How to Overcome Difficulties With a Doctorate or Graduate School
Aug 12, 2009 By Alistair McCulloch
When students begin to study for a PhD or enters a new graduate school, they begin a highly significant period in their life. A PhD, or doctorate, involves making a contribution to knowledge and can also result in significant changes in the way the student thinks about the world.
However, because a PhD involves this sort of transformational experience, and also because it takes a minimum of three years and sometimes more than ten, real problems can emerge during the period of study. Here are the main problems with studying for a PhD.
Problems With Identifying a Good Research Topic
A good research topic is vital to a good PhD. It must be rooted in the academic literature, as this is one of the functions of the literature review, to identify an issue that no-one else has approached in the way the student intends.
Loss of Motivation
The best way to avoid loss of motivation is for students to really know why they want to do a PhD and to select a topic, institution and supervisor that they know will excite them for the long run. Understanding motivation will also help the institution (through the supervisor) support and encourage the student in the best way.
Experiments That Don’t Work
All scientists know the frustration that can arise when an experiment doesn’t work. If this happens to a research student, he should seek help from his fellow students, from the supervisor, from other researchers in the department and also from scientists working at other Universities and research laboratories.
Problems With Data-Collection
Sometimes social science research students face problems collecting their data. This can be due to difficulties in recruiting respondents. Sometimes their responses will not allow students to address their research questions in the way they had expected. If numbers of respondents are insufficient, the selection criteria may be extended or the number of case studies increased. If the responses are not what was expected, there may be an opportunity to revise the theory or model being tested and this may result in a better thesis.
Isolation
The further a student gets into his PhD, the more likely he is to experience isolation, a consequence of studying something completely original. Joining student societies and trying to maintain a life outside university are two ways of combating this. If there’s no doctoral student society in place, a student could try to organise one.
Running out of Money
PhD students on grants often find it very difficult to complete their studies once their grant has run out, and part-time students are likely to find it hard to make ends meet throughout their studies. It is possible to suspend a doctoral registration, but it is not uncommon that students who do this find it hard to re-start. If it is essential that a suspension be put in place, students should keep in touch with both their research project and their supervisor during the period of suspension. Maintaining links improves the chances of eventual completion considerably.
Problems With a Supervisor
Problems with a research degree supervisor can arise from overwork, personality clashes, a change of topic or a multitude of reasons. Whatever the reason, it is important that the student discuss the situation openly with the supervisor. If that is impossible, go to the Departmental Research Degree Coordinator or the Graduate School. If none of these are possible, talk with the student union or student association. It is a student’s right to have good supervision and a university’s failure to provide it leaves it liable.
What if Someone Publishes a Student’s Original Idea Before He Does?
A student shouldn’t worry about this because it very rarely happens. If it does, and it is worth reiterating that the chance of someone reproducing independently a student’s programme of work is very, very slim. In that case, re-orientate the thesis so that it focuses on a different aspect of the research. Think about drawing on a different aspect of the literature review or do some additional empirical work to move the idea on. The latter option may not be a satisfactory situation, but it may well save the doctorate the student has put so much work into.
Read more at Suite101: Overcoming Problems in Studying for a PhD: How to Overcome Difficulties With a Doctorate or Graduate School http://graduate-schools.suite101.com/article.cfm/overcoming_problems_in_studying_for_a_phd#ixzz0wpSpZoQ1
Showing posts with label PhD Preparation. Show all posts
Showing posts with label PhD Preparation. Show all posts
Tuesday, August 17, 2010
Tuesday, May 4, 2010
Factor Analysis Tutorial Class
Posted by Prof. Ghazali
Dear All,
For the 23rd Mei meeting, I am trying to get someone to show us how to run factor analysis. So I'd like to encourage those of you who are interested to bring your laptop, with the SPSS software installed. There shall be a short presentation by Mr Lum on his thesis.
Prof Ghazali
Dear All,
For the 23rd Mei meeting, I am trying to get someone to show us how to run factor analysis. So I'd like to encourage those of you who are interested to bring your laptop, with the SPSS software installed. There shall be a short presentation by Mr Lum on his thesis.
Prof Ghazali
OUM Phd(BA) Discussion Session (May-Oct 2010)
Posted by Chia Bee Ching
Dear All,
For those who have attended the 25th April Informal discussion session, we all found the session beneficial, relevant and engaging. For many of us, the junior learners, that was the first time we ever listened to a mock viva presentation (by Richard Ng).
The short technical report writing talk (a copy of the the ppf was forwarded to all by member, Sallah Jantan) given by Prof. Junaidah is neat, clear and precise. The message for us is to start writing, rewriting and rewriting.
Following the Q&A in the April's discussion session, the participants' consensus to keep the learning momentum on, we are happy to announce the confirmed dates for next six months informal discussion session are as stated below. Prof Ghazali will be there to monitor and co-ordinate.
May 23-05-10 9am-12.30pm
June 27-06-10 9am-12.30pm
July 25-07-10 9am-12.30pm
August no meeting OUM Colloquium
September 26-09-10 9am-12.30pm
October 24-10-10 9am-12.30pm
Venue: The seminar room (GA-01, beside theatrette) OUM main campus.
RSVP: beeching chia (mobile 0173402968)
Remark: If any member who likes to present a mock QE or to share the research problems, progress and issues, please liaise with Prof Ghazali, our Phd(BA) Coordinator.
Thank you
Dear All,
For those who have attended the 25th April Informal discussion session, we all found the session beneficial, relevant and engaging. For many of us, the junior learners, that was the first time we ever listened to a mock viva presentation (by Richard Ng).
The short technical report writing talk (a copy of the the ppf was forwarded to all by member, Sallah Jantan) given by Prof. Junaidah is neat, clear and precise. The message for us is to start writing, rewriting and rewriting.
Following the Q&A in the April's discussion session, the participants' consensus to keep the learning momentum on, we are happy to announce the confirmed dates for next six months informal discussion session are as stated below. Prof Ghazali will be there to monitor and co-ordinate.
May 23-05-10 9am-12.30pm
June 27-06-10 9am-12.30pm
July 25-07-10 9am-12.30pm
August no meeting OUM Colloquium
September 26-09-10 9am-12.30pm
October 24-10-10 9am-12.30pm
Venue: The seminar room (GA-01, beside theatrette) OUM main campus.
RSVP: beeching chia (mobile 0173402968)
Remark: If any member who likes to present a mock QE or to share the research problems, progress and issues, please liaise with Prof Ghazali, our Phd(BA) Coordinator.
Thank you
Monday, January 19, 2009
A STRUCTURED APPROACH TO PRESENTING PhD THESES: NOTES FOR CANDIDATES AND THEIR SUPERVISORS
Posted by Chad Perry
Paper presented to the ANZ Doctoral Consortium, University of Sydney, February 1994, with later additions to 18 September 1995.
