The Structure of the PhD

views updated

The Structure of the PhD

What does a PhD Thesis Actually Look Like?
Chapter 1: Introduction
Chapter 2: Literature Review
Chapter 3: Research Methodology
Chapter 4: Results
Chapter 5: Conclusion
Other Sections

What does a PhD Thesis Actually Look Like?

The quickest way to find out what a PhD thesis actually looks like is to go to a university library and take a look at a few PhD theses. This will give you an idea of the content and standard required. However, you may be puzzled at the differences in the format and layout. The best thing would be to look at PhD theses in the subject area that you propose to do your research. Look at the more recent ones and try not to go back more than ten years. You may find that the layout for a PhD in management may differ in detail, but generally contain the following chapters:

  • Abstract
  • Chapter 1: Introduction
  • Chapter 2: Literature Review
  • Chapter 3: Research Methodology
  • Chapter 4: Results
  • Chapter 5: Conclusion
  • Bibliography or References

The above represents the minimum number of chapters and their logical sequence. However, you may come across theses with more than five chapters. For instance, my thesis (Ahmad, 2001) is divided into eight chapters as follows:

  • Chapter 1: Introduction
  • Chapter 2: Literature Review
  • Chapter 3: Research Methodology
  • Chapter 4: Selection of Companies and Data Collection
  • Chapter 5: Results (1)—Descriptive Statistics and Correlation
  • Chapter 6: Results (2)—Complementary Fit: The use of hierarchical logistic and multiple regression in the analysis of the interaction effect between person and environment
  • Chapter 7: Results (3)—Supplementary Fit: Co-worker congruence and superior—subordinate congruence
  • Chapter 8: Discussion and Conclusion
  • Bibliography

Notice that I have three results chapters and a separate chapter on the selection of companies and data collection. I conducted many statistical analyses and they were far too lengthy to be contained in a single chapter. So I divided them into three separate chapters according to themes. ‘Complementary fit’ and ‘supplementary fit’ are technical terms that I do not wish to elaborate in this book. Suffice to say that these are related but separate issues, and it is therefore logical to separate them into different chapters. Selection of companies and data collection are usually described in the research methodology chapter. However, I wrote about them in a separate chapter immediately after research methodology. It was more logical to do so as there were specific criteria for the selection of companies, due to the nature of my research. These criteria had to be described in some detail. Combining the chapters would make the resultant chapter too lengthy.

Thus, the issue of whether one should have more than five chapters, and what the additional chapters should be, would vary from thesis to thesis. Essentially, the main criteria would be one of length. If you find that your results chapter is too long, break it down into different chapters according to themes. If you feel that you should describe something in a separate chapter in order to draw special attention to it, by all means do so. This is of course provided that the new chapter is not too short. Generally there should not be chapters that are too long or too short. However, it is generally accepted for the introduction and conclusion chapters to be shorter compared to the other chapters in the thesis. Literature reviews and research methodology chapters tend to be long. It is always useful to look at other theses and compare them with your own. However, do not abandon your judgement and common sense. If you are completely lost, then ask your supervisor for his opinion!

On a general note, all the chapters should have an introduction and a conclusion as subtopics. For instance, the literature review should have an introduction as the first heading and a conclusion as the final heading. The same applies to the research methodology and other chapters. This enables the reader to grasp the essence of the chapter, by merely reading the introduction and the conclusion, without having to read the entire chapter.

I will now explain in detail what the chapters are that should be written and what each should contain.

Chapter 1: Introduction

The introduction sets the scene and generally hints at why the current research is necessary. I can safely say that I have yet to read a thesis that does not have an introduction. So, what should an introduction contain? My thesis (Ahmad, 2001) had the following sub-headings in the introduction chapter:

1.1 Historical development and current importance

1.2 Main purpose of this research

Another question that might be asked is, ‘What is the purpose of an introduction?’

The purpose of the introduction is to set the scene. It first gives a brief overview of the topics involved and more or less tries to justify why the current research is important. It gives an overview of the subject area of the research, the constructs involved, the problems with past research and to what extent the current research aims to address these problems. Now this might sound like a lot. Perhaps the reader can get a better idea of this by simply reading several theses to see exactly how other students have written their introductions.

