Tag Archives: tips

The abstract


The curious thing about an abstract is that although, after the title, it is the first text to be read, it is usually the last thing to be written in the dissertation. The reason is quite simple. Writing an abstract is a highly developed skill. On one page, or less, the author needs to summarise the entire body of the research work, describing the research question(s), the methods used to gather new evidence, how this evidence was analysed, list the key findings, and say why these are important. This is a tall task, demanding a number of difficult decisions about what to include and what to leave out of the text. The added pressure is that this might be the one and only part of your research that a browsing researcher of the future will read, so you need to captivate their interest in half a page or so. On websites such as https://ethos.bl.uk/ which is the British Library catalogue the entire output of completed UK PhDs, are the abstracts that researchers consult to decide whether to read the whole PhD dissertation, or not. This is a good site to consult to gain an idea of what is needed, but creating your own takes practice.

For this reason, a good supervisor will encourage the research student to finesse their skill at abstract writing by trying several versions before the culminating attempt. It is sometimes said that to ask a research student what their PhD is about at the beginning of their studies is to get a verbal paragraph in response, but to ask the same question at the end gets a succinct response of 5 or 6 words. This is because over the intervening period, the researcher has honed their analytical skills and (hopefully) their ability to separate what is really important, from that which is interesting but incidental to the main research question. The abstract is about what the reader needs to know, rather than the wider perspective on what might be nice to know.

Writing a concise abstract is a skill that will also serve an author well if/when they progress to submitting a paper to an academic journal. Again, the objective is to capture the essence of the article and grab the attention of the prospective reader. In a society awash with information, it is the ability of information to attract our attention that will distinguish it from the things that do not get noticed, and do not get passed on. In ‘the attention economy’ getting noticed is perhaps even more important that the information itself. If no-one ever reads your brilliant idea, it slowly moves to the graveyard of good ideas. There is a careful balance to be achieved between sensationalist headlines and dry-as-dust reporting, and though the title needs to reflect this, the real meat of what the text is about is contained in a cleverly worded abstract. Ask yourself, what does this abstract actually tell us? For this reason, it is almost never too early for a research student to begin studying the structure of a useful abstract. According to Polonius (in Hamlet) ‘brevity is the soul of wit’ and it is also a very powerful academic skill.


In the hot seat – defending the thesis

hot seat

One of the unusual aspects in studying for a PhD is that the final examination of competence (and quality) is based not simply on the written dissertation but, more importantly, by giving a verbal defence of the work under external scrutiny. Normally this takes the form of a final extended question-and-answer discussion over a couple of hours with an External Examiner from another university and an internal (to the host university) examiner. The student is tested to ensure their authorship of the dissertation and to justify the methods of data-collection, analysis and the formulation of conclusions. What exactly is the new contribution made by this piece of knowledge to the discipline as a whole? Is it really new primary research? Across the university network, the regulations might be applied slightly differently, ranging from a quiet discussion with just the examiners present (the supervisors are not admitted) to a full public audience (as in Scandinavian universities) with almost anyone who has an interest in the subject being able to spectate.

As students will not have any previous experience of the viva voce – the oral defence – of their work, it goes without saying that the supervisory team have an obligation to prepare the student about what to expect. This can be done either as a series of conversations, or as a full “mock viva” in which academic colleagues of the supervisor will role-play and raise the sorts of questions or problems that the student might expect to encounter during the real viva. Student responses can be explored and rehearsed.  Normally the viva is not a confrontational event, but it can certainly be ‘robust’ and very demanding for the student. Almost any aspect of the research can be explored, and the student needs to be able to explain and justify what they did (and did not do) to reach the conclusions of their thesis. Common questions ask the student to summarise the research, to indicate their unique contribution made to the subject, to interrogate the quality of the results, and to explain in detail how those results have been achieved. The selection of External Examiner is usually as a result of a nomination to the university by the research supervisors of a shortlist of potential academics that have an expertise in the subject area. The student has a right to expect that the examiners will be objective and fair, but almost nothing is off limits for commentary, from simple errors in spelling or grammar, through gaps in the literature review, to the logic of data-collection and the presentation of the results.