Abstract
This paper addresses the problem: how should a PhD candidate in management or a similar field (and his or her supervisor) present the thesis? The structure developed provides a starting point for understanding what a PhD thesis should set out to achieve, and also provides a basis for communication between a candidate and his or her supervisor. Firstly, criteria for judging a PhD thesis are reviewed and justification for its structure is provided. Then writing style is considered. Finally, each of the five chapters and their sections are described in some detail: introduction, literature review, methodology, analysis of data, and conclusions and implications.
Acknowledgments
These notes were originally based on ideas of Drs Geoff Meredith, Bert Cunnington and Mike Watkins and also on University of Oregon (n.d.). However, views and errors are the writer's own. He has written the paper with a beginning PhD candidate in mind, and so has presented some positions as starting points for drafting a thesis rather than as the only positions that can be adopted. He thanks Drs Kwaku Atahuene-Gima, Robert Brown, Alan Buttery, Gail Craswell, Hank Johnson, Di Lewis, Estelle Phillips, John Roberts and John Rossiter, and Barry Bell, Diana Best, Claudia Hope and Tony Ward for commenting on earlier drafts, and thanks Barry Bell, Len Coote, June Dunleavy, Marliyn Healy, John Jackson, Ben Lyttle, Cec Pederson, Tony Ward and Vicky Schinkel for ideas for some examples.
INTRODUCTION
Ideally, PhD research in management or a related field should:
* cover a field which fascinates the candidate sufficiently for him or her to endure years of hard and solitary work;
* build on the candidate's previous studies, for example, his or her course work in a Master's degree;
* be in an area of `warm' research activity rather than in a `cold', overworked area or in a `hot', too-competitive, soon-to-be extinguished area;
* be in an area near the main streams of a discipline and not at the margins of a discipline or straddling two disciplines - being near the main streams makes it easier to find thesis examiners, to gain academic positions, and to get acceptance of journal articles about the research;
* be manageable, producing interesting results and a thesis in the shortest time possible;
* have accessible sources of data;
* open into a program of research projects after the PhD is completed; and
* provide skills and information for obtaining a job in a non-research field, if a research or academic job is not available or not desired.
Whatever research a PhD candidate finally chooses to do, he or she must record the research in a thesis. This note is written for PhD candidates in management or a related field and their supervisors and outlines a structure for a five chapter PhD thesis. The structure is summarised in figure 1 and in table 1. Other writers have provided general procedures for the many parts of the PhD research process (for example, Davis & Parker 1979; Phillips & Pugh 1987; Perry 1990), but these notes concentrate on the PhD thesis itself and do so more comprehensively and with far more examples than other writers (for example, Clark 1986; Pratt 1984; Witcher 1990). That is, this paper addresses the problem:
How should PhD candidates and their supervisors present the thesis?
This problem is clearly important for PhD candidates. Many universities provide little guidance to candidates, prompting the criticism that, at one university, `the conditions for the award of degrees in the Graduate Study section of the calendar give more precise information on the size of the paper to be used and the margins to be left on each side of the sheet than on the university's understanding of what a thesis is' (Massingham 1984, p. 15). By using the structure developed below, a candidate will ensure his or her PhD thesis demonstrates the key requirements of a PhD thesis (Moses 1985):
* a distinct contribution to a body of knowledge through an original investigation or testing of ideas, worthy in part of publication (see chapter 5 described below) -this is usually the most important criterion for a PhD;
* competence in research processes, including an understanding of, and competence in, appropriate research techniques and an ability to report research (see chapters 3 and 4, plus the whole report format); and
* mastery of a body of knowledge , including an ability to make critical use of published work and source materials (see chapter 2) with an appreciation of the relationship of the special theme to the wider field of knowledge (see chapters 2 and 5).
The candidate should ask to see a copy of the letter sent to examiners to determine the priorities of his or her faculty for the three criteria above and if the faculty has additional criteria (Nightingale 1992). As well, a supervisor may be able to produce copies of previous examiners' reports.
The foundations for the structured approach were the writer's own doing, supervising, examining and adjudicating conflicting examiners' reports of many master's and PhD theses in management and related fields at several Australian universities, and examining requests for transfer from master's to PhD research, together with comments from the people listed in the acknowledgments section.
The paper has two parts. Firstly, the five chapter structure is introduced, possible changes to it are justified and writing style is considered. In the second part, each of the five chapters and their sections are described in some detail: introduction, literature review, methodology, analysis of data, and findings and implications.
Delimitations.
The structured approach may be limited to PhDs in management areas such as marketing and strategic management which involve common quantitative and qualitative methodologies. That is, the structure may not be appropriate for PhDs in other areas or for management PhDs using relatively unusual methodologies such as historical research designs. Moreover, the structure is a starting point for thinking about how to present a thesis rather than the only structure which can be adopted, and so it is not meant to inhibit the creativity of PhD researchers. Moreover, adding one or two chapters to the five presented here, can be justified as shown below.
Another limitation of the approach is that it is restricted to presenting the final thesis. This paper does not address the techniques of actually writing a thesis (however, appendix I describes three little-known keys to writing a thesis). Moreover, the approach in this paper does not refer to the actual sequence of writing the thesis, nor is it meant to imply that the issues of each chapter have to be addressed by the candidate in the order shown. For example, the hypotheses at the end of chapter 2 are meant to appear to be developed as the chapter progresses, but the candidate might have a good idea of what they will be before he or she starts to write the chapter. And although the methodology of chapter 3 must appear to be been selected because it was appropriate for the research problem identified and carefully justified in chapter 1, the candidate may have actually selected a methodology very early in his or her candidature and then developed an appropriate research problem and justified it. Moreover, after a candidate has sketched out a draft table of contents for each chapter, he or she should begin writing the `easiest parts' of the thesis first as they go along, whatever those parts are - and usually introductions to chapters are the last to written (Phillips & Pugh 1987, p. 61). But bear in mind that the research problem, limitations and research gaps in the literature must be identified and written down before other parts of the thesis can be written, and section 1.1 is one of the last to be written. Nor is this structure meant to be the format for a PhD research proposal - one proposal format is provided in Parker and Davis (1979), and another related to the structure developed below is in appendix II. How to write an abstract of a thesis is described in appendix VI.
