Wednesday, May 05, 2010

Multipaper Dissertations - Considering an Alternative to Traditional Dissertations

I'm almost finished my first year of my PhD studies. I still haven't narrowed down my topic. I have concluded the format of my dissertation. I was considering the multipaper format that appears to be popular in computer science. I don't think it is the right option for me (or for my faculty).

In trying to decide if this alternative format would be appropriate and helpful, I found that there wasn't much written about this format. So here is background info I put together, which may help others in determining if this format is right for them.

What is a multipaper dissertation?
A traditional dissertation often relates the details of one large study (which may have various sub-components) and is normally comprised of an introduction, literature review, methodology, findings/results, discussion, and conclusion. An alternative is to publish a small number of papers in traditional journals and include them with a contextualization chapter or two to form a dissertation.

I encountered various names for this type of dissertation (some of which have shades of difference), such as multipaper, sandwich, alternative, stapled-papers, Scandinavian, paper-based, PhD by publication(s), thesis by publication, and PhD by published works. I prefer the term multipaper dissertation as it is suitably descriptive of its defining quality.

While not attempting to relate a definitive history of dissertations, a brief background is helpful to position the novel format. Our notion of the traditional doctoral degree and accompanying dissertation format was largely imported from Germany to Canada and the United States in the mid-nineteen century. This format has not changed significantly since that time. It appears the multipaper format was introduced into the United Kingdom and the United Statues in the 1960s and has grown slowly but progressively since then. I was unable to find any accounts of the format’s usage in Canada, but it is not uncommon to find it used in Scandinavia or Australia. In the UK and US, it is becoming increasingly used in the fields of chemistry, physics, biology, geology, computer science, finance, economics, marketing, and medicine

Components of a multipaper dissertation
Badley (2009) analyzed universities in the United Kingdom offering this format and the academic literature surrounding the topic; he found few commonalities and a lack of scholarship on the topic. Based on the literature I did find, there appears to be little consensus on a standard structure for a multipaper dissertation. According to Badley’s research, some UK universities do not require any documentation beyond the publications, while others may require a new document, a “critical appraisal”, be written to synthesize the collected works (p. 333, 2009).

Grant and Reed offer their take on what a multipaper dissertation should contain: an abstract, an introduction, an explanation or summary of the included papers, the published papers, a conclusion, and a literature review as an appendix (2006). Including the full contents of the published papers in the master document is not necessarily an obligatory component (although discussing them, no doubt, is). Multipaper dissertations do tend be shorter than the traditional format as it is felt that the published work alone is indicative of significant scholarship. Another variance is in the number of papers to include. Badley found no universities were explicit on the required number, but he offered that best practice indicates it should be not less than six and no more than ten.

Pros and cons of the multipaper format
The multipaper format paper has engendered criticism and disrepute arising largely from universities that allow students to submit any work published prior to enrollment. In the United Kingdom, the "PhD by publication" process was initially used to allow existing university faculty (for example, heads of laboratories, librarians, scientists) who had a distinguished record of publishing in academic journals to acquire a doctorate more easily. This practice remains despite criticisms of it being an insider perk. Further criticisms of this format arose upon some universities opening this option to any qualifying student (provoking derogatory remarks that anyone who pays tuition, qualifies).

Some universities also allow any pre-published or artistic work, including those neither peer-reviewed nor from an academic source. Starrs (2008) argues that evaluating the films of a director for a PhD in film studies makes sense, but allowing trade literature for any PhD program may not. Badley is further critical that accepting trade literature is more appropriate for professional doctorates than a PhD. It is this lack of traditional, advanced academic performance metrics that spurs criticisms that doctorates attained this way amount to degrees for purchase. Starrs equates degrees achieved in such as fashion as essentially honourary degrees, at times as dubious as the one the Long Island University awarded to Kermit the Frog. It is therefore important to distinguish between a dissertation comprised of peer-reviewed papers published in academic sources during one’s enrollment and a dissertation comprised of papers or work published before starting a PhD program.

The traditional dissertation format, however, would not have survived for such a long time if it did not offer strong advantages. The familiar format provides a standard template and shared experience for all doctoral students. This makes the doctoral process more consistent and understandable. It also provides training in delivering an in-depth piece of original research. This training is a singular opportunity in most academics’ careers that allows them to acquire the experience and evidence of mastering a topic and methods.

However, Duke and Beck believe traditional formats do not offer sufficient practice in writing the journal or trade articles that students will be expected to master upon graduation (1999). They posit that a traditional dissertation is a unique genre in both style and in readership. Dissertations are thus like a will; one only writes one once. Mastering this genre is consequently of minimal value. They also argue that the style and circumstances of a traditional dissertation are such that few are suitable for publishing or apt to be reworked into more than a couple publishable papers at best. Very few dissertations thus result in any publications. Duke and Beck therefore advocate for the multipaper format as it offers the opportunity for more real-life practice and increased readership.