A valuable tip that I can give you is the introduction should never be the first chapter that you write . In fact, a colleague of mine recommended to me that it should be written last. I feel that it can be written simultaneously with the conclusion after most of the other chapters have been written. The rationale for doing so is that the PhD process is rarely sequential. It is only after experimentation that you finally decide on what should be reported in your thesis. Of course, before embarking on a research you must be clear on what is it that you are trying to do: what constructs will be involved in your research, what research methodology you will be using and what statistical analyses you will be carrying out. However, the PhD research is very often fine-tuned as it progresses. What you have proposed earlier on may not be feasible now. You may have found a better method along the way. As I mentioned earlier, the PhD process is a process of discovery. You discover and learn things progressively up to the end, rather than learn it all from the beginning. You are constantly improving yourself every step of the way. Consequently, it should come as no surprise that things that you proposed in the early stages can be substituted by something better. In fact, this was what happened to me. During my research proposal, I merely planned to look at three-dimensional graphs to investigate the interacting effects of group work environment on the relationship between personality and satisfaction. Subsequently, I found a better method, i.e. using a statistical test of hierarchical regression to investigate the moderating effect of the work environment on the personality-satisfaction relationship.

Another crucial point which I found recently is that you should address issues relevant to the study at hand and not describe at length, factors or variables which are not part of the study. If you do so, you are actually inviting the examiners to question you about their relevance, or possibly worse, ask you to justify why you have not included these variables in your study.

Chapter 2: Literature Review

This is probably one of the first chapters that you will write. However, you probably have to keep amending the same throughout your entire PhD period. The literature review should discuss the constructs (e.g. job satisfaction, personality) which are used in the current research. A literature review is a way of demonstrating that you have a full professional grasp of the background theory relating to your subject and that you are aware of the controversies, developments and breakthroughs that are currently exciting or engaging leading practitioners in the subject. Thus the literature review should not be a mere recitation of what other researchers have done in the area of the proposed research; instead, it should be a critical evaluation of what other researchers have done. It should identify a gap in the literature (i.e. an area or issue that no previous reported research has addressed) that the proposed research intends to fill. The literature review should identify the strengths and weaknesses of previous research. The articles should be arranged in such a way that the writer can build up his case from one article to the next. The literature review should also end by showing how these different variables are related. In addition, it should end with a summary of all the weaknesses and omissions of previous research—weaknesses and omissions that the current research will address. In other words, you are not only criticising previous research but you are able to devise a research design that will overcome the criticism that you are making. So, the literature review chapter may end with the proposal of a new model that will capitalise on the strengths of previous research while avoiding their weaknesses.

Of all the chapters in the thesis, the literature review probably contains the most cited articles/books—generally around 200 in total.

Chapter 3: Research Methodology

The research methodology chapter should contain the following:

  • The research design.
  • An enumeration of the predictor variables.
  • An enumeration of the dependent variables.
  • Demographic data, e.g. age, sex, etc.
  • Rationale for the use of questionnaires (or whatever method used).
  • Detailed description of the questionnaires used. A brief description of alternatives available and why they were not chosen.
  • Possible sources of error.
  • Development of theory and hypotheses.

The research methodology is similar to the literature review because articles of previous studies are also cited here. However, whereas the literature review focuses on the constructs, the research methodology focuses on the instruments that measure these constructs. The research methodology should describe in detail the instruments/questionnaires that were actually used in the study. It should also include a brief description of alternative instruments or methods available and why they were not chosen. Students should resist the temptation to enumerate in their theses all the different methods of data collection that are available, as well as detailed descriptions of the pros and cons of each method. In short, the student should not reproduce the textbook. I have seen instances where students were not awarded PhDs due to this reason. This is the difference between PhDs and MBAs. PhD students are expected to be already well acquainted with the existence and the pros and cons of the different methods of research. What is expected in a thesis is evidence that a rational decision has been made. Thus a thesis should contain a detailed description of the method chosen, and only a brief statement that the student is aware that there are other alternatives available, but that they were not chosen for the reasons given. If the reasons for the choice given were illogical, rest assured this issue will be raised in the viva.

I will now explain briefly the different research methods available. There are basically two main types of research—experimental and ex post facto (or after-the-fact research). Experimental research is one where the conditions are controlled so that one variable can be manipulated at a time in order to test a hypothesis. Put in another way, only one variable at a time (the independent variable) is manipulated and its effect on another variable (the dependent variable) is measured while all other variables that may confound the relationship are held constant or controlled. Only experimental research can allow the evaluation of causal relationships among variables. It usually involves two groups of subjects—the test group and the control group. Measurements of both groups are taken immediately prior to the test. Thereafter, in the case of the test group, one independent variable is manipulated while holding all other variables constant. In the case of the control group, no changes in any variables are made. Thereafter, measurements are made of both groups. Any changes in readings in the test group that were not in the control group can therefore be attributed to the change in the independent variable. One of my classmates used experimental research to measure the impact of certain advertisements on consumers' perceptions. He called in two sets of people—the test group and control group. First, he took measurement of their perceptions. Certain advertisements were shown to only the test group but not to the control group. He then took measurements of perceptions of both groups. He found that there were some changes in perceptions of the test group, but not of the control group, after he showed them the advertisements. Thus he could confidently say that the changes in perceptions were due to the advertisements. The objection towards experimental research is that it lacks external validity. This is because people are subjected to artificial conditions. They may not react the same way in real life. The next type of research, ex post facto, is more popular among social science researchers.