Table 1 Sequence of a five chapter PhD thesis
Title page
Abstract (with keywords)
Table of contents
List of tables
List of figures
Abbreviations
Statement of original authorship
Acknowledgments
1 Introduction
1.1 Background to the research
1.2 Research problem and hypotheses
1.3 Justification for the research
1.4 Methodology
1.5 Outline of the report
1.6 Definitions
1.7 Delimitations of scope and key assumptions
1.8 Conclusion
2 Literature review
2.1 Introduction
2.2 (Parent disciplines/fieldss and classification models)
2.3 (Immediate discipline, analytical models and research questions or hypotheses)
2.4 Conclusion
3 Methodology
3.1 Introduction
3.2 Justification for the paradigm and methodology
3.3 (Research procedures)
3.4 Ethical considerations
3.5 Conclusion
4 Analysis of data
4.1 Introduction
4.2 Subjects
4.3 (Patterns of data for each research question or hypothesis)
4.4 Conclusion
5 Conclusions and implications
5.1 Introduction
5.2 Conclusions about each research question or hypothesis
5.3 Conclusions about the research problem
5.4 Implications for theory
5.5 Implications for policy and practice
5.5.1 Private sector managers
5.5.2 Public sector policy analysts and managers
5.6 Limitations
5.7 Further research
Bibliography
Appendices
BASICS OF STRUCTURE AND STYLE
A five chapter structure can be used to effectively present a PhD thesis, and it is summarised in figure 1 and table 1.
Figure 1 Model of the chapters of a PhD thesis
In brief, the thesis should have a unified structure (Easterby-Smith et al. 1991). Firstly, chapter 1 introduces the core research problem and then `sets the scene' and outlines the path which the examiner will travel towards the thesis' conclusion. The research itself is described in chapters 2 to 5:
* the research problem and hypotheses arising from the body of knowledge developed during previous research (chapter 2),
* methods used in this research to collect data about the hypotheses (chapter 3),
* results of applying those methods in this research (chapter 4), and
* conclusions about the hypotheses and research problem based on the results of chapter 4, including their place in the body of knowledge outlined previously in chapter 2 (chapter 5).
This five chapter structure can be justified. Firstly, the structure is a unified and focussed one, and so addresses the major fault of postgraduate theses in a survey of 139 examiners' reports, that is, it clearly addresses those examiners' difficulty in discerning what was the `thesis' of the thesis? `Supervisors need to emphasise throughout students' candidacies that they are striving in the thesis to communicate one big idea' (Nightingale 1984, p. 174), and that one big idea is the research problem stated on page 1 or 2 of the thesis and explicitly solved in chapter 5. Easterby-Smith et al. (1991) also emphasise the importance of consistency in a PhD thesis, and Phillips and Pugh (1987, p. 38) confirm that a thesis must have a thesis or a `position'. Secondly, the structure carefully addresses each of the 31 requirements of an Australian PhD thesis outlined by an authority in a publication of the Higher Education Research and Development Society of Australia (Moses 1985, pp. 32-34). Thirdly, the structure is explicitly or implicitly followed by many writers of articles in prestigious academic journals such as The Academy of Management Journal and Strategic Management Journal (for example, Datta et al. 1992). Fourthly, the structure has been the basis of several PhD and masters theses at Australian universities that were completed in minimum time and passed by examiners with none or negligible revisions required. Fifthly, the structure is much like that which will be used by candidates later in their career, to apply for research grants (Krathwohl 1977; Poole 1993). Finally, by reducing time wasted on unnecessary tasks or on trying to demystify the PhD process, the five chapter structure provides a mechanism to shorten the time taken to complete a PhD, an aim becoming desired in many countries (Cude 1989).
Justified changes to the structure. Some changes to the five chapter structure could be justified. For example, a candidate may find it convenient to expand the number of chapters to six or seven because of unusual characteristics of the analysis in his or her research; for example, a PhD might consist of two stages: some qualitative research reported in chapters 3 and 4 of the thesis described below, which is then followed by some quantitative research to refine the initial findings reported in chapters 5 and 6; the chapter 5 described below would then become chapter 7. In addition, PhD theses at universities that allow its length to rise from a minimum length of about 50 000 to 60 000 words (Phillips & Pugh 1987), say, through a reasonable length of about 70 000 to 80 000 words, up to the upper limit 100 000 words specified by some established universities like the University of Queensland and Flinders University, may have extra chapters added to contain the extended reviews of bodies of knowledge in those huge theses. In brief, in some theses, the five chapters may become five sections with one or more chapters within each of them, but the principles of the structured approach should remain. That is, PhD research must remain an essentially creative exercise. Nevertheless, the five chapter structure provides a starting point for understanding what a PhD thesis should set out to achieve, and also provides a basis for communication between a candidate and his or her supervisor.
As noted above, the five chapter structure is primarily designed for PhDs in management or related fields using common methodologies. Qualitative methodologies such as case studies and action research (Perry & Zuber-Skerritt 1992; 1994) can fit into the structure, with details of how the case study or the action research project being presented in chapter 3 and case study details or the detailed report of the action research project being placed in appendices. In theses using the relatively qualitative methodologies of case studies or action research, the analysis of data in chapter 4 becomes a categorisation of data in the form of words, with information about each research question collected together with some preliminary reflection about the information. That is, the thesis still has five chapters in total, with chapter 4 having preliminary analysis of data and chapter 5 containing all the sections described below. Appendix III discusses in more detail the difficult task of incorporating an action research project into a PhD thesis.
Links between chapters. Each chapter described below should stand almost alone. Each chapter (except the first) should have an introductory section linking the chapter to the main idea of the previous chapter and outlining the aim and the organisation of the chapter. For example, the core ideas in an introduction to chapter 3 might be:
Chapter 2 identified several research questions; chapter 3 describes the methodology used to provide data to investigate them. An introduction to the methodology was provided in section 1.4 of chapter 1; this chapter aims to build on that introduction and to provide assurance that appropriate procedures were followed. The chapter is organised around four major topics: the study region, the sampling procedure, nominal group technique procedures, and data processing.
The introductory section of chapter 5 (that is, section 5.1) will be longer than those of other chapters, for it will summarise all earlier parts of the thesis prior to making conclusions about the research described in those earlier parts; that is, section 5.1 will repeat the research problem and the research questions/hypotheses. Each chapter should also have a concluding summary section which outlines major themes established in the chapter, without introducing new material. As a rough rule of thumb, the five chapters have these respective percentages of the thesis' words: 5, 30, 15, 25 and 25 percent.
Style
As well as the structure discussed above, examiners also assess matters of style (Hansford & Maxwell 1993). Within each of the chapters of the thesis, the spelling, styles and formats of Style Manual (Australian Government Publishing Service 1988) and of the Macquarie Dictionary should be followed scrupulously, so that the candidate uses consistent styles from the first draft and throughout the thesis for processes such as using bold type, underlining with italics, indenting quotations, single and double inverted commas, making references, spaces before and after side headings and lists, and gender conventions. Moreover, using the authoritative Style Manual provides a defensive shield against an examiner who may criticise the thesis from the viewpoint of his or her own idiosyncratic style. Some pages of Style Manual (Australian Government Publishing Service 1988) which are frequently used by PhD candidates are listed in appendix IV.