The following additional benefits can be listed: it can accelerate a career by improving one’s curriculum vitae and connections with journal editors; it provokes comments from reviewers that can be invaluable and may come from scholars more knowledgeable on one’s specific domain than may otherwise be available; it encourages collaboration; it mitigates against poor results or arriving at a project failure too late in one’s dissertation process by focusing on smaller projects and getting progressive feedback through the publication process.

Multipaper dissertations have disadvantages: the process is slower due to publication timeframes, and if full publication is required this can amount to a substantial delay; one gets experience with several smaller projects but not the benefit of working on one big, in-depth project; peer review may be tougher and less open to discussion than one’s committee; and there may be a lack of experience at one’s university with this format that may provoke resistance or result in insufficient support.

To get around the time delays that academic publishing often entails, Duke and Beck suggest that a multipaper dissertation be allowed to include papers that are ready for consideration or in the process of consideration for publication (1999). Another way to combat publishing and feedback delays is to publish in conference proceedings. Further consideration should also be made to account for an individual student’s personality, circumstances, faculty, and topic that may – or may not – be conducive to such an approach.

Evaluation criteria
This nonstandard dissertation format introduces the problem of finding an evaluation criteria. Traditional quality metrics may not apply to such a format with profound differences in structure and goals.
The traditional format appears to have a general consensus on quality hallmarks. Garson is representative of the standard advice, stating that dissertations should start with framing an analytic question and offering a roadmap; have a robust, well-organized literature review; develop a model through operationalizing variables and a rigourous research design; present the case through clear writing and valid findings; and conclude by drawing inferences, making generalizations, and pointing ahead. More specialized advice for qualitative dissertations is offered by Piantanida and Garman (1999). They outline high-level qualities for optimal qualitative work, which are integrity (structural soundness), verité (rings true), rigour, utility, vitality, aesthetics, and ethics.

Unlike with the traditional format, there appears to be little guidance for multipaper formats; I was able to find only one source for any discussion of evaluation. Badley, in his overview of the status of the format in the UK does offer some quality benchmarks (2009). He believes such dissertations should strive for originality, rigour, significance, and coherence.

Badley acknowledges that these criteria are not necessarily unique, objective, or well-defined. He even questions whether a discussion of quality is necessary since the articles have already been peer-reviewed and accepted by journals. He notes that the multipaper format is evaluated more on process rather than the product.

Ascertaining a work’s originality within an entire subject is hard to assess for any work. With multipaper dissertations traditional ways to assess originality also applies, but a unique issue arises when the associated papers are jointly authored. This raises the question of individual contribution, which is rarely an issue for dissertations otherwise. In a traditional dissertation the work is assumed to be new work originating from the PhD student. With multipaper dissertations the master dissertation document would always be authored by an individual PhD student, but it is not uncommon for the associated publications to be comprised of jointly authored papers. This raises the question of how much work is done by the individual. Robins & Kanowski (2008) find this issue to be one of the main contentions with this format at many universities. They advise that universities should require a declaration from the PhD student regarding their level of contribution. They do not, however, advise how to determine sufficient individual effort. Universities accepting this format, it would seem, must be comfortable with collaborative rather than entirely individual effort.

Again traditional evaluation criteria (e.g. appropriate sampling strategy, validity of findings, etc.) do also apply. Where the unique aspects of the multipaper format arises in regards to this criteria is in the length of reporting on methodology. Although methodology sections can be unsatisfactorily brief in traditional formats, in multipaper dissertations they are almost brief by definition. To further compound the problem are articles published in conferences proceedings as such articles are even shorter than conventional journals. This further reduces the opportunity to supply sufficient methodology details. Wilson encountered this situation when his university began accepting the format (2002). This posed a roadblock to committee members feeling they were able to adequately evaluate candidates' work. Wilson notes that this was rectified by requiring candidates to augment their dissertations with further methodology details.

As multipaper format dissertations are comprised of existing published works, evaluating significance can be done by examining the papers publishing records and feedback. This can be assessed based on impact factors, download totals, citation counts, prestige of publication source, reviews, and informal feedback. This data is useful to demonstrate that the work had importance and influence in the field – and also indicates that a dissertation was read by more than just a committee and a student’s spouse.

