At first, you think you know a lot. As a postgraduate student, you are probably researching an area that you are familiar with. You probably did coursework related to your area of research when you did your Bachelors and/or Masters degree. So building up a bibliography should not be too painful an experience for you. You go to the library, go through the catalogue, select articles and books, retrieve them from the shelves, make photocopies, and start reading.
Yet doubts begin to creep in. Other people seem to have written everything that is interesting in your area of research already. You fear that your thesis or dissertation is going to be as original as a tribute band.
You need to have faith. Supervisors know what they are doing when they endorse PhD proposals. Your supervisor knows there is a little corner somewhere in your research area you can claim as your own. Your model or framework, which seems so basic and unoriginal at an early moment, will be modified and improved until the time your thesis/dissertation is submitted.
It may help to postpone reading specialist articles and books for a while. Start reading in books that give a general description of your research area. If there is a cross-disciplinary aspect to your research area, books about related areas can be useful. The day will come when different pieces fall together, and you will have a clearer picture of what your framework is, why the dependent and independent variables of your research have been so chosen, and other matters that are not clear early on. There will still be some missing pieces of the picture, but as your research progresses, they will come to your attention.
Don’t delay. Our suggestion is that you should start writing on your thesis or dissertation as early as possible. Note that we did not say “start writing your thesis or dissertation” but “start writing on it”. Begin by writing critical reports or analyses of the books and articles you have read. What you write at this stage may not be up to Masters or Doctoral standard but if you don’t start writing down what you have read and any insights arising from this reading, it is not likely that you will remember the detail of the articles nor your insights later, when you need them, that is, when you are trying to put together draft chapters. Writing can also help you think. It can keep your ideas flowing and evolving towards the ultimate goal of finishing your degree. It is generally easier to organise and assess your thoughts after they have been written down and you can study them. Writing is moreover a learning process in itself. Thus, the sooner you start, the longer you will have to learn what you need to know in order to be able to write a really good thesis or dissertation.
Get to know your own discipline. Each discipline has its own writing conventions. Consult your supervisor, and find out more from published work related to your area of research how you should write to fit those conventions.
Prepare for your own writing by reading other theses/dissertations. Read theses or dissertations close to your own area of research to get an idea of the acceptable range of structures and styles and to develop a schema for your own writing.
A dedicated computer. Ideally, you should have a computer dedicated to your thesis/dissertation writing. If you use it for net-surfing and e-mailing, chances are that your computer will be infected by a virus at some time during your (long) PhD study and destroy a file of your thesis/dissertation or even the file, if you have the whole thing in one very big file.
A non-dedicated computer. If you have to use the same computer for writing and net-surfing, you need to back up the files. Having backup files on a floppy disk is good but keeping a hardcopy as backup is better. Your backup floppy disk might be infected next because it is likely you insert it into the drive the moment you realise something is wrong with the files on the hard disk containing your thesis/dissertation draft or your data. A hardcopy can never be infected. When you have your computer fixed, you can use a scanner and transfer everything back onto the hard disk. Another way to be sure you do not lose too much if/when a file becomes infected, is to save individual chapters or parts of your work as separate files. Increasingly, computers come with writeable CD-ROMs, and these can be used for backup on a regular basis instead of disks. If you use the cheap “write once” ones, they cannot get infected when put in later. Of course, the most important thing is to have a very good virus-checker program installed on your computer and make sure it is updated on a regular basis, weekly or even daily if possible.
Inexpensive books. A way to look for inexpensive books is to go on the web. Another way is to browse second-hand bookshops. Both of these activities can be a relaxing change from writing. The books that you need may also be available in bargain (first-hand) bookshops which have a section dedicated to academic titles. If you live in a city, you should be able to find many of these (e.g. in London they are many along Charing Cross Road) and it would be worthwhile to go to them now and then. This PhD student once bought a book (with a slightly damaged cover) he needed in one of those bookshops for £0.90 that would have cost £20 in other bookshops.
Obscure articles. You may have to get hold of an article in a journal that your university library does not have a copy of. An Internet search reveals that it is not available in any other libraries in your country; or even if it is, you can’t wait for the article to be sent to you via inter-library loan. This should not be the end of your quest. Why not try a Google search to see if the author has a website? Or if you know his/her university affiliation, do a Google search to get to their website, and from there do a search for the author and get email address and/or website. In this way, you can contact the author directly and request (politely, with offer to pay for copying and postage) a copy of the article you want. There is also a possibility that the article you need has already been re-published as part of a collection of articles in book form. Go to the shelves containing the books of your area of research and those of related areas. Check the title pages of each of these books one by one. It is not as bad as it sounds. There are clues to help you in your search. An important clue is publishing date. It is unlikely that the article was published before the hypothetical book. Your knowledge of your area of research also helps. For example, if the article you need is about language and group identity, it is unlikely to be included in a collection of articles on formal linguistics. And even if you do not find the exact article you wanted, you may find other useful materials in the table of contents of recently published books in your area of research.