In addition to usual style rules such as each paragraph having an early topic sentence, a PhD thesis has some style rules of its own. For example, chapter 1 is usually written in the present tense with references to literature in the past tense; the rest of the thesis is written in the past tense as it concerns the research after it has been done, except for the findings in chapter 5 which are presented in the present tense. More precisely for chapters 2 and 3, schools of thought and procedural steps are written of in the present tense and published researchers and the candidate's own actions are written of in the past tense. For example: 'The eclectic school has [present] several strands. Smith (1990) reported [past] that...' and `The first step in content analysis is [present] to decide on categories. The researcher selected [past] ten documents...'
In addition, value judgements and words should not be used in the objective pursuit of truth that a thesis reports. For example, `it is unfortunate', `it is interesting', `it is believed', and `it is welcome' are inappropriate. Although first person words such as `I' and `my' are now acceptable in a PhD thesis (especially in chapter 3 of a thesis within the interpretive paradigm), their use should be controlled - the candidate is a mere private in an army pursuing truth and so should not overrate his or her importance until the PhD has been finally awarded. In other words, the candidate should always justify any decisions where his or her judgement was required (such as the number and type of industries surveyed and the number of points on a likert scale), acknowledging the strengths and weaknesses of the options considered and always relying upon as many references as possible to support the decision made. That is, authorities should be used to back up any claim of the researcher, if possible. If the examiner wanted to read opinions, he or she could read letters to the editor of a newspaper.
Moreover, few if any authorities in the field should be called `wrong', at the worst they might be called `misleading'; after all, one of these authorities might be an examiner and have spent a decade or more developing his or her positions and so frontal attacks on those positions are likely to be easily repulsed. Indeed, the candidate should try to agree with the supervisor on a panel of likely people from which the university will select the thesis examiner so that only appropriate people are chosen. After all, a greengrocer should not examine meat products and an academic with a strong positivist background is unlikely to be an appropriate examiner of a qualitative thesis, for example (Easterby-Smith et al. 1991), or an examiner that will require three research methods is not chosen for a straightforward thesis with one. That is, do not get involved in the cross fire of `religious wars' of some disciplines. Moreover, this early and open consideration of examiners allows the candidate to think about how his or her ideas will be perceived by likely individual examiners and so express the ideas in a satisfactory way, for example, explain a line of argument more fully or justify a position more completely. That is, the candidate must always be trying to communicate with the examiners in an easily-followed way.
This easily-followed communication can be achieved by using several principles. Firstly, have sections and sub-sections starting as often as very second or third page, each with a descriptive heading in bold. Secondly, start each section or sub-section with a phrase or sentence linking it with what has gone before, for example, a sentence might start with `Given the situation described in section 2.3.4' or `Turning from international issues to domestic concerns,...' Thirdly, briefly describe the argument or point to be made in the section at its beginning, for example, `Seven deficiencies in models in the literature will be identified'. Fourthly, make each step in the argument easy to identify with a key term in italics or the judicious use of `firstly', `secondly', or `moreover', `in addition', `in contrast' and so on. Finally, end each section with a summary, to establish what it has achieved; this summary sentence or paragraph could be flagged by usually beginning it with `In conclusion,..' or `In brief,...' In brief, following these five principles will make arguments easy to follow and so guide the examiner towards agreeing with a candidate's views.
Another PhD style rule is that the word `etc' is too imprecise to be used in a thesis. Furthermore, words such as `this', `these', `those' and `it' should not be left dangling - they should always refer to an object; for example, `This rule should be followed' is preferred to `This should be followed'. Some supervisors also suggest that brackets should rarely be used in a PhD thesis - if a comment is important enough to help answer the thesis' research problem, then it should be added in a straightforward way and not be hidden within brackets as a minor concern to distract the examiner away from the research problem.
As well, definite and indefinite articles should be avoided where possible, especially in headings; for example, `Supervision of doctoral candidates' is more taut and less presumptuous than `The supervision of doctoral candidates'. Paragraphs should be short; as a rule of thumb, three to four paragraphs should start on each page if my preferred line spacing of 1.5 and Times Roman 12 point font is used to provide adequate structure and complexity of thought on each page. (A line spacing of 2 and more paragraphs per page make a thesis appear disjointed and `flaky', and a sanserif font is not easy to read.) A final note of style is that margins should be those nominated by the university or those in Style Manual (Australian Government Publishing Service 1988).
The above comments about structure and style correctly imply that a PhD thesis with its readership of three examiners is different from a book which has a very wide readership (Derricourt 1992), and from shorter conference papers and journal articles which do not require the burden of proof and references to broader bodies of knowledge required in PhD theses. Candidates should be aware of these differences and could therefore consider concentrating on completing the thesis before adapting parts of it for other purposes. However, it must be admitted that presenting a paper at a conference in a candidature may lead to useful contacts with the `invisible college' (Rogers 1983, p. 57) of researchers in a field, and some candidates have found referees' comments on articles submitted for publication in journals during their candidacy, have improved the quality of their thesis' analysis (and publication has helped them get a job). Nevertheless, several supervisors suggest that it is preferable to concentrate on the special requirements of the thesis and adapt it for publication after the PhD has been awarded or while the candidate has temporary thesis `writer's block'.
The thesis will have to go through many drafts (Zuber-Skerritt & Knight 1986). The first draft will be started early in the candidature, be crafted after initial mindmapping and a tentative table of contents of a chapter and a section, through the `right', creative side of the brain and will emphasise basic ideas without much concern for detail or precise language. Supervisors and other candidates should be involved in the review of these drafts because research has shown that good researchers ` require the collaboration of others to make their projects work, to get them to completion' (Frost & Stablein 1992, p. 253), and that social isolation is the main reason for withdrawing from postgraduate study (Phillips & Conrad 1992). By the way, research has also shown that relying on just one supervisor can be dangerous (Conrad, Perry & Zuber-Skerritt 1992; Phillips & Conrad 1992).
Facilitating the creative first drafts of sections, the relatively visible and structured `process' of this paper's structure allows the candidate to be more creative and rigorous with the `content' of the thesis than he or she would otherwise be. After the first rough drafts, later drafts will be increasingly crafted through the `left', analytical side of the brain and emphasise fine tuning of arguments, justification of positions and further evidence gathering from other research literature.
Further reading: http://www.elec.uq.edu.au/doc/Thesis_guide/phdth1.html
Paper presented to the ANZ Doctoral Consortium, University of Sydney, February 1994, with later additions to 18 September 1995.
Abstract
This paper addresses the problem: how should a PhD candidate in management or a similar field (and his or her supervisor) present the thesis? The structure developed provides a starting point for understanding what a PhD thesis should set out to achieve, and also provides a basis for communication between a candidate and his or her supervisor. Firstly, criteria for judging a PhD thesis are reviewed and justification for its structure is provided. Then writing style is considered. Finally, each of the five chapters and their sections are described in some detail: introduction, literature review, methodology, analysis of data, and conclusions and implications.
