they get it wrong, then that isn’t your problem, because you’ve taken reasonable steps. You’ll still need to understand what the statistical results mean, though.
Academic style: an example
The example below is from a paper by one of the authors and a colleague. It has been chosen speciﬁcally because its topic will be unfamiliar to most readers, making it easier to demonstrate the way in which language is being used at both the explicit and the ‘between the lines’ levels.
114 THE UNWRITTEN RULES OF PhD RESEARCH Direct evidence for hemispheric asymmetry in human and protohominid brains is in principle obtainable via endocranial fossil remains.
Unfortunately, the number of suitable skulls is limited, particularly for older species. Much the same problem applies to attempts to infer handedness from the weight and size of hand bones, on the principle that a more extensively used hand will be more robust (and therefore have larger bones) than the non-preferred one (Roy, Ruff and Plato 1994). There is evidence of such asymmetry in the long bones of the upper limb and shoulder girdle of Neanderthal specimens (Trinkhaus, Churchill and Ruff 1994; Vandermeersch and Trinkhaus
1995) but the paucity of suitable pre-Neanderthal material limits the scope for this approach.
There are several speciﬁc types of coded language in use here. These include the following.
Technical terms signal membership of the relevant research community:
‘hemispheric asymmetry, protohominid, endocranial, robust, non-preferred, long bones, upper limb, shoulder girdle’. ‘Robust’ is a technical term, the converse of ‘gracile’. ‘Long bones’ is also a technical term, contrasting with ‘short bones’.
References References signal familiarity with relevant literature: (Roy, Ruff and Plato 1994); (Trinkhaus, Churchill and Ruff 1994); (Vandermeersch and Trinkhaus 1995). All three are from within six years of the submission date of the article, and all are specialist journal articles.
General academic language General academic language signals membership of the general academic community: ‘in principle, via, much the same, infer, there is evidence of such, paucity’. Note how speciﬁc claims about inferring handedness via speciﬁc methods are supported by references, while a broad statement about lack of suitable fossil material is not supported by a reference. Within this research community, the lack of suitable fossil materials is a generally agreed truth which does not require a supporting reference; the authors have already demonstrated their membership of this research community via their familiarity with its literature and technical terms, so can make the broad statement without a supporting reference. Writing for a different research community, a supporting statement might well be needed (for instance, if the different WRITING STYLE 115 research community had not reached a consensus about the lack of suitable fossil remains, or was completely unaware of the issue).
Note also how an entire approach is described and rejected in two sentences.
The authors throughout assume that their readers will be familiar with a range of technical terms relating to physiology, such as ‘endocranial’ – this assumption is made because the paper is for publication in a journal catering for a well-deﬁned research community (laterality researchers). This assumption reduces the need to unpack and explain terms; this in turn means that the writing can be terse and efﬁcient. If the authors deal with terms unlikely to be known to the readers, then it is necessary to explain each term on the ﬁrst occasion when it is used (as happens later in this paper when the authors describe ﬂint artefact manufacture).
Writing as expressive behaviour Style and voice – and indeed the selection of content – do have a role in expounding your central argument, but they have an even larger role in conveying what sort of author you are.
Bad things to communicate are:
• I am ignorant, clueless and in despair;
• I am lazy, dishonest and rude, and I deserve to be hanged and ﬂogged.
Good things to communicate are:
• I know what I’m doing;
• I’m a professional with the right attitude.
How do you communicate these things? They vary in easiness.
I’m a professional with the right attitude (easy concepts, but hard work)
• I pay attention to detail in things like spelling and the layout of the references
• I’ve done all the work I should have done, and demonstrated this in the write-up
• I’ve done a meticulous job of work and demonstrated this in the write-up
• I’ve presented this work neatly and exhaustively, following the conventions of this area 116 THE UNWRITTEN RULES OF PhD RESEARCH I know what I’m doing (requires knowledge of your chosen ﬁeld and hard work)
• I know all the key texts, have read them and have cited them correctly
• I have read other relevant things as well and have cited them correctly
• I know and understand the technical concepts in this area and have been careful to use all the relevant ones somewhere in my write-up I am ignorant, clueless and in despair (all too easy to communicate, especially in the second year of a PhD)
• I have not read the key texts
• I have made classic mistakes without even realizing it, and have not had the wit to show my draft to a reliable mentor who would have spotted them ages ago
• My work contains apologies and pleas for mercy I am lazy, dishonest and rude, and I deserve to be hanged and ﬂogged (effortless, if you’re a sinner)
• My work contains hardly any references
• My references are all from the internet, popular magazines or textbooks
• My spelling, grammar and presentation are dreadful
• I have attempted to conceal the dreadfulness of my spelling, grammar and presentation with jokes, clip art, a fancy binder, coloured pie charts and a grovelling acknowledgement to my supervisor
• I have misspelled my supervisor’s name and got their title wrong in the grovelling acknowledgement to them
• There is no evidence in what I have written that I have done any work
• Some paragraphs of my text are much better written than others, and bear a strong resemblance to articles on the internet
• I have done things which my supervisor speciﬁcally told me not to do
• My text compares theory and academia unfavourably with the ‘real world’ but I have not put my money where my mouth is and gone away into the ‘real world’ If these things apply to you, then you will probably not even be allowed to start a PhD – the initial selection process will almost certainly detect you and hurl you into the outer darkness (though you might manage to bluff your way onto an MSc, only to be failed when your dissertation erupts onto an unwilling world). It is unlikely that anyone ﬁtting this description will be conscientious enough to bother reading this book anyway; we have included this section largely as a reassurance to virtuous but insecure students, so that they know they aren’t as bad as they sometimes fear.