Traditional dissertations can suffer from an inability to maintain a clear focus, but this problem is exacerbated when a dissertation is also comprised of multiple, possibly external, papers. To further compound the problem, the research could have been conducted through a series of separate projects conducted years apart. The goal of a dissertation is to demonstrate mastery of conducting research on a single topic, not to patch together a collection of assorted work that one has already conveniently published. A dissertation can explore various dimensions of a subject, but at some point trying to marry divergent works into one document becomes difficult. Cinderella’s wicked stepsister managed to get most of her foot in the glass slipper, but it clearly did not fit and it wasn’t pretty. Thus achieving coherence in a multipaper dissertation is an important, if difficult, undertaking.

PhD as process, not product
I heard a saying that the only good dissertation is a completed one. If dissertations are a journey then some see the destination as less important than how one got there. Badley notes that shares the journey analogy as it “should help candidates shift attention away from the traditional PhD emphasis on the research product – the thesis itself – towards the development of the autonomous scholar who is capable of undertaking further research journeys” (2009, p. 340).

Samples of multipaper format
The two dissertations below offer two different styles using this format. Harper uses the more conventional (of this unconventional format) approach of including his published papers in his dissertation with a framing introduction and conclusion. Roto, on the other hand, lists her published papers and then synthesizes from them to offer an entirely new paper in a conventional format.

Harper, F. (2009). The impact of social design on user contributions to online communities. University of Minnesota, Department of Computer Science and Engineering.

Roto, V. (2006). Web browsing on mobile phones – characteristics of user experience. Helsinki University of Technology, Department of Computer Science and Engineering.

Badley, G. (2009). Publish and be doctor-rated: The PhD by published work. Quality Assurance in Education, 17(4), 331-342.

Duke, N. K., & Beck, S. W. (1999). Education should consider alternative formats for the dissertation. Educational Researcher, 28(3), 31-36.

Garson, G. D. (2001). Guide to writing empirical papers, theses, and dissertations. New York: M. Dekker.

Grant, D., & Reed, A. (2006, April 21). Multi-paper Dissertation. Retrieved April 23, 2010 from multi- paper%2520dissertation%2520presentation%2520042106. ppt+multipaper+dissertation&cd=1&hl=en&ct=clnk&gl=ca

Piantanida, M., & Garman, D. N. (1999). The qualitative dissertation: A guide for students and faculty. Thousand Oaks, CA: Corwin Press.

Robins, L., & Kanowski, P. (2008). PhD by publication: A student's perspective. Journal of Research Practice, 4(3).

Starrs, B. (2008). Publish and graduate?: Earning a PhD by published papers in Australia. M/C Journal, 11(4).

Wilson, K. (2002). Quality assurance issues for a PhD by published work: A case study. Quality Assurance in Education, 10(2), 71-78.


Max said...

Hi webslinger.

I just discovered that you wrote a lengthy paper with a detailed critique of my dissertation, which is unexpected.

FWIW, my assumption is (was?) that conference papers have impact, while nobody will read the dissertation outside of my committee. Also, I believe that dissertations are tied to institutional culture more than personal preferences or a desire for influence over the research community. It strikes me as backwards to begin thinking about the dissertation format first, but perhaps this is your institutional culture.

Good luck with your journey.


Glen Farrelly said...

Hi Max:

My paper was an assignment for a class, hence the structure of the paper. I took the paper down as my analysis of the dissertations isn't important. The consideration of the format, I think, is (which is why I summarized it for this blog post).

I haven't encountered sufficient consideration of this format either in literature (almost nothing published on it that I could turn up) or in academic circles. I believe it can have tremendous impact on doctoral students' careers and the entire nature of a PhD.

There are questions of fairness and consistency across regions and universities. That's a larger educational issues, which I didn't really examine. I looked at the issues from a personal perspective.

I completely agree with you that institutional culture and one's field of study are the deciding factors. Multipaper dissertations are standard in some regions and discliplines - in which case following those considerations is a primary concern. Also, some faculties are not supportive (or are against) the format, in which case fighting to have a different format recognized added to the already stressful and difficult process of completely a PhD is probably unwise.

If one has a choice of formats, however, then I think the format is the primary concern as it is linked to one's research agenda and career. Considering the format helps one determine which path would be best suited to them.

Considering the format, then will help one consider the importance of graduating with a few publications (which is more crucial in some fields than others), whether they want experience with many projects or one big one, whether their research goals and questions are conducive to multiple projects or one, and whether they want to have more people read their work than just their committee and spouse.

These are pretty foundational considerations for new scholars so I think considering this format is a great way to assess one's goals for their PhD and subsequent career.

Ultimately there are practical constraints. So it may be great to actually want one's work to be read, but then it is even more great to do what is necessary to finish the program (considering the stat I heard that less than approximately less than 50% of people who start actually finish a PhD).