Acknowledgments
These notes were originally based on ideas of Drs Geoff Meredith, Bert Cunnington and Mike Watkins and also on University of Oregon (n.d.). However, views and errors are the writer's own. He has written the paper with a beginning PhD candidate in mind, and so has presented some positions as starting points for drafting a thesis rather than as the only positions that can be adopted. He thanks Drs Kwaku Atahuene-Gima, Robert Brown, Alan Buttery, Gail Craswell, Hank Johnson, Di Lewis, Estelle Phillips, John Roberts and John Rossiter, and Barry Bell, Diana Best, Claudia Hope and Tony Ward for commenting on earlier drafts, and thanks Barry Bell, Len Coote, June Dunleavy, Marliyn Healy, John Jackson, Ben Lyttle, Cec Pederson, Tony Ward and Vicky Schinkel for ideas for some examples.
INTRODUCTION
Ideally, PhD research in management or a related field should:
* cover a field which fascinates the candidate sufficiently for him or her to endure years of hard and solitary work;
* build on the candidate's previous studies, for example, his or her course work in a Master's degree;
* be in an area of `warm' research activity rather than in a `cold', overworked area or in a `hot', too-competitive, soon-to-be extinguished area;
* be in an area near the main streams of a discipline and not at the margins of a discipline or straddling two disciplines - being near the main streams makes it easier to find thesis examiners, to gain academic positions, and to get acceptance of journal articles about the research;
* be manageable, producing interesting results and a thesis in the shortest time possible;
* have accessible sources of data;
* open into a program of research projects after the PhD is completed; and
* provide skills and information for obtaining a job in a non-research field, if a research or academic job is not available or not desired.
Whatever research a PhD candidate finally chooses to do, he or she must record the research in a thesis. This note is written for PhD candidates in management or a related field and their supervisors and outlines a structure for a five chapter PhD thesis. The structure is summarised in figure 1 and in table 1. Other writers have provided general procedures for the many parts of the PhD research process (for example, Davis & Parker 1979; Phillips & Pugh 1987; Perry 1990), but these notes concentrate on the PhD thesis itself and do so more comprehensively and with far more examples than other writers (for example, Clark 1986; Pratt 1984; Witcher 1990). That is, this paper addresses the problem:
How should PhD candidates and their supervisors present the thesis?
This problem is clearly important for PhD candidates. Many universities provide little guidance to candidates, prompting the criticism that, at one university, `the conditions for the award of degrees in the Graduate Study section of the calendar give more precise information on the size of the paper to be used and the margins to be left on each side of the sheet than on the university's understanding of what a thesis is' (Massingham 1984, p. 15). By using the structure developed below, a candidate will ensure his or her PhD thesis demonstrates the key requirements of a PhD thesis (Moses 1985):
* a distinct contribution to a body of knowledge through an original investigation or testing of ideas, worthy in part of publication (see chapter 5 described below) -this is usually the most important criterion for a PhD;
* competence in research processes, including an understanding of, and competence in, appropriate research techniques and an ability to report research (see chapters 3 and 4, plus the whole report format); and
* mastery of a body of knowledge , including an ability to make critical use of published work and source materials (see chapter 2) with an appreciation of the relationship of the special theme to the wider field of knowledge (see chapters 2 and 5).
The candidate should ask to see a copy of the letter sent to examiners to determine the priorities of his or her faculty for the three criteria above and if the faculty has additional criteria (Nightingale 1992). As well, a supervisor may be able to produce copies of previous examiners' reports.
The foundations for the structured approach were the writer's own doing, supervising, examining and adjudicating conflicting examiners' reports of many master's and PhD theses in management and related fields at several Australian universities, and examining requests for transfer from master's to PhD research, together with comments from the people listed in the acknowledgments section.
The paper has two parts. Firstly, the five chapter structure is introduced, possible changes to it are justified and writing style is considered. In the second part, each of the five chapters and their sections are described in some detail: introduction, literature review, methodology, analysis of data, and findings and implications.
Delimitations.
The structured approach may be limited to PhDs in management areas such as marketing and strategic management which involve common quantitative and qualitative methodologies. That is, the structure may not be appropriate for PhDs in other areas or for management PhDs using relatively unusual methodologies such as historical research designs. Moreover, the structure is a starting point for thinking about how to present a thesis rather than the only structure which can be adopted, and so it is not meant to inhibit the creativity of PhD researchers. Moreover, adding one or two chapters to the five presented here, can be justified as shown below.
Another limitation of the approach is that it is restricted to presenting the final thesis. This paper does not address the techniques of actually writing a thesis (however, appendix I describes three little-known keys to writing a thesis). Moreover, the approach in this paper does not refer to the actual sequence of writing the thesis, nor is it meant to imply that the issues of each chapter have to be addressed by the candidate in the order shown. For example, the hypotheses at the end of chapter 2 are meant to appear to be developed as the chapter progresses, but the candidate might have a good idea of what they will be before he or she starts to write the chapter. And although the methodology of chapter 3 must appear to be been selected because it was appropriate for the research problem identified and carefully justified in chapter 1, the candidate may have actually selected a methodology very early in his or her candidature and then developed an appropriate research problem and justified it. Moreover, after a candidate has sketched out a draft table of contents for each chapter, he or she should begin writing the `easiest parts' of the thesis first as they go along, whatever those parts are - and usually introductions to chapters are the last to written (Phillips & Pugh 1987, p. 61). But bear in mind that the research problem, limitations and research gaps in the literature must be identified and written down before other parts of the thesis can be written, and section 1.1 is one of the last to be written. Nor is this structure meant to be the format for a PhD research proposal - one proposal format is provided in Parker and Davis (1979), and another related to the structure developed below is in appendix II. How to write an abstract of a thesis is described in appendix VI.
Table 1 Sequence of a five chapter PhD thesis
Title page
Abstract (with keywords)
Table of contents
List of tables
List of figures
Abbreviations
Statement of original authorship
Acknowledgments
1 Introduction
1.1 Background to the research
1.2 Research problem and hypotheses
1.3 Justification for the research
1.4 Methodology
1.5 Outline of the report
1.6 Definitions
1.7 Delimitations of scope and key assumptions
1.8 Conclusion
2 Literature review
2.1 Introduction
2.2 (Parent disciplines/fieldss and classification models)
2.3 (Immediate discipline, analytical models and research questions or hypotheses)
2.4 Conclusion
3 Methodology
3.1 Introduction
3.2 Justification for the paradigm and methodology
3.3 (Research procedures)
3.4 Ethical considerations
3.5 Conclusion
4 Analysis of data
4.1 Introduction
4.2 Subjects
4.3 (Patterns of data for each research question or hypothesis)
4.4 Conclusion
5 Conclusions and implications
5.1 Introduction
5.2 Conclusions about each research question or hypothesis
5.3 Conclusions about the research problem
5.4 Implications for theory
5.5 Implications for policy and practice
5.5.1 Private sector managers
5.5.2 Public sector policy analysts and managers
5.6 Limitations
5.7 Further research
Bibliography
Appendices
BASICS OF STRUCTURE AND STYLE
A five chapter structure can be used to effectively present a PhD thesis, and it is summarised in figure 1 and table 1.
