Tag Archives: Dissertation

Amazon’s Kindling 2 for Poor Research Habits in the New Generation of Young Researchers

Posted on by
This entry is part 1 of 2 in the series Scholarly Research

When I began work on my honours thesis in receptor-ligand kinetics during my biology and biotechnology degree, in 1998, electronic journals were not on the minds of legitimate mainstream research scientists.  At the time, Carleton University introduced an extremely-beneficial agreement it had reached with the National Research Council to allow fourth year honours students working on their theses access to the NRC CISTI Library Stacks–as a government agency, the NRC CISTI library did provide services to the public, but the actual stacks were out of bounds, public patrons could only access the extensive resources in their East Ottawa library

004

facilities by submitting reference lists on paper and waiting for library staff to retrieve them.  As fourth year honours thesis students, we benefitted from the more intimate time we got to spend with the references, and the ability to follow the chain of research articles much faster.  Remember, this was all during a time when there was no way to conduct research comprehensively in one pass;

you simply had to look up a set of initial promising-looking references, find only about 75% of them are in the library collection to begin with, and a further 25% are misplaced, out for binding, or waiting to be returned from the photocopy room.  And even then, you had very little idea whether those articles were of any value at all until you actually read them.  Only then, could you begin to make a second generation of references to look for, based on that first pass.

Normal research papers would take several passes of this process, and although the library was open every night until 23h (Walden University’s Librarians still provide personalized research reference services by telephone and email until 1am!).  This is one aspect of the rigorous research process that consumed the lion’s share of time it took to conduct research, and emphasized in us as developing researchers, the importance of the iterative critical evaluation process.

The young students nowadays probably can’t even conceive of what tremendous effort this research process necessarily was, and that you couldn’t rush journals to come back from the bindery or being located misplaced on shelves.  Today’s generation of young students has not only the belief, but the expectation, that an entire legitimately-researched and supported scholarly paper can be–indeed, many of them even believe that it should be–researched defensibly in one pass.  Although databases like ProQuest, LexisNexis, and ScienceDirect do indeed facilitate and vastly improve research productivity.  But the new generation of computer-based journal databases and their reliance on keyword searches and abstracts undermines the significantly-more valid critical process of evaluating references for their actual findings and content before dismissing a reference, considering it further, or determining what articles to follow next in the chain.  I encourage Interested students to visit the following resources to on literature surveys and research skills:

Research Resources

 
 

Dear Robin Cheung,

Thank you for taking part in our eBook survey at the Springer booth at Academy of Management 2010 in Montreal.kindle 2

Your feedback was a valuable source in getting more information about the experiences with eBooks our readers already have. I am happy to announce that you are the winner of the Kindle 2. Congratulations!

Please inform us about your shipping address so we can send it to you.

You then will receive the Kindle within the next few weeks by mail. Enjoy reading Springer eBbooks and eJournals! More than 20.000 Springer eBooks are available in Kindle format at Amazon. Please make also sure you will have access to our eProducts through your library. You and your librarian can find more information here: http://www.springer.com/librarians/e-content/ebooks?SGWID=0-40791-0-0-0

Best regards,

Margit Dann, Springer Product Manager

Sony E-ReaderAfter I had tried, for a few months, to use Sony’s PRS-505 eBook reader, the frustration of being unable to highlight research or add marginalia meant I was still printing out research and annotating by hand; the eBook reader’s ability to bring around myriad research papers to read whilst out of the home or office is entirely eclipsed by the inability to highlight or even recall later which reference had the interesting bit of information you wanted to remember.

But the Amazon Kindle 2, which has wireless support here in Canada, does support highlighting, though by a slightly-annoying five-direction “joystick” that you can really only reliably manipulate by reoriented the edge of a fingernail to catch on. And it does have a keyboard, though the linear layout is quite uncomfortable to type much more than a few words, unlike the curved keyboard of the original 1990s Blackberries.

All that remains now is for Mendeley to improve the core paradigm that it based its schema on, and someone to leverage the open API access they recently launched and write a value-added application to let us accomplish more than just reading and annotating on Kindle, but then due to the Kindle’s non-support of native PDF, manually re-annotating on Mendeley, where we might eventually be managing both the PDF and the citation. And for journal publishers, like Springer, to facilitate adoption of these devices into the research workflow; they allow much more ready-access, portability, and confidence in citations–everything that is research.

The Kindle provides a way for established researchers to improve portability, availability, and accountability by facilitating the migration to electronic journals.  In my bachelors honours thesis days, I could legitimately get away with photocopying all the journal articles that supported my research (the rule of thumb most professors expected was approximately one scholarly reference cited directly for each  page your paper was in length), since I might have one or two 30-page (article-length) papers for a whole term, and only one thesis around 100 pages in length.  I could easily handle annotating in the margins by hand, putting a phrase or two in big letters on the first page of each so I could find each reference at a glance, and filed in file boxes.