Ex post facto research takes measurements only once. No control of variables is attempted. Therefore, no causal relationship can be tested. Why is it then the more popular choice of research? This is probably because it is much simpler to conduct and respondents can be more easily accessed and at lower cost compared with experiments. Furthermore, there are a wide variety of ways in which data can be collected in ex post facto research. For instance, data can be collected via survey or observation. Via survey, data can be collected by questionnaires or by personal interviews. Self-administered questionnaires can be sent through the post or via email, or they can be completed in the presence of the researcher. Personal interviewing can be done face to face or by telephone. Face-to-face interviewing can be conducted by mall intercept, door to door or at the office by prior appointment. There are obvious advantages and disadvantages of each method. For instance, questionnaires sent by post can be relatively cheap, time saving and can cover wide geographical regions. However, they are notorious for their low response rate. The benchmark, i.e. acceptable response rate, has been said to be around 20 per cent. While this may be more easily achieved in countries such as Britain, it is an almost impossible task in Malaysia. I have heard some of my colleagues at University of Malaya complaining that they had sent out around 1,000 questionnaires and only managed to get around twenty-five questionnaires completed and returned (i.e. 2.5 per cent). Obviously, the low response rate can have some implications of response bias and a lack of validity, i.e. whether there are differences in characteristics between the few people who did respond, and the many who did not.

A student of mine at the University of Malaya tried to distribute her questionnaires by email. She also complained of an extremely low response rate. The filling of forms is an activity loathed by many Malaysians and this culture is not helping PhD students and other researchers at all. I feel that getting respondents to complete questionnaires in front of you is the best method of data collection, from the point of view of not only ensuring a high response rate, but also in terms of accuracy of results. This is because respondents will find it more difficult to refuse if you are physically present. Another advantage is that if the respondents are unclear of anything in the questionnaire, they can immediately seek clarification from you. In fact, I used this method of data collection in my PhD research. My collection of questionnaires took respondents approximately three hours to fill! Therefore, a lot of coaxing and persuasion had to be done on my part to ensure that the respondents completed all sections of all the questionnaires. Also, any questions raised by the respondents were immediately answered. In this way I managed to maintain the motivation and co-operation of the respondents.

Some of my classmates at the Manchester Business School collected their data by interview. They would interview the respondents at their workplace at an appointed date and time. Questions were structured (i.e. the same questions for all respondents) but open ended (i.e. respondents were free to phrase their answers as they wish). Although this method allows flexibility and the preparation of such questionnaires does not take as long (as answers do not have to be thought out by the researcher beforehand), a lot of time has to be spent subsequently in coding (i.e. translating the answers into numbers for use in a statistical software). In fact, I had only a few questions in my questionnaire that were open-ended. However, I spent much time trying to categorise certain words for the purposes of coding. It is an extremely tedious and time-consuming process and I would recommend that you should try to avoid asking open-ended questions, as much as possible.

In addition to carrying out a survey, additional data can also be collected by way of observation. Observation can be overt or covert. In overt observation the subject of observation knows that he or she is being observed, whereas in covert observation, the subject does not. Such data can be used to enhance and explain the data that were collected by survey. In my research (Ahmad, 2001), I used both overt and covert observations. After completing the questionnaire, I would follow the respondents to the shop floor and observe them at work. I also took note of the working conditions, the seating arrangements, the physical space, noise levels and the size of the team. From my observations, I was able to see that the physical arrangements allowed the assembly workers to interact with each other as a team. They were working sufficiently close to each other and the noise levels were sufficiently low to enable them to have a decent conversation with each other. However, what I also observed through overt and covert observations, was that although they were close to each other, the workers tended not to interact very much with each other. In fact, one of the supervisors said that the team members were working rather independently of each other. Thus valuable information can be obtained through observation, which can be used to support or supplement data collected by survey. So, even though most of my data were collected by closed-ended survey questionnaires, I supplemented the same with data collected by observation and even by some open-ended interviews with a few supervisors.