Figure 1 Model of the chapters of a PhD thesisIn brief, the thesis should have a unified structure (Easterby-Smith et al. 1991). Firstly, chapter 1 introduces the core research problem and then `sets the scene' and outlines the path which the examiner will travel towards the thesis' conclusion. The research itself is described in chapters 2 to 5:
* the research problem and hypotheses arising from the body of knowledge developed during previous research (chapter 2),
* methods used in this research to collect data about the hypotheses (chapter 3),
* results of applying those methods in this research (chapter 4), and
* conclusions about the hypotheses and research problem based on the results of chapter 4, including their place in the body of knowledge outlined previously in chapter 2 (chapter 5).
This five chapter structure can be justified. Firstly, the structure is a unified and focussed one, and so addresses the major fault of postgraduate theses in a survey of 139 examiners' reports, that is, it clearly addresses those examiners' difficulty in discerning what was the `thesis' of the thesis? `Supervisors need to emphasise throughout students' candidacies that they are striving in the thesis to communicate one big idea' (Nightingale 1984, p. 174), and that one big idea is the research problem stated on page 1 or 2 of the thesis and explicitly solved in chapter 5. Easterby-Smith et al. (1991) also emphasise the importance of consistency in a PhD thesis, and Phillips and Pugh (1987, p. 38) confirm that a thesis must have a thesis or a `position'. Secondly, the structure carefully addresses each of the 31 requirements of an Australian PhD thesis outlined by an authority in a publication of the Higher Education Research and Development Society of Australia (Moses 1985, pp. 32-34). Thirdly, the structure is explicitly or implicitly followed by many writers of articles in prestigious academic journals such as The Academy of Management Journal and Strategic Management Journal (for example, Datta et al. 1992). Fourthly, the structure has been the basis of several PhD and masters theses at Australian universities that were completed in minimum time and passed by examiners with none or negligible revisions required. Fifthly, the structure is much like that which will be used by candidates later in their career, to apply for research grants (Krathwohl 1977; Poole 1993). Finally, by reducing time wasted on unnecessary tasks or on trying to demystify the PhD process, the five chapter structure provides a mechanism to shorten the time taken to complete a PhD, an aim becoming desired in many countries (Cude 1989).
Justified changes to the structure. Some changes to the five chapter structure could be justified. For example, a candidate may find it convenient to expand the number of chapters to six or seven because of unusual characteristics of the analysis in his or her research; for example, a PhD might consist of two stages: some qualitative research reported in chapters 3 and 4 of the thesis described below, which is then followed by some quantitative research to refine the initial findings reported in chapters 5 and 6; the chapter 5 described below would then become chapter 7. In addition, PhD theses at universities that allow its length to rise from a minimum length of about 50 000 to 60 000 words (Phillips & Pugh 1987), say, through a reasonable length of about 70 000 to 80 000 words, up to the upper limit 100 000 words specified by some established universities like the University of Queensland and Flinders University, may have extra chapters added to contain the extended reviews of bodies of knowledge in those huge theses. In brief, in some theses, the five chapters may become five sections with one or more chapters within each of them, but the principles of the structured approach should remain. That is, PhD research must remain an essentially creative exercise. Nevertheless, the five chapter structure provides a starting point for understanding what a PhD thesis should set out to achieve, and also provides a basis for communication between a candidate and his or her supervisor.
As noted above, the five chapter structure is primarily designed for PhDs in management or related fields using common methodologies. Qualitative methodologies such as case studies and action research (Perry & Zuber-Skerritt 1992; 1994) can fit into the structure, with details of how the case study or the action research project being presented in chapter 3 and case study details or the detailed report of the action research project being placed in appendices. In theses using the relatively qualitative methodologies of case studies or action research, the analysis of data in chapter 4 becomes a categorisation of data in the form of words, with information about each research question collected together with some preliminary reflection about the information. That is, the thesis still has five chapters in total, with chapter 4 having preliminary analysis of data and chapter 5 containing all the sections described below. Appendix III discusses in more detail the difficult task of incorporating an action research project into a PhD thesis.
Links between chapters. Each chapter described below should stand almost alone. Each chapter (except the first) should have an introductory section linking the chapter to the main idea of the previous chapter and outlining the aim and the organisation of the chapter. For example, the core ideas in an introduction to chapter 3 might be:
Chapter 2 identified several research questions; chapter 3 describes the methodology used to provide data to investigate them. An introduction to the methodology was provided in section 1.4 of chapter 1; this chapter aims to build on that introduction and to provide assurance that appropriate procedures were followed. The chapter is organised around four major topics: the study region, the sampling procedure, nominal group technique procedures, and data processing.
The introductory section of chapter 5 (that is, section 5.1) will be longer than those of other chapters, for it will summarise all earlier parts of the thesis prior to making conclusions about the research described in those earlier parts; that is, section 5.1 will repeat the research problem and the research questions/hypotheses. Each chapter should also have a concluding summary section which outlines major themes established in the chapter, without introducing new material. As a rough rule of thumb, the five chapters have these respective percentages of the thesis' words: 5, 30, 15, 25 and 25 percent.
Style
As well as the structure discussed above, examiners also assess matters of style (Hansford & Maxwell 1993). Within each of the chapters of the thesis, the spelling, styles and formats of Style Manual (Australian Government Publishing Service 1988) and of the Macquarie Dictionary should be followed scrupulously, so that the candidate uses consistent styles from the first draft and throughout the thesis for processes such as using bold type, underlining with italics, indenting quotations, single and double inverted commas, making references, spaces before and after side headings and lists, and gender conventions. Moreover, using the authoritative Style Manual provides a defensive shield against an examiner who may criticise the thesis from the viewpoint of his or her own idiosyncratic style. Some pages of Style Manual (Australian Government Publishing Service 1988) which are frequently used by PhD candidates are listed in appendix IV.
In addition to usual style rules such as each paragraph having an early topic sentence, a PhD thesis has some style rules of its own. For example, chapter 1 is usually written in the present tense with references to literature in the past tense; the rest of the thesis is written in the past tense as it concerns the research after it has been done, except for the findings in chapter 5 which are presented in the present tense. More precisely for chapters 2 and 3, schools of thought and procedural steps are written of in the present tense and published researchers and the candidate's own actions are written of in the past tense. For example: 'The eclectic school has [present] several strands. Smith (1990) reported [past] that...' and `The first step in content analysis is [present] to decide on categories. The researcher selected [past] ten documents...'
