An Agile Approach to
Doctoral Research and
Dissertations
Doctor of Professional Studies in Computing Class of 2016
Seidenberg School of Computer Science and Information Systems
Pace University, White Plains, NY
Student/Faculty Research Day, May 2, 2014
This research sets out to:
• compare traditional, or “non-agile” doctoral programs and their
respective dissertation processes with an Agile professional studies
doctoral program.
• compare and contrast the core differences of the two program types.
• summarize enrollment and completion rates from various private,
public, and online institutions
• conclude with a summary and ideas for future work.
2
Why is Doctoral Research Important?
• Although there are inherent differences, traditionally, university-led research programs devote a percentage
of resources to the attainment of a doctoral-level degree, which affords both student and university with
what can be considered leading-edge research. This research thus becomes a vital and continuous source of
funding for universities due to the reputation gained of discovering, sharing, and creating real-world
solutions (measured via various metrics such as student enrollment, graduation and retention rates, reputation,
endowment, number of patents, et al)
What are the drivers of value?
• Given the breadth, depth, and successes of ongoing doctoral-level research programs, interest has arisen in
discovering a better understanding as to which components in these respective research programs are
responsible for the tremendous value that continues to be attributed to university-led doctoral research.
Do results vary based on non-agile vs. Agile research and dissertation processes?
• In this study, we examined the historical make-up of non-agile doctoral-level research programs with a
focus on U.S. universities that offer degrees in information technology, information systems, computer science
and engineering, and computational and computing disciplines such as informatics and information theory.
• The common framework and practices in these programs were discussed and analyzed. The researchers then
compared and contrasted these non-agile programs which employ a top-down, approach of directed
research from theory to applications, with those programs that adopt an Agile-based approach.
3
An Agile approach to doctoral research and dissertations can be compared to Agile values in
software development
Using Agile values can allow students to:
 Complete their dissertations
 Complete their dissertations sooner
 Maintain a sustainable pace during research and dissertation writing
The important Agile principles for a research and dissertation process include:
 Early and continuous delivery of valuable product
 Welcoming changed requirements, even late in development
 Delivering working product frequently, from a couple of weeks to a couple of months, with a preference
for the shorter timescale
 Measuring progress primarily through useful deliverables
 Using Agile processes for sustainable development, enabling sponsors, developers, and users to
maintain a constant pace indefinitely
 Valuing simplicity as the art of maximizing the amount of work not done.
4
• The DPS 2016 cohort comprised of 14 doctoral students authored and
conducted the study.
• The authors operated on the Agile practice of “frequent delivery of quality
product” by setting up a framework for the study paper and matrices, which
were filled in as drafts of sections were completed.
• Authors made phone calls and sent emails to Universities and searched websites
to find program descriptions and statistics for Computer Science and
Information Technology doctoral programs.
5
Research was focused on answering these questions for public, private, and online doctoral programs:
• The number of applicants
• The number of acceptances
• The number of students who have passed the written qualifying exam
• The number of students who have passed the oral qualifying exam
• The number of graduates
• The length of time that it took these students to complete their studies
• The number of students leaving the program ABD (All But Dissertation)
• If this could be broken down by gender, and race
• And if possible, by year
The authors built matrices and statistics tables as results were provided, and all
authors were asked for their input on the conclusions.
6
Private
• Indiana University – School of Informatics and Computing
• Stanford University – Engineering Computer Science
• Pace University – Doctor of Professional Studies in Computing
Public
• Penn State University – College of Information Science and Technology
• University of Virginia – Computer Science and Engineering
• University of California, Los Angeles – Computer Science
• Virginia Tech – Computer Science and Information Technology
• City College of New York - Engineering
• George Mason University – Information Technology and Computational Sciences
• Rutgers University – Computer and Information Science
Online
• University of Phoenix – not specified
• Nova Southeastern – not specified
• Walden University – not specified
• Capella University – Information Technology
• Edinburgh University – Science and Engineering
7
Characteristics
Type of Student
Non-agile Research Program
Generally young Masters degree graduates
with little to no experience.
Agile Research Program
Seasoned professionals. (Agile teams hire those
with strong skillsets.)
Lifestyle of
Student
Usually residential, full time, and inflexible
schedules. Students are usually not
permitted to have full time jobs outside of
program.