Over-recruit subjects. You may have to recruit participants for your research. Depending on the participant profiles and the data elicitation methods, the time it takes to recruit the right number of participants can be lengthy. It took this student two years to recruit enough informants that fitted the necessary profile of participants for research of a specific group in London. Contact with some of them was lost over time for various reasons. Some changed their mobile phone number without telling me. Some were not available for the data collection. Realise that there is going to be attrition, and recruit more participants than you need; then keep your fingers crossed.
Develop a system. Everyone is using a word processor these days, and we assume that you are no exception. As the number of files increases, so do the chances of forgetting what you have written in the files created a year ago. Some students do not find creating sub-directories on the hard disk helpful. In this student’s case, by Year Two of the thesis work, there were so many of them in his computer that he did not want to bother with them any more. He now thinks he should have kept a log, writing a brief summary (a good exercise for writing academic abstracts) for every file created. The log would have provided him an easier access to what he had written ages ago.
Spell-check in small doses. Spellcheckers can be a pain. This student does not like those wavy red lines underneath names and non-English words any more than the next person. Spellcheckers are not always reliable, and they can also pack up. This student’s did after his thesis draft had grown to more than 250 pages. A message came up and stated that the spellchecker no longer functioned for that specific file. Your thesis or dissertation is divided into chapters, right? As already suggested, store these in separate files and only put them together into one big file at the last stage before you submit your final draft.
Use the spellchecker to check references. You might want to consider not adding names of authors to the electronic dictionary of your spellchecker. You can learn to live with the red lines. They come in handy when you are doing the final check of what you have put in your bibliography. Is that name with the wavy red line underneath it properly referenced in the body of the thesis/dissertation itself and in the bibliography? And are all those author names in the text matched by the same ones, and also spelled the same, in the bibliography, and vice versa?
Use hardcopies. First and foremost, check for errors with a hardcopy. Research has shown that it is more likely for errors to be missed if you are reading from the screen. It has to do with efficiency of scanning and glare, among other things. Most of all, how much does a screen show? Probably half a page. You have to scroll up and down to make sure the points you have made two pages apart do not contradict each other, for example. It is tiring and frustrating.
Avoid some typical problems. There are language features that are considered to be errors in all kinds of academic writing. Here are some of them.
· Vagueness: Vagueness is defined as information or ideas not made explicit in a piece of writing. They are often not made explicit because the reader mistakenly believes they are in there, i.e. that they can be derived from the words that have been written down already. You need to make all the ideas explicit even if you think the result will read as boring, style-less, or repetitive. Ellipsis is one of your worst enemies. Everyone can read between the lines, but they will all read something different into the text. If you are not precise, you are taking a chance on a misreading and a misunderstanding.
· Punctuation errors: Punctuation is hard, particularly comma, dash, colon, and semi-colon. Don’t be discouraged. Find a good grammar book that has detailed sections on the use of punctuation marks, based on authentic data from written works that have been computer-analysed and not just on the intuition of the writers. Compared with some books on writing style, these grammar books are more up-to-date and make easier reading.
· Complete sentences: Writing in complete sentences is not mandatory for every genre. Fiction, for example, is one of the genres that allow incomplete sentences. However, it is unlikely that it is acceptable in academic writing, whatever the discipline. A typical error is to start a new thought and with a conjunction (e.g., And, But, Because), as a new sentence, instead of putting the conjunction between two linked thoughts as one larger sentence. For example:
After the reorganisation, the department no longer followed these guidelines. Because they wanted to develop their own new structure and guidelines.
This could simply be changed to:
After the reorganisation, the department no longer followed these guidelines because they wanted to develop their own new structure and guidelines.
· Run-on sentences. A common problem is that sentences run on and on, that is, they just keep on going without completion for several lines of text. A common problem in creating run-on sentences is the linking of ideas by comma (,) or and, or by a combination of these. Here is an example from a Computing thesis:
The network node frequency list algorithm accepts as input a web page, in this specific example I use the root of http://www.xyz.net as input and the output is a frequency list of all the words encountered in the web page apart from the words defined by the user in the stop word list.
A run-on sentence can usually be spotted simply by counting the number of words between full-stops, or periods (.). As a general rule of thumb, check for a run-on if the length of a sentence is over 25 words or more than two lines of text. The example above is 53 words long and over three lines of text. A simple sentence is usually under 15 words long, one line of text or less. Sentences of 25-30 words, approximately two lines of text, are not uncommon in academic writing. In a thesis or dissertation, even three-line sentences of 40-45 words occur on occasion. Sentences any longer than this are less common and may be quite taxing on the reader, though there are exceptions (see first sentence of the next section, for example). If written academic English is not your strong point, it is probably a good idea to aim for sentences that will not run over three lines of text, to keep yourself out of grammatical trouble.
· Number: There are many nouns in English that have two different senses, as either countable or non-countable (e.g. I heard a sound; the Concorde travels at the speed of sound), and this causes trouble in deciding whether it should be singular or plural and whether it should be preceded or not by the (definite article) or a (indefinite article). This area of grammar is especially problematic for those for whom English is not the mother tongue, as the system of countable and non-countable nouns and the rules for use of articles are complex. The following example shows four different possibilities, each with a different meaning:
The study of language is fascinating. (in general)
a language (any one)
the language (a specific one)
languages (any of them)
Again, there are good dictionaries on the market based on computer-analysed authentic data. There is no shortcut. You have to check every suspicious case to make sure that you have the correct specific meaning.
Revise with readers in mind. Remember that you are not writing for yourself but for an audience. Also keep in mind that this audience is not only your supervisor, who is likely to be more familiar with your topic and more sympathetic to you than other readers of your work. Always keep other readers in mind while you are writing, and try to imagine problems readers might have understanding your ideas. Aim for maximal clarity and transparency.
Be ready to write several drafts of your work. Do not delude yourself that you can write down your ideas once in a perfect form or that you can finish your postgraduate study in record time by writing only one draft of your thesis/dissertation. Experienced writers often write dozens of drafts of their works. While even one dozen drafts may be more than you can bear to think about, there is no doubt that you will need to write more than one draft of your thesis, and probably more than two. This is because redrafting has a very important purpose: to make sure what you have written will be understood by the reader exactly.
Revise as a step-by-step process. Be aware of the types of errors that you might make and revise for each of these in a series of steps. Make your own revision protocol or list of common errors and ways to correct these. Here is an extended example of the decision process you might undergo in stages of revising the Computing student’s run-on sentence above.
· Correct any cases of comma splice. A “comma splice” links two sentences merely by comma punctuation. Thus, the comma following the phrase a web page in the Computing student’s text above must be corrected. The easiest thing to correct this is to finish the first sentence a web page by placing a full stop or period there, thus creating two sentences.
The network node frequency list algorithm accepts as input a web page. In this specific example I use the root of http://www.xyz.net as input and the output is a frequency list of all the words encountered in the web page apart from the words defined by the user in the stop word list.
· Make decisions aimed at improving readability. After the comma splice has been corrected in the text above, the two parts of the second sentence joined by and are still quite long. Punctuation can help to signal to the reader the parts of this long sentence. When the two parts of a sentence joined by and are more than a few words long, a comma is normally inserted before the conjunction to indicate the separation of the two main thoughts. It is also possible to add additional punctuation to make the sentence easier for the reader to parse, that is, to divide into component parts. This could be done by putting a comma after the introductory phrase, In this example and another comma before the phrase beginning with apart. The text would then be:
The network node frequency list algorithm accepts as input a web page. In this specific example, I use the root of http://www.xyz.net as input, and the output is a frequency list of all the words encountered in the web page, apart from the words defined by the user in the stop word list.
· Show the main divisions of ideas clearly. It is the job of the writer to make the text clear and understandable for a reader. In the above example, there are three different parts of the long second sentence, each one separated by comma. In the first use, the comma separates an introductory phrase from the rest of the sentence. In the second use, the comma separates the two main ideas. In the third use, the comma separates an added piece of information at the end of the sentence. The fact that the main division of ideas is between (i) the part of the sentence before and and (ii) the part of the sentence after and is obscured by using the same punctuation mark (,) for three different purposes.
· Break up complex ideas into simple sentences. The writer could therefore decide to make the text a bit easier to understand by removing the and to make three sentences instead of two.
The network node frequency list algorithm accepts as input a web page. In this specific example, I use the root of http://www.xyz.net as input. The output is a frequency list of all the words encountered in the web page, apart from the words defined by the user in the stop word list.
This three-sentence version is simpler and so should be easier for a reader to comprehend.
· It all depends on what you want to say. The final decision about how to divide up and punctuate ideas lies with the writer, who must decide on matters of style and the grouping of ideas. In the example above, the three-sentence version would be preferable if the writer considers that the main division of ideas is between the topic of the input, as developed in the first two sentences, and that of the output, as developed in the third sentence. The two-sentence version with and might be preferred if the writer wants to combine the two pieces of information about this specific example.
We would be very interested in hearing about your own experience or recommendations, and others may also want to know about this too. Please email us with your story or comments, and let us know if we can use your remarks on the website. If so, please give your name, country of origin, the University where you are studying, and your field of study.