In addition, value judgements and words should not be used in the objective pursuit of truth that a thesis reports. For example, `it is unfortunate', `it is interesting', `it is believed', and `it is welcome' are inappropriate. Although first person words such as `I' and `my' are now acceptable in a PhD thesis (especially in chapter 3 of a thesis within the interpretive paradigm), their use should be controlled - the candidate is a mere private in an army pursuing truth and so should not overrate his or her importance until the PhD has been finally awarded. In other words, the candidate should always justify any decisions where his or her judgement was required (such as the number and type of industries surveyed and the number of points on a likert scale), acknowledging the strengths and weaknesses of the options considered and always relying upon as many references as possible to support the decision made. That is, authorities should be used to back up any claim of the researcher, if possible. If the examiner wanted to read opinions, he or she could read letters to the editor of a newspaper.
Moreover, few if any authorities in the field should be called `wrong', at the worst they might be called `misleading'; after all, one of these authorities might be an examiner and have spent a decade or more developing his or her positions and so frontal attacks on those positions are likely to be easily repulsed. Indeed, the candidate should try to agree with the supervisor on a panel of likely people from which the university will select the thesis examiner so that only appropriate people are chosen. After all, a greengrocer should not examine meat products and an academic with a strong positivist background is unlikely to be an appropriate examiner of a qualitative thesis, for example (Easterby-Smith et al. 1991), or an examiner that will require three research methods is not chosen for a straightforward thesis with one. That is, do not get involved in the cross fire of `religious wars' of some disciplines. Moreover, this early and open consideration of examiners allows the candidate to think about how his or her ideas will be perceived by likely individual examiners and so express the ideas in a satisfactory way, for example, explain a line of argument more fully or justify a position more completely. That is, the candidate must always be trying to communicate with the examiners in an easily-followed way.
This easily-followed communication can be achieved by using several principles. Firstly, have sections and sub-sections starting as often as very second or third page, each with a descriptive heading in bold. Secondly, start each section or sub-section with a phrase or sentence linking it with what has gone before, for example, a sentence might start with `Given the situation described in section 2.3.4' or `Turning from international issues to domestic concerns,...' Thirdly, briefly describe the argument or point to be made in the section at its beginning, for example, `Seven deficiencies in models in the literature will be identified'. Fourthly, make each step in the argument easy to identify with a key term in italics or the judicious use of `firstly', `secondly', or `moreover', `in addition', `in contrast' and so on. Finally, end each section with a summary, to establish what it has achieved; this summary sentence or paragraph could be flagged by usually beginning it with `In conclusion,..' or `In brief,...' In brief, following these five principles will make arguments easy to follow and so guide the examiner towards agreeing with a candidate's views.
Another PhD style rule is that the word `etc' is too imprecise to be used in a thesis. Furthermore, words such as `this', `these', `those' and `it' should not be left dangling - they should always refer to an object; for example, `This rule should be followed' is preferred to `This should be followed'. Some supervisors also suggest that brackets should rarely be used in a PhD thesis - if a comment is important enough to help answer the thesis' research problem, then it should be added in a straightforward way and not be hidden within brackets as a minor concern to distract the examiner away from the research problem.
As well, definite and indefinite articles should be avoided where possible, especially in headings; for example, `Supervision of doctoral candidates' is more taut and less presumptuous than `The supervision of doctoral candidates'. Paragraphs should be short; as a rule of thumb, three to four paragraphs should start on each page if my preferred line spacing of 1.5 and Times Roman 12 point font is used to provide adequate structure and complexity of thought on each page. (A line spacing of 2 and more paragraphs per page make a thesis appear disjointed and `flaky', and a sanserif font is not easy to read.) A final note of style is that margins should be those nominated by the university or those in Style Manual (Australian Government Publishing Service 1988).
The above comments about structure and style correctly imply that a PhD thesis with its readership of three examiners is different from a book which has a very wide readership (Derricourt 1992), and from shorter conference papers and journal articles which do not require the burden of proof and references to broader bodies of knowledge required in PhD theses. Candidates should be aware of these differences and could therefore consider concentrating on completing the thesis before adapting parts of it for other purposes. However, it must be admitted that presenting a paper at a conference in a candidature may lead to useful contacts with the `invisible college' (Rogers 1983, p. 57) of researchers in a field, and some candidates have found referees' comments on articles submitted for publication in journals during their candidacy, have improved the quality of their thesis' analysis (and publication has helped them get a job). Nevertheless, several supervisors suggest that it is preferable to concentrate on the special requirements of the thesis and adapt it for publication after the PhD has been awarded or while the candidate has temporary thesis `writer's block'.
The thesis will have to go through many drafts (Zuber-Skerritt & Knight 1986). The first draft will be started early in the candidature, be crafted after initial mindmapping and a tentative table of contents of a chapter and a section, through the `right', creative side of the brain and will emphasise basic ideas without much concern for detail or precise language. Supervisors and other candidates should be involved in the review of these drafts because research has shown that good researchers ` require the collaboration of others to make their projects work, to get them to completion' (Frost & Stablein 1992, p. 253), and that social isolation is the main reason for withdrawing from postgraduate study (Phillips & Conrad 1992). By the way, research has also shown that relying on just one supervisor can be dangerous (Conrad, Perry & Zuber-Skerritt 1992; Phillips & Conrad 1992).
Facilitating the creative first drafts of sections, the relatively visible and structured `process' of this paper's structure allows the candidate to be more creative and rigorous with the `content' of the thesis than he or she would otherwise be. After the first rough drafts, later drafts will be increasingly crafted through the `left', analytical side of the brain and emphasise fine tuning of arguments, justification of positions and further evidence gathering from other research literature.
Further reading: http://www.elec.uq.edu.au/doc/Thesis_guide/phdth1.html
Tuesday, January 13, 2009
P values. One-tail or two-tail?
Jan 13, 2009
When comparing two groups, you must distinguish between one- and two-tail P values. Some books refer to one-sided and two-sided P values, which mean the same thing.
What does one-sided mean?
It is easiest to understand the distinction in context. So let’s imagine that you are comparing the mean of two groups (with an unpaired t test). Both one- and two-tail P values are based on the same null hypothesis, that two populations really are the same and that an observed discrepancy between sample means is due to chance.
A two-tailed P value answers this question:
Assuming the null hypothesis is true, what is the chance that randomly selected samples would have means as far apart as (or further than) you observed in this experiment with either group having the larger mean?
To interpret a one-tail P value, you must predict which group will have the larger mean before collecting any data. The one-tail P value answers this question:
Assuming the null hypothesis is true, what is the chance that randomly selected samples would have means as far apart as (or further than) observed in this experiment with the specified group having the larger mean?
If the observed difference went in the direction predicted by the experimental hypothesis, the one-tailed P value is half the two-tailed P value (with most, but not quite all, statistical tests).
When is it appropriate to use a one-sided P value?
A one-tailed test is appropriate when previous data, physical limitations, or common sense tells you that the difference, if any, can only go in one direction. You should only choose a one-tail P value when both of the following are true.
* You predicted which group will have the larger mean (or proportion) before you collected any data.