In some cases, the examiners might challenge the student about what they have written, while at the same time being in broad agreement with the student – but they want to gauge the student response. The examiners want to be confident that the PhD student really does have an intimate understanding of both the subject matter and the processes of advanced research. The viva report that is fed back to the university will not only make a recommendation of a pass, or ‘pass with amendments’ (it is possible, but rare, to have absolutely no corrections) there will also be recommendations that need to be met before the award is given. These recommendations might simply be spell-checking or entering a missing reference or two, but they might also be a requirement to re-write, extend or remove of some aspect of the dissertation – the addition of more up-to-date references, a clarification of technique, or a re-working of the conclusions. Whatever the recommendations might be, the student now has an unambiguous written list of things that they need to address in order to gain the PhD and a time deadline for these changes to be made. It is perhaps the clearest guidelines that they will have had during the entire PhD study, and a small price to pay for the award of the highest academic degree.

Reviewing and revising


One of the strange but common occurrences in producing large pieces of writing is that the writer frequently becomes so close to the text that small (and even some large) errors get completely unnoticed. Supervisors have different ways of dealing with this. Normally I give a detailed commentary chapter by chapter, and then quickly read a revised version, but do not revisit that unless a later chapter forces some sort of re-think. It is usually emphasised from the very start of a PhD that the research project should belong to the student, not to their supervisors, and as the final draft of the dissertation approaches completion, this is a crucial time for the student to assert their ownership. In defending the thesis at a viva, it is the student who will be held responsible for any errors and misspellings, but the supervisors can effectively support this process by timely guidance.

Firstly, in addition to re-reading every chapter as it is drafted, I encourage students to review and revise the entire dissertation just before they start to write the final chapter that brings everything together. In this way writers can check for any small typos and at the same time refresh their memory about what they have written earlier. (It can be a relatively long time between the start and the end of the writing process, and memory can play tricks!) Next, it is usually a good idea to get an extra person (apart from the writer and the supervisor) to read through a document (in stages) to give some feedback. Although it helps to have someone who is knowledgeable about the subject material, the main thing is to have someone that can be trusted to tell you the hard truth. A friend or partner can be a great source of guidance to clarify writing (“what exactly do you mean by this sentence?”). Thirdly, it is a good idea to re-read the dissertation (again!) after you think that it is finished – perhaps not every single page, but certainly to dip into sections at random and check that the detail still makes sense. Do not skim over the small things such as tables or the caption of diagrams, these are just as likely to contain errors as any other paragraph.

It seems superfluous to say, but as each section or chapter is backed up for security, it is important that each saved copy has a date and/or version control number on every page. With multiple back-ups and multiple versions of revised copies, it can be very easy to create confusion. Ultimately, however, there comes a time to stop tinkering or tweaking the text and let it stand on its own merit. In some universities, the submission of the dissertation requires to be countersigned by the supervisor to agree that it is now in a fit state to be sent to an External Examiner for evaluation, but in other institutions the supervisors are simply informed. Either way, the student is responsible for the final contents and its appearance, and the supervisor is responsible for helping the student to produce the best submission under the prevailing circumstances.

Appendices and archives


As with every piece of substantial research, it can be a problem to decide what needs to stay in the main text and what can be left out without substantially impacting upon the ability to understand the narrative. This is where appendices can be useful, and an important role of the supervisor is to give gentle guidance on what needs to go into an appendix and what simply best kept in an archive. The temptation of the early career researcher is to believe that everything is necessary, and in the classic “can’t see the wood for the trees” mentality, to cram loads and loads of supplementary data into appendices that are rarely (if ever) read subsequently. The golden rule of an appendix is that it should contain information that is not so important that it needs to be in the main text, but that it can still substantially contribute to understanding the background detail of that main text by providing supplementary evidence. A good example of this would be a large table of numerical results (in a quantitative study) or a key interview transcript (in a qualitative study). Both of these types of appendix can furnish crucial raw data that can enable an experienced reader to ‘get behind’ that research results and help them to make their own interpretations of the results. (Or understand the decisions made by the research student).