The research methodology chapter should usually end with the development and enumeration of hypotheses. Having said this, however, I do recall having seen PhD theses without a single hypothesis stated. So it appears that hypotheses are not mandatory in PhD theses. However, I have been informed by several professors that the more ‘modern’ approach for want of a better word, is to have the hypotheses stated prior to data analyses. As a rule of thumb, a thesis should contain at least two but not more than seven hypotheses. However, this is only a general guide. A thesis can have more than seven hypotheses if the student can show that they are actually related and can be grouped in say, three themes. For example, there can be three main hypotheses but each hypotheses is subdivided into four hypotheses, making twelve sub-hypotheses in total (Ahmad, 2001). A thesis that has thirty hypotheses or more will probably leave the examiner with an impression that the researcher lacks focus. The rationale behind insisting that there should not be too many hypotheses is that a PhD thesis should be focussed, i.e. it have considerable depth, not width.

The population, sampling method and the sample itself are usually described in the research methodology chapter. It may however, be placed in a different chapter if the student wishes to elaborate on the issue at length and the research methodology chapter is already too lengthy. For instance, I placed it in a different chapter with the headings as follows:

Chapter 4: Selection of Companies and Data Collection
Part 1: Criteria for selection
4.2 Criteria for selection
4.3 Rationale for the choice of only one occupation and a few companies
Part 2: Data collection
4.4 Introduction: Chronology of events
4.5 Pilot testing
4.6 Selection of companies
4.7 Brief description of the companies
4.8 Collection of qualitative data
4.9 Selection of the sample
4.10 Data collection methods

4.11 The data
4.12 Conclusion

A common mistake by students is to cite articles relating to methodology (such as the questionnaires) in the literature review when they should be in the research methodology chapter. Similarly, students sometimes make the mistake of citing articles relating to the variables in the research methodology chapter when they should be placed in the literature review chapter.

Chapter 4: Results

This usually begins with a statement of the preliminary analyses i.e. checking the missing data, linearity of distribution and homogeneity of variance. It should briefly describe the analyses used; for example, if the researcher used regression analyses, the regression equation should be stated. Results of analyses should be reported in tables and salient points reported in the text.

If analyses are many and detailed, it is better to present them in separate chapters. Thus there may be more than one results chapter. For instance, in Ahmad (2001), there were three separate results chapters and these were divided according to themes. For instance, the first results chapter contains the report of the preliminary analyses, basic statistics (frequencies, mean) and correlation. The second results chapter contains the hierarchical regression analyses pertaining to complementary fit. The final results chapter contains analyses pertaining to supplementary fit.

Graphs and tables should obviously be included in these chapters complete with numbered headings. However, what is less obvious is, how much should be included? Should the SPSS printouts be included? The answer is no, SPSS printouts should never be inserted directly in the chapters. Rather, only salient or pertinent figures should be inserted and presented in tables in the chapters. Computer printouts should only be included in the appendices. Furthermore, you may need to exercise your discretion on what printouts to include if there are too many. As a rule of thumb, appendices should never form more than half of the bound volume of the thesis. You cannot impress examiners with volumes of computer printouts. Essentially, appendices should be provided with the purpose of enabling the examiner to search for more information if he or she so desires. Thus the student should be able to discern the more important ones from those less so.

Chapter 5: Conclusion

This should be a summary of all the previous chapters. The conclusion also reiterates what you set out to do from the beginning (i.e. your objectives) and assesses to what extent you were successful. Often the conclusion ends with a statement of the limitations of the study as well as proposals for future research. Usual limitations mentioned are:

  • Having a small sample size.
  • The study is conducted in a limited number of companies and occupations.
  • The study is conducted in one country only.

The above limitations usually mean that the results of the findings cannot be generalised to the rest of the population, or to other populations. However, this in itself is not fatal. In fact, you may be pleasantly surprised by the fact that the lack of external validity (inability to generalise findings) is a rather common limitation among PhDs in the social sciences.

A very important point that I wish to stress is that you should only make conclusions based on your data. I have seen students' theses, where the data points to ‘A’ but they conclude ‘B’ based on their own feelings and intuitions. If you are going to rely on intuitions, why collect data in the first place? So, do not make conclusions beyond your data, or worse still, do not make conclusions which actually contradict your results. This might sound terribly obvious and no one will make that mistake, you say. Believe me, I have seen that happen more than once.