When I began working on my PhD, I began to find that my old system was too inefficient. It was simply taking too much time and to try to locate a single reference that I needed to support a single fact I needed to assert.  Even worse: when I tried to find a reference from which to synthesize and deduce a specific application of theory, requiring more than one paper to support and qualify restrictions on applicability, it become virtually impossible to locate all the references in a reasonable amount of time.  And managing the mountain of file-boxes in my spartan accommodations was more than my meagre Toronto basement rental could accommodate.

(I think people have a bit of a misconception of what a doctorate is intended to do and what it involves.  Writing doctoral-level papers does not simply mean more complex; in fact, complexity is not a requirement for doctoral work–it’s more a symptom.  Doctoral work must be appropriately-specific so that a reasonable-length 200-page dissertation can be authoritative and comprehensive, but stay within scope. Doctoral work, in start opposition to my MBA work, requires not only every assertion you make to be supported with reliable references, but that you demonstrate that your assertion is supported by all relevant accepted theory. So when people believe that doctoral level work is simply the same as bachelors, undergraduate work, that is simply not the case.  Whilst your Bachelors degree aims to develop higher order reasoning with a certain known introduction to a body of knowledge, to earn a doctorate, candidates must develop rigorous research skills, develop rigorous scholarly reasoning to ensure research is maximally-valid, and apply research to produce a dissertation that must demonstrate a significant, scientifically-valid contribution to humanity’s body of knowledge that was not shown before.)

.

100_1666 (800x600) North Park and Residency Checkin 7Jan2010 026 (800x217) Walden Splash at Sheraton Dallas Entrance Day 3 Lunch, Validity, Junk English 037 (800x600) Walden DFW Residency Day 3 027 (800x600) Walden DFW Residency Day 3 030 (800x600) 100_1662 (800x600) Walden DFW Residency Day 3 015 (800x600) North Park and Residency Checkin 7Jan2010 038 (600x800)

Publisher Interested in my Dissertation as a Book

Posted on by

Just 20 minutes ago, I was still dragging my feet around, trying to do anything but get my day under way.   But then an email came:

Dear Robin Cheung,

I am writing on behalf of an international publishing house, Lambert Academic Publishing.

In the course of a research on the Walden University, I came across a reference to your thesis on “The effects of Basel II and OSFI regulations that saved Canadian Banking: Key learnings for other banks and regulators.”.
We are an international publisher whose aim is to make academic research available to a wider audience.
LAP would be especially interested in publishing your dissertation in the form of a printed book.

Your reply including an e-mail address to which I can send an e-mail with further information in an attachment
will be greatly appreciated.

I am looking forward to hearing from you.
Kind regards,
Daniela M.
Acquisition Editor

LAP LAMBERT Academic Publishing AG & Co. KG
Saarbrücken
Dudweiler Landstraße 99, 66123 Saarbrücken Germany

When I came up the the topic idea–and I’m sure many other people also did–but I probably had a better grounding in the subject matter than the others who thought of it.  When I was a Financial Analyst at the Canadian Imperial Bank of Commerce, I had a chance to pursue a very entrepreneurial attitude, looking for interesting dotted line-reporting projects.  One of them  I completed was making a “mini-HRMS” system for the recruiters in CIBC HR that hired me.  It was in mid-2002 when the  Director of Investor Relations approached me to design a comparative analysis database and series of reports.  I was also allowed to research what performance and stability metrics were most appropriate, investigate who the owners of those data were, and populate the model.

I do so hope to hear back from her.  My dissertation topic can’t even be approved yet–until I finish more formal KAM research projects.  But it was still a great way to start my day.

Some of the important metrics were only known amongst analysts who specifically dealt with an international series of recommendations, named Basel II accord.  Among other things, Basel II recommended how much liquid funds the banks should have on hand, as a matter of stability

Regulatory capital categorizes bank assets according to how risky the assets are as well as how much control over are.  Tier 1 Capital is regarded conferring the most strength to a bank from a financial point of view.  It generally comprises the core capital (common stock) as well as any disclosed resserve that comprise part of the retained earnings.

Although the credit crisis had many more contributing factors than simply regulatory capital (for examplle, one striking difference between the Canadian and US banking system is how fragmented the US banking systems seems to be.  Canada had also traditionally limited Canadian finacial institutions from participating in  all the functions of a bank we take for granted the do today since regulators “knocked down” three of the four pillars  in 1986.  Previously, banks were not conduct business in all four pillars: banking, trust, brokerage, and insurance.

Structurally, Canadian banks seemed to find strength in their sheer size–with essetially ive providung service across all of Canada–a surface with greeater area than the US.  Meanwhile, another factor generally attributed to the US bank difficulties was the 1999 repeal of Glass-Steagall.  But we’ll discuss that on its own  another day.

Although Canada adopted both Basel I and Basel II, which stipulated 4% Tier 1 capital; the Office of the Superintendent of Financial Institutions decided, however, that Canadian abnks should provide 7%.   Even without applying any amount of statistics, we can see that in 2009, all of the Big Five banks kept their Tier 1 capital ratios (ratio of core equity to