An appendix is not an excuse to dump all the information that has been collected in the research but has not been able to find a place in the main text. Crucially, the appendices (and footnotes/endnotes) are included in the word count for a dissertation submission, so weighty appendices risk robbing space for the more substantial (and important) presentation of the main text arguments. If a point is critical to the development of the research conclusions or interpretations, then it should probably be in the main text; if it is important but not crucial to see in detail and can be summarised in the main text, then perhaps a fuller account can be included in an appendix. There will also be some information that has a background relevance but should neither be included in the main text nor the appendices, but this does not necessarily mean that it can be thrown away. There will be information such as lists of consultees, or anonymised participants codings, or transcripts of (most) interviews that might be needed in the months following completion of the research but reading these are not germane to understanding the narrative of the main text.

In some cases, for example if the researcher does not intend to continue with the research topic, some of this background research might be archived with the university library, or with the research supervisor. Increasingly, it is common for research that has received public funding to require that the raw data is made publicly available, and this creates new opportunities and new difficulties. If the data is required to be publicly accessible for ten years (or ten years from the last time the data was accessed) then it is conceivable that the raw data will be openly available for far longer than any individual research project, and possibly even longer than the lives of individual researchers. This places an important new responsibility on the researcher to be very organised and very transparent in their collection and use of data. It also requires an accentuated awareness by the supervisor (and then the student) about the inclusion of what information is relevant for the successful completion of the dissertation, what can go into an appendix, and what should be kept in an archive.

Checking references and citations


It is amazing how many students get into a muddle over the simple process of ensuring accurate links to the supporting evidence for their claims. Let us get this right; it is not the supervisors job to check that citations and references are correct, but the External Examiner of the degree certainly will check this. For that reason, it is the supervisor’s job to make sure that the research student gets it right. It is not a difficult task, but it can be time-consuming, so the task needs meticulous care.

To lay down some ground rules, when researchers make a claim or a statement of ‘fact’ in their writing, we need to establish the source of that claim. There are two ways to do this; either the information is new, i.e. as a result of the new research, or it is derived from previous research. When it is the latter, the normal way to credit the source of the evidence is to include a citation in the statement, (such as, “There are …. (Rennie and Smyth, 2019)” or “Rennie and Smyth (2019) claimed….” This then flags the full reference, listed in alphabetical order at the end of the document, which in this case is, “Rennie, F. and Smyth, K. (2019) Digital Learning: The Key Concepts. London: Routledge”. (I also usually include the ISBN – International Standard Book Number – for books, and the DOI – Digital Object Identifier – for journal articles, but that is just me being extra-sure that subsequent readers can locate the item.) Crucially, if a claim or a ‘fact’ is given without a citation to the supporting evidence, then it is assumed that the information source is the writer. If it is not the writer, then the missing citation is regarded either as shoddy workmanship at best, or plagiarism at worst. Plagiarism – knowingly misrepresenting some other persons direct words/ideas as your own – is regarded a major misdeed in academia, so an important role of the supervisor is to ensure from an early stage that the student treats accurate referencing very seriously.

I encourage my students to adopt the same rigor that I do, which is, firstly, start to compile the list of references right from the very start of the research project. Always write out the references in full, and do not leave any information out – the chances are that you will forget to go back to correct it. When I find new relevant articles, books, or other resources that I know I want to include in my writing, I add the references to the master list as I read them. Secondly, when I have completed my final draft text (and usually a couple of times before then) I sit down with a printout of the main text on one side, and the list of references on the other. I go through the main text, marking with a highlighting pen every citation that I come across. I then turn to the list of references and highlight it there too. By the time that I have read the whole of the text, every citation should be highlighted, and every reference should also be highlighted. If I have missed any references, or included references that I have not actually mentioned in my text, then this is the opportunity to update the reference list by either adding or removing the relevant items. It is a laborious process, but it is fool proof.

The danger of not doing this is when a particular citation catches the eye of the External Examiner, and they turn to the reference list for the full details. If the reference is missing, they do not know whether the writer has made a one-off mistake, or if there are many more missing references. The result is an almost mandatory viva condition to “Check all references” before passing the dissertation, rather than getting a “no corrections required”.

Conclusions and recommendations

QuestionsFor a PhD, indeed for most academic research, the researcher is judged more on the quality and justification of the research method, rather than getting “the answer”, but the interpretation of the results, the conclusions, and any possible recommendations are a pretty important part of the outcome too. There are three common mistakes made by early career researchers for which supervisors must be on the lookout, but basically they all revolve around the one question – do the conclusions relate directly to the evidence produced by the research?