* If the other group had ended up with the larger mean – even if it is quite a bit larger – you would have attributed that difference to chance and called the difference 'not statistically significant'.
Here is an example in which you might appropriately choose a one-tailed P value: You are testing whether a new antibiotic impairs renal function, as measured by serum creatinine. Many antibiotics poison kidney cells, resulting in reduced glomerular filtration and increased serum creatinine. As far as I know, no antibiotic is known to decrease serum creatinine, and it is hard to imagine a mechanism by which an antibiotic would increase the glomerular filtration rate. Before collecting any data, you can state that there are two possibilities: Either the drug will not change the mean serum creatinine of the population, or it will increase the mean serum creatinine in the population. You consider it impossible that the drug will truly decrease mean serum creatinine of the population and plan to attribute any observed decrease to random sampling. Accordingly, it makes sense to calculate a one-tailed P value. In this example, a two-tailed P value tests the null hypothesis that the drug does not alter the creatinine level; a one-tailed P value tests the null hypothesis that the drug does not increase the creatinine level.
The issue in choosing between one- and two-tailed P values is not whether or not you expect a difference to exist. If you already knew whether or not there was a difference, there is no reason to collect the data. Rather, the issue is whether the direction of a difference (if there is one) can only go one way. You should only use a one-tailed P value when you can state with certainty (and before collecting any data) that in the overall populations there either is no difference or there is a difference in a specified direction. If your data end up showing a difference in the “wrong” direction, you should be willing to attribute that difference to random sampling without even considering the notion that the measured difference might reflect a true difference in the overall populations. If a difference in the “wrong” direction would intrigue you (even a little), you should calculate a two-tailed P value.
Recommendation:
I recommend using only two-tailed P values for the following reasons:
* The relationship between P values and confidence intervals is more straightforward with two-tailed P values.
* Two-tailed P values are larger (more conservative). Since many experiments do not completely comply with all the assumptions on which the statistical calculations are based, many P values are smaller than they ought to be. Using the larger two-tailed P value partially corrects for this.
* Some tests compare three or more groups, which makes the concept of tails inappropriate (more precisely, the P value has more than two tails). A two-tailed P value is more consistent with P values reported by these tests.
* Choosing one-tailed P values can put you in awkward situations. If you decided to calculate a one-tailed P value, what would you do if you observed a large difference in the opposite direction to the experimental hypothesis? To be honest, you should state that the P value is large and you found “no significant difference.” But most people would find this hard. Instead, they’d be tempted to switch to a two-tailed P value, or stick with a one-tailed P value, but change the direction of the hypothesis. You avoid this temptation by choosing two-tailed P values in the first place.
When interpreting published P values, note whether they are calculated for one or two tails. If the author didn’t say, the result is somewhat ambiguous.
Source: http://www1.graphpad.com/faq/viewfaq.cfm?faq=1318
When comparing two groups, you must distinguish between one- and two-tail P values. Some books refer to one-sided and two-sided P values, which mean the same thing.
What does one-sided mean?
It is easiest to understand the distinction in context. So let’s imagine that you are comparing the mean of two groups (with an unpaired t test). Both one- and two-tail P values are based on the same null hypothesis, that two populations really are the same and that an observed discrepancy between sample means is due to chance.
A two-tailed P value answers this question:
Assuming the null hypothesis is true, what is the chance that randomly selected samples would have means as far apart as (or further than) you observed in this experiment with either group having the larger mean?
To interpret a one-tail P value, you must predict which group will have the larger mean before collecting any data. The one-tail P value answers this question:
Assuming the null hypothesis is true, what is the chance that randomly selected samples would have means as far apart as (or further than) observed in this experiment with the specified group having the larger mean?
If the observed difference went in the direction predicted by the experimental hypothesis, the one-tailed P value is half the two-tailed P value (with most, but not quite all, statistical tests).
When is it appropriate to use a one-sided P value?
A one-tailed test is appropriate when previous data, physical limitations, or common sense tells you that the difference, if any, can only go in one direction. You should only choose a one-tail P value when both of the following are true.
* You predicted which group will have the larger mean (or proportion) before you collected any data.
* If the other group had ended up with the larger mean – even if it is quite a bit larger – you would have attributed that difference to chance and called the difference 'not statistically significant'.
Here is an example in which you might appropriately choose a one-tailed P value: You are testing whether a new antibiotic impairs renal function, as measured by serum creatinine. Many antibiotics poison kidney cells, resulting in reduced glomerular filtration and increased serum creatinine. As far as I know, no antibiotic is known to decrease serum creatinine, and it is hard to imagine a mechanism by which an antibiotic would increase the glomerular filtration rate. Before collecting any data, you can state that there are two possibilities: Either the drug will not change the mean serum creatinine of the population, or it will increase the mean serum creatinine in the population. You consider it impossible that the drug will truly decrease mean serum creatinine of the population and plan to attribute any observed decrease to random sampling. Accordingly, it makes sense to calculate a one-tailed P value. In this example, a two-tailed P value tests the null hypothesis that the drug does not alter the creatinine level; a one-tailed P value tests the null hypothesis that the drug does not increase the creatinine level.
The issue in choosing between one- and two-tailed P values is not whether or not you expect a difference to exist. If you already knew whether or not there was a difference, there is no reason to collect the data. Rather, the issue is whether the direction of a difference (if there is one) can only go one way. You should only use a one-tailed P value when you can state with certainty (and before collecting any data) that in the overall populations there either is no difference or there is a difference in a specified direction. If your data end up showing a difference in the “wrong” direction, you should be willing to attribute that difference to random sampling without even considering the notion that the measured difference might reflect a true difference in the overall populations. If a difference in the “wrong” direction would intrigue you (even a little), you should calculate a two-tailed P value.
Recommendation:
I recommend using only two-tailed P values for the following reasons:
* The relationship between P values and confidence intervals is more straightforward with two-tailed P values.
* Two-tailed P values are larger (more conservative). Since many experiments do not completely comply with all the assumptions on which the statistical calculations are based, many P values are smaller than they ought to be. Using the larger two-tailed P value partially corrects for this.
* Some tests compare three or more groups, which makes the concept of tails inappropriate (more precisely, the P value has more than two tails). A two-tailed P value is more consistent with P values reported by these tests.
* Choosing one-tailed P values can put you in awkward situations. If you decided to calculate a one-tailed P value, what would you do if you observed a large difference in the opposite direction to the experimental hypothesis? To be honest, you should state that the P value is large and you found “no significant difference.” But most people would find this hard. Instead, they’d be tempted to switch to a two-tailed P value, or stick with a one-tailed P value, but change the direction of the hypothesis. You avoid this temptation by choosing two-tailed P values in the first place.
When interpreting published P values, note whether they are calculated for one or two tails. If the author didn’t say, the result is somewhat ambiguous.
Source: http://www1.graphpad.com/faq/viewfaq.cfm?faq=1318
Subscribe to:
Posts (Atom)