Tends to be isolated. Students work alone
on their research with little community
support.
Students are fully employed in the area of their
studies in order to develop professionally.
Dissertation
Methodology
Dissertations must be proposed in a large
formal planning process and drafts are
reviewed only occasionally.
Dissertations are designed briefly and evolve over
the lifetime of the document. Regular check-ins
(frequent iterations) ensure the product is of high
quality without waste.
Time to
Completion
Format
Teaching
Experience
Examinations
5+ years
~3 years
Individual, in person
Most students teach as part of their
program.
Comprehensive examinations
Cohort, hybrid
No teaching required.
Community
Support
Cohort-based. Students support each other and
the community’s goals. Students work together
on teams (pairing) to accomplish milestones in
the program.
No comprehensive examinations.
8
Process
Non-agile Approach
Dissertation development Mostly individual with guidance from an
dissertation advisor
Qualifying Exams
Must pass all qualifying exams in order to
qualify to continue with your dissertation
Internal Review Boards
Required for all dissertations conducting
human research
Course Work completion All course work must be completed prior
to submission of dissertation proposal
Document name
Document length
Document scope
(Research Approach and
Plan)
Agile Approach
DPS candidate can work with both an advisor,
input from other DPS students and teachers and
colleagues from work.
No qualifying exams
Required for all dissertations conducting human
research
DPS is now requiring Idea Paper at the end of the
first year of program. There is at least another
year of classes.
Dissertation proposal
Idea Paper
Dissertation proposal average length is 30- Idea paper is 7-10 pages
50 pages long
Non-agile approach dissertation proposal Idea paper is created using Agile methodology to
is created to cover the entire proposal in an deliver a working idea paper developed
all or nothing approach. The full thesis
frequently and at a sustainable pace. The idea
proposal is required by doctoral committee paper is continually honed and focused during
the second year. The focus of the idea paper is
presented and reviewed in an iterative style in
class.
9
Process
Dissertation idea
(Research Area)
Average Time to
completion
Oral defense /
dissertation defense
Literature Review
Dissertation Status
Sessions
Dissertation Manuscript
Non-agile Approach
Can be chosen by advisor
Agile Approach
Can be an idea that you are interested in.
Average student takes 8.2 years to
complete
Required
Average Agile Pace DPS is 4.05 years
Student reviews literature relevant to the
problem.
Meet regularly with your advisor to discuss
your research idea and progress.
No recommended amount of meetings
Student reviews literature relevant to the
problem.
Regularly scheduled meetings, usually with
PowerPoint presentations, Includes:

Brief Elevator description of your research
problem

What you did since the last meeting.

What you are currently doing.

What you intend to do by the next session.
Idea paper is expanded to a full document that
will become the dissertation. Student and advisor
decide where the gaps are and their resolutions.
Completed thesis is given to advisor for
final approval, prior to the oral defense.
Required
10
The typical research process in a non-agile
approach consists of the following steps:
1. Planning the Study
2. Literature Review
3. Study Implementation and Data
Gathering
4. Analysis and Interpretation
5. Reporting
The typical research process in an Agile
approach that consists of the following
steps:
1. Begin with the problem
2. Research defines the goals
3. Divide into sub-problems
4. Create Hypothesis as proposed
solutions to the problem
5. Look for data directed by your
hypothesis, collect and organize
6. Interpret the meaning of the data,
resolve the problem, or create new
ones
7. Start back at [1]
11
• Data shows a higher level of
completion for Pace DPS
students in the first three to
five years of study, while
some of the other programs do
not show their students
completing until the six to eight
year mark.
• While Pace DPS students tend
to complete their degree
earlier as compared to
traditional programs, the Pace
faculty are always trying new
methods to improve the
completion rate and the
dissertation quality
12
• We found that because the Pace DPS program leverages professional
experience in the program, this enables quicker completion whereas nonagile programs spend anywhere from three to five years preparing students with
basic coursework before they even begin the dissertation process.
• When speaking with several of the traditional institutions, we also found that
they seek and attract a much different type of doctoral candidate. Most
institutions would not allow students to have careers or full time jobs,
required residential status, and required students to teach during their
program.