This might seem a rather obvious question, but it is important to address first. Frequently the writer of the conclusions will have very weak, generic conclusions that seem to fade out and say nothing. Sometimes it seems that an experienced researcher could write these sort of conclusions before even starting the research – they lack clarity and don’t really say very much. Secondly, is the opposite extreme, the temptation to read too much into the data and make conclusions or predictions for which there is no real evidence. This is almost worse than understating the results, for a critical reader would begin to wonder if the whole of the research project had been influenced by this optimistic speculation. It would certainly make me re-read the data analysis more carefully to see if the researcher displays any suggestion that they “knew the results of the research” in advance, and looked favourably on the data in order to “find” the answers that they wanted. In this phase of the research, the supervisor has a crucial role as a critical friend, to challenge the research student into justifying their conclusions, and relating this directly back to the evidence displayed by the analysis of the data.

The third common error in writing the research conclusions is that the writer describes very plausible conclusions, which actually have little or nothing to do with the research project that has been undertaken. The researcher has become so immersed in the research that s/he has lost a sense of the boundaries of the project. Everything seems to be connected to everything else, and there is a lack of focus on what is really relevant, or evidence-led.

In my role as a supervisor, I call this “the so-what?” moment. The student has designed a research project, identified a key question and related it to the current knowledge of the subject, then gathered a load of new primary data which has been analysed to reveal some “results”. So what? What does this actually mean? Sometimes the results show what can not be proved, and this can be almost as important as getting “an answer”. Knowing what the evidence does not show, or where there are blind-alleys in the data gathering, can be critical in the design of a new research project that advances our knowledge a stage further. What do the results really say? What claims can be based on the research and what does it tell us about the research question that is an original contribution to the subject?

In this respect, the advice from a supervisor needs to be offered carefully, in order not to discourage or demoralise the student, for this is a time for honest self-reflection. It is better to be slightly less ambitious in the research aims but be more robust in the collection and interpretation of data, rather than strive for an idealistic but very ambitious research aim that is undermined by careless data collection, too many assumptions instead of hard evidence, and joining the dots to make speculative predictions rather than making comments based on solid evidence that can be justified by the data. I always tell research students to take a pause after they have written the penultimate draft of their data interpretation chapter, then go back to the very start of the dissertation; read every section afresh, as if for the first time, and make the final tweaks to the narrative. Then, fresh with this knowledge of where they are and how they got there, try to write the conclusions by answering the “so what?” question.

Interpreting the data


Building on existing knowledge

A key role of any supervisor is helping the research student to bridge the gap between the fundamentals about what is currently known about the research topic, and the new results which have been generated through the research activities of the student. All research is built upon some level of pre-existing knowledge on the subject, even if existing knowledge is patchy or otherwise insubstantial. In the literature review chapter, the student will have built up the profile on what is already known about the research topic and how that information can be backed-up by evidence from the academic literature available. In the analysis chapter, the first task is to provide some interpretation for the new, primary research conducted by the student, but a significant secondary task is to relate this back to the previously discussed evidence and underpinning theories which were explained in the earlier chapter(s). This can be a tricky task because the new research results might either fully support earlier work (in which case, what’s new about the research?) or else directly contradict it (in which case how do you prove the superiority of the new results?).

It is a useful tip to bear in mind that hindsight is a wonderful perspective, so try to avoid feeling too smug about the wonderful flashes of insight produced by the new research. Always assume (unless proven otherwise beyond doubt) that the earlier researchers did the best job that they could with the information, equipment, and currency of information at the time they were doing their research. It is easy to look back in history and wonder why our predecessors could ever have believed some of the accepted wisdom and “common sense” of the time, but in fact we are no different: we simply have much more information in a greater level of detail, but it would be a fool or a knave who would claim to know every last thing about the chosen subject. In most circumstances the research will both tweak prior definitions, and then throw clearer light on an existing area or understanding, or it will provide new data to enable the researcher to propose a new way of thinking about the existing data and justifying that new approach by producing new evidence (or a new way of interpreting the existing evidence).