Other Sections

In addition to the main sections mentioned above, your thesis should contain, at the beginning, an abstract that summarises your work. My supervisor advised me that the abstract should be written in such a way that readers would be able to know and understand what is contained in your thesis without having to read it fully! This means that it must be a concise, yet a complete summary of what you have done. The usual mistake made in the abstract is not summarising the results. The abstract should summarise the entire thesis. I feel that the easiest way for you to understand this is to have a look at other theses.

In addition to the abstract, there are the acknowledgements, layout of title, and contents pages. Normally there should be separate contents pages for tables, diagrams and appendices. Also there are requirements as to the width of the margins, paragraph spacing, type and size of fonts. You should consult your university's regulations regarding these. Another important section is the bibliography or references section. The importance of this is often underestimated by the students. It is without a doubt the most boring part of preparing a thesis. Yet time, care and effort have to be expended in preparing it as examiners are extremely fond of picking out mistakes in it. The rationale is that if you are relying on a particular theory that has been put forward by somebody else, then due credit should be given to him or her. Also, as a courtesy to your readers, you should give them a means to be able to find more information on the study, should he or she so desire. Students should consult previous PhD theses in their universities for the acceptable format. This is because formats used in PhD theses can differ slightly among universities.

A common complaint of supervisors is that their students' theses are ‘dry as dust’ or good cures for insomnia. I remember that when I was doing my PhD in Manchester, my late father told me that he was very interested in reading a PhD thesis. So I lent him one that was written on leadership. However, after a few days he complained that it was extremely dry. He stopped after having read only one third of the thesis. It is a fair statement to make that in general, theses are boring to the layman as they are rather technical. They are nothing like novels. So, do not be put off if your supervisor complains that your thesis is boring. It may be boring because he or she has to read it instead of being able to take a walk down to the pub. However, what you should be more concerned with is whether your thesis has sufficient merit for you to be awarded a PhD.

Here is a checklist of what must be included in the different chapters of a thesis:

  1. The introduction should contain the uniqueness, significance and contribution of the study and why the study should be conducted—i.e. there should be a theoretical contribution and a practical contribution.
  2. The main variables should be introduced here, setting the scene.
  3. Do not discuss at length in the introduction, variables which are not investigated in the current research.
  4. The literature review should discuss in depth all the variables which are subjects of the study.
  5. The literature review should also explain why and how all the dependent and independent variables in the study are related to one another.
  6. The literature review should be a critical review rather than mere recitation of previous research.
  7. The literature review should identify a gap in previous research. 8. If possible, identify a methodological weakness in previous research and how that has been overcome in your research. (In my opinion, this is preferable but not mandatory. However, some examiners may feel that it is mandatory and it is better to be safe than sorry).
  8. The research methodology should contain a diagram of the conceptual model, which should be described.
  9. The research methodology should cite articles reporting previous research which have made use of the questionnaires you have adopted in your research (if you use questionnaires)—this will justify the use of the questionnaires.
  10. The research methodology should also cite articles reporting previous research from which you have adopted your research design—this will justify your choice of the research design.
  11. The hypotheses must be formulated in accordance with your research questions and theoretical model.
  12. The items in your questionnaires should relate to your hypotheses.
  13. Descriptive statistics such as frequencies, means, standard deviation and normality of distribution should be conducted. 15. Reliability and validity tests should also be conducted.
  14. More importantly, your results chapter should describe statistical tests that you have conducted in order to find support for your hypotheses.
  15. Report explicitly whether the results support the hypotheses or not—do not leave it to the examiners to figure it out for themselves.
  16. Report whether the results are consistent or contrary to your expectations; if they are contrary, explain why.
  17. The conclusions and recommendations should be based purely on the results obtained and not on the student's intuition. However, the student may attribute certain factors to be the cause, but the student must make it explicit that it is only a suggestion on the part of the student and the research can in no way confirm that. In other words the student can only offer a suggestion as a possible explanation of the events, although the research cannot confirm the same.
  18. The bibliography conforms to APA (American Psychological Association) style or whatever stipulated style.
  19. Ensure that all the articles mentioned in the thesis are listed in the bibliography and there are no articles which have been mentioned in the text but are not in the bibliography (NIBs—Not in bibliography).
  20. Also ensure that the bibliography does not give the citation of articles which have not been mentioned anywhere in the thesis (NITs—Not in text).

Although at the time of writing this checklist I have tried to come up with a comprehensive list, I cannot help but feel that I may have left something out or that new criteria may have arisen. So my advice to you is that the above list should be treated as containing the more obvious criteria, but should not be treated as exhaustive. It is absolutely critical that you should at the very least, make sure that you have all the items in this checklist, before submitting your thesis.

About this article

The Structure of the PhD

Updated About content Print Article