WRITING STYLE 117
Getting it right
Avoiding mistakes is all very well, but what can you do that is positive?
One useful method is to go through your text with a highlighter, highlighting any words or phrases which would not be familiar to the average person on the street. This is particularly useful for your ﬁrst page, and especially for the ﬁrst paragraph, where ﬁrst impressions count. An alternative is to delete any sentences which fail this test and see what is left.
Here are two examples. The ﬁrst is ﬁctitious, to protect the guilty, but closely based on horrible experience with MSc students. The second is real, to reward the virtuous, and comes from one of our MSc students. Both deal with the design of good web pages for commercial sites on the internet.
Example 1 The Internet is the fastest-growing technology that the world has seen.
Its now possible to download movie clips and the latest music via the Internet from sites across the world, and to watch events like a Shuttle launch happening in real time. The Internet clearly offers many opportunities to a company to advertise its products around the world for a fraction of what that would previously have cost. However, with so many companies competing for attention on the Web, it is particularly important for companies to have Web pages which are eye-catching and memorable, and which convey the right impression to the customer.
Otherwise, even if a customer does view the organization’s Web site, there is the risk of conveying the wrong impression and losing sales and wasting money in the process. Although the design and advertising communities have considerable experience of doing this sort of task via traditional media, it is not safe to assume that the same principles apply to Web page design as to printed page design, and this area clearly needs to be researched.
Example 2 At the core of software engineering lies the issue of software quality and this has resulted in the growth of the software metrics ﬁeld of research.
Software metrics is based on measurement theory and there exists a comprehensive literature relating to both measurement theory and software metrics, [Fenton and Pﬂeeger, 1997], [Kitchenham et al., 1995], [Hall and Fenton, 1997]. Traditionally software metrics research has been in relation to application software. However, the rapid expansion of the Internet and corresponding growth in the number of commercial web sites [Rupely, 1997], [Jones, 1997], has given rise to a new set of problems relating to the development of metrics for a new class of software products including web pages, epitomised in the question ‘How good was my web site?’.
118 THE UNWRITTEN RULES OF PhD RESEARCH Numerous guidelines have been published with regard to the design of web sites but no recognised standards have yet been set with regard to the assessment of the quality of web pages [Nielsen and Molich, 1990], [Pfaffenberger, 1997], [Berk and Devlin, 1991].
Analysis If we strip out the terms which would not be familiar to the person in the
street, here is what we are left with:
The ﬁrst example contains absolutely no evidence that the writer has ever studied the area. Depending on the marker’s world view, this answer might get a mark ranging from ‘just fail’, on the grounds that it was true and relevant but devoid of advanced knowledge, to zero, on the grounds that it contains no evidence of having studied the topic and could have been written by a member of the public who reads the occasional newspaper.
The second example shows that the writer has done his or her homework thoroughly, to the extent of quoting eight different sources, including a classic text and some advanced journal articles, in the ﬁrst paragraph alone. This is the sort of thing which creates a good ﬁrst impression. You don’t need to write the whole piece at this level of density, but you do need to establish your credibility early on. If you put the heavy stuff on page 2, it’s usually too late; the reader will already have formed a low opinion of you, and you will have the doubly hard job of undoing that low opinion and then persuading them to change it to a high opinion. (They’ll probably think that if you’re silly enough to have a lightweight ﬁrst page, then there’s no point in changing their opinion of you.) Note that this example does not contain any WRITING STYLE 119 ‘buzzwords’ and that the English is pretty plain apart from the technical terms. It could only have been written by an expert, but can be read by a non-expert.
Spelling and punctuation Get your spelling and punctuation right. If necessary, buy a dictionary and/or go on a training course. They will probably be the best investments you will ever make. If you’re claiming to be a highly educated professional and you can’t spell or use punctuation correctly, then you’re off to a bad start.
Rhetoric and rigour Most students know that a PhD requires good theory, good documentary evidence, good science or good engineering. Many forget that it also requires good storytelling. Both rigour and rhetoric are essential ingredients of a successful dissertation. At most institutions, rhetoric is an explicit criterion for a PhD, expressed as ‘good presentation’ or ‘publishability’.
Books will tell you that the purpose of writing is to communicate. This is true. At the most crude and mercenary level, if you don’t communicate anything about your knowledge and ability, you won’t get any marks. (A blank sheet of paper, or a sheet of impenetrable wafﬂe, get zero for being content-free, not full marks for having no mistakes...) The dissertation
There are certain things that you’re demonstrating through your dissertation:
• mastery of your subject;
• research insight;
• respect for the discipline;