Either way, the first stage of research analysis is to compare the new information with what has already been understood, and then go beyond this to open up a new area that is worthy of further research (and/or a proposing a different way of understanding the subject). Two common failings at this stage of the research process are either to appear to present the conclusions as if nothing important had ever preceded the current research (thereby inventing a whole new branch of epistemology) or else failing to restrict the conclusions to the actual results of the current research and instead attempting to make grand conclusions for the whole of the discipline (rather than just for the current research project). Either way, getting to grips to understand the importance of the new research, and using it to build upon earlier research results to improve our knowledge of the subject, is a fundamental step in the dissertation.

Establishing the storyline


One of the first things that both the supervisors and the research students need to remember is that although the dissertation is the justification of an academic thesis, it also needs to tell a good and convincing story. There is little point in making a wonderful discovery if you cannot properly communicate with other people to tell them about it. Research is about discovering something unknown, and like any good mystery story, there needs to be an introduction to set the scene for your readers, (the literature review) there needs to be a storyline to develop the research agenda (starting with the methodology) there need to be clues discovered as the story develops (the results chapter(s)) and there needs to be a moment of final revelation of the object of your search (the analysis and conclusions).

The delivery of this story requires a certain writing style – it needs to follow the academic conventions of the subject discipline, it is not a novel – but that should not mean that the ‘story’ that the researcher wants to tell should not be easy to read. There are some simple tips, such as to first lay-out using numbered sub-headings, the main headlines in the ‘story’. The separate sections can then be written under these headings and linked together to form the chapters. A good idea is to begin by drafting the contents page of the dissertation, listing the chapter headings (1. Literature Review, 2. Methodology… etc.) and then entering the various headings of likely sub-sections. In addition to helping to establish a coherent storyline (which can be amended as the writing progresses) this enables the dissertation to be written in a manner which is not necessarily linear (sub-sections can be skipped and returned to a later date) and built up piece-by-piece while still keeping within the framework of the story. It is also a good tool for discussions between the research student and the supervisor about stage-by-stage progress (I have used this with red (not-done) green (completed) and amber (working on it) highlights, to help students prioritise what bits of writing need to be tackled next).

The bottom line is that the research student needs to craft a good story to introduce, explain, and discuss their research project, and if this is easy to read, then it will be easier for readers to follow and perhaps build-upon in subsequent activities. This includes correct spelling, good grammar, and simple tactics such as to avoid l-o-n-g and cumbersome sentences (I had once student who wrote a sentence containing the word ‘and’ seven times! This really was three separate sentences and would have been far easier to understand if it had been written in a simpler style.) Another avoidable error is to include sentences which give ambiguous comments. If there is a way in which your comments can be misinterpreted, it is human nature that someone will take the wrong meaning, and this can be easily avoided by actually saying what you really mean, and keeping this in simple language that cannot be misunderstood. Using hierarchical numbering for the chapters, sections, and sub-sections not only helps to create a clear storyline, it also helps to allow cross-reference to earlier (or future) comments in the dissertation.

It is the role of the research supervisor to read and give comments to help improve the direction of the writing process. The student does not need to like these comments (and indeed, at their own risk, may chose to ignore them) but they should heed them because it is the duty of the supervisor to direct the work of the student to ensure that they give the very best presentation possible of their work for examination and further scrutiny.

How to Illustrate your results


When I was writing up my own PhD (in the antediluvian days before personal computers, desktop publishing software, or graphics packages!) I was given a very useful lesson by the Prof. who was my supervisor. I was agonising about how good my hand-drawn graphs and maps needed to be, how precise the individual, hand-printed, iron-on lettering needed to look. He informed me, rather drily, for that was his preferred style, that I was “… training to be a geologist, not a draughtsman!

From that response I understood, correctly, that if my diagrams are clear and accurate enough to convey my key point(s) then a point of diminishing returns is quickly reached on the time spent labouring over them. There is no need to produce a “work of art” – it is about “communication”. The situation is slightly different now, for there are lots of clever software packages, in Excel and elsewhere, which can quickly produce lots of impressive diagrams which can be “cut-and-pasted” into the text with minimal effort – but the two fundamental points remain the same. Firstly, if the initial data is weak and/or disorganised, then any resulting illustration is hardly worth the effort of trying to interpret with any degree of real meaning. As computer programmers are taught early – GIGO – (Garbage in, garbage out)! Secondly, a diagram (or a map, or a graph) needs to convey something meaningful. It is a visual expression of something that the author is trying to convey to the reader, so if this can be communicated clearly and simply, that is sufficient. There are far too many elaborate diagrams that are over-designed, and the result can appear so complicated that it is the diagram, rather than the results, that needs to be explained to the reader.