• Moreover, the dissertation process was mostly self-guided, instead of a true
collaborative work as it is in the Agile research and dissertation process.
13
• Survey tool sent to all DPS students (alumni and current) asking them about
their satisfaction with the program, if they would suggest the program to
other perspective students, and suggestions for possible improvement of the
DPS program.
• Additional research around graduation and attrition statistics. Due to the lack of
response from many of the researched institutions, it was difficult to obtain a
consistent set of statistics across the board from public, private, and online
institutions.
14
•
•
•
•
•
•
•
Gilbert Alipui
Claude Asamoah
Richard Barilla
Leigh Anne Clevenger
Alecia Copeland
Sam Elnagdy
Hugh Eng
•
•
•
•
•
•
•
Michael Holmes
Saravanan Jayaraman
Kevin Khan
Steven Lindo
Javid Maghsoudi
Mantie Reid
Michael Salé
15
•
[1] Agile Manifest Site; http://agilemanifesto.org/ , accessed April 2014.
•
[2] X. Chen (2009). Students Who Study Science, Technology, Engineering,
and Mathematics (STEM) in Postsecondary Education. Stats in Brief. NCES
2009-161. National Center for Education Statistics.
•
[3] Computing Research Association Taulbee Survey;
http://cra.org/resources/taulbee/ , accessed April 2014.
•
[4] Graduate Division Program Profile Report: Computer Science Dept.,
UCLA;
http://www.gdnet.ucla.edu/asis/progprofile/result.asp?selectmajor=0201 ,
accessed April 2014.
•
•
•
•
•
•
[5] F. Grossman, C. Tappert, J. Bergin & S. M. Merritt (2011). A research
doctorate for computing professionals. Communications of the ACM, 54(4),
133-141.
• [10] K. Kuldeep, R. Welke & R. Weber (2007). Restoring the Viability of PhD
Programs in Information Systems: Getting Past Denial and Targeting Non-Nonagile Markets. ICIS 2007 Proceedings.
• [11] E. Mansfield & J. Y. Lee (1996). The modern university: contributor to
industrial innovation and recipient of industrial R&D support. Research policy,
25(7), 1047-1058.
• [12] L. McAlpine & J. Norton (2006). Reframing our approach to doctoral
programs: an integrative framework for action and research. Higher Education
Research & Development, 25(1), 3-17.
• [13] E. L. McWilliam, P. Taylor, P. Thomson, B. Green, T. Maxwell, H. Wildy &
D. Simons (2002). Research Training in Doctoral Programs-What can be
learned from Professional Doctorates? Commonwealth of Australia.
• [14] S. M. Merritt, J. Bergin, H. Blum, R. Frank, D. A. Sachs, A. Stix & S. Varden
(2001). The Doctor of Professional Studies in Computing: An Innovative
Professional Doctoral Program. Life Sciences, 9(6.4), 8-7.
[6] F. Grossman, C. Tappert, J. Bergin, S.M. Merritt, “A Research Doctorate
for Computing Professionals”, in Communications of the ACM, April 2011, Vol.
54, No. 4, pp.133-141.
• [15] S. M. Merritt, A. Stix, J. E. Sullivan, F. Grossman, C. Tappert & D. A. Sachs
(2004, June). Developing a professional doctorate in computing: a fifth-year
[7] F. Grossman, DPS Dissertation Completion Rate, Pace University DPS in
assessment. In ACM SIGCSE Bulletin (Vol. 36, No. 4, pp. 42-46). ACM.
Computing Program, 2014
• [16] Council of Graduate Schools PhD Completion Project;
[8] H. Klein & F. Rowe (2007). Marshalling the professional experience of
http://www.phdcompletion.org , accessed April 2014.
doctoral students: Towards bridging the gaps between theory and practice.
• [17] J. Swazey, M. Anderson & K. Louis (1993). Ethical Problems in Academic
[9] D. Kohrell, “Agile Principle 10 – Maximize work NOT done!” ;
Research; A survey of doctoral candidates and faculty raises important
http://tapuniversity.com/2011/02/13/agile-principle-10-maximize-workquestions about the ethical environment of graduate education and research.
not-done/, Feb. 13, 2011, accessed April 2014.
American Scientist, 81(6), 542.
16