In some subjects, there are more-or-less standard conventions for diagrammatic representations, such as histograms, bar-charts, tolerance diagrams, or pie-charts. It usually makes sense to abide by these conventions because it can help comparison with similar studies elsewhere. Usually, simple is best. Let the eloquence of the diagram communicate the data for you. Sometimes, particularly due to the speed and ease with which computer-generated diagrams can be generated, there can be a tendency to “graph every variable against every other variable” in the hope that a stunning correlation is unexpected revealed. While this can happen, it is more likely that simply a blinding flash of the obvious is revealed, without contributing anything more than confusion to the current understanding of the topic. As with the use of statistics, it is always better if the author actually understands what they are trying to do before attempting the activity. It is too easy to drop into the text a “pretty photograph” or a diagram of a rather obvious feature without actually conveying much real information (e.g. a pie-chart of the male/female split of respondents; it is probably better just to give the percentage figures).

In a few cases, the use of a few clever diagrams, such as fishnet images of topography, or bar-chart information superimposed on a map to show geographical abundance, can produce a stunningly visual interpretation, but these should be used sparingly. While it is true that a (good) picture can say a thousand words, the tokenistic use of photographs, diagrams, or graphs can simply clutter up the main text, and require further text to explain the image to the reader. A good illustration actually says something clearly and makes a positive contribution to help the reader understand the accompanying text and data.

Using statistics in research


For me, the golden, unbreakable rule is, never use any statistics if you do not really understand what they mean! This might seem obvious, but it is surprising how frequently statistics get misused or misinterpreted to support an argument that actually has no real basis in fact. The famous saying that there are three types of lies – “lies, damned lies, and statistics” is attributed to Prime Minister Disraeli, but when properly used, statistics can be clear, unequivocal, and very supportive in communicating complex data simply. The main problem(s) are often started by people deliberately selecting the information that they want to hear and then seeking for statistical back-up which looks impressive and difficult to challenge. Secondary problems occur when people either do not properly understand what the real statistics are saying, or when people choose to deliberately select or twist the statistical information which seems to support their preferred point of view. Ultimately, all three problem areas both devalue the use of statistics and also take us even further away from a clearer understanding of the situation that we are trying to interpret and communicate.

A good “rule of thumb” is to stick to the simplest means of statistically expressing the results that you want to communicate. Percentages, pie-charts, and histograms might look fairly unadventurous if you are trying to impress an examiner, but they have the advantage that they are quick to produce, clear to interpret, and easy to understand. Fancy calculations may look more impressive, but they are frequently harder to produce, more difficult to fully understand, and have a greater chance of either the creator or the reader making errors of interpretation. A common error is to quote percentages rather than give simple numerical values for small population samples. If 7 out of 12 of your interviewees agree, say “7 out of 12 agree that…” rather than “58.33%…” A difference of 1 person immediately gives an 8% error and is clumsier. Keeping it simple gives both a truer impression of the data and an easier comparison with other results. Similarly, I have seen some very impressive and complex diagrams, complete with 3-D shading and vector trends, which actually do not tell us very much at all because the detail is lost in the artistic flamboyance. They look fancy but add nothing to the discussion.

Quite often, certain disciplinary areas will have their own conventions as to which statistical procedures are common, or preferred, and how they are presented. The supervisors should be able to advise on these common standards, and the benefit is usually that the new research data can both make use of earlier research results, and can be easily contrasted and compared with already published data in the discipline. Some statistical procedures can look complicated to calculate, but are actually quite straightforward to use. All universities will have opportunities for research students to attend courses on statistical methods which are appropriate for different subject areas of research, so students should get training early in the research to avoid any subsequent false starts. There are lots of self-guide short courses on the web, and of course in text-books, but self-tuition is also open to misunderstanding, so it is always best, in the first exploration at least, to work through the procedure(s) alongside someone who is already very familiar with the statistical technique(s). Bear in mind, contrary to some current political rhetoric, there are no “alternative facts” simply facts that you acknowledge and facts that you might prefer to ignore. Research is about improving knowledge, not picking just the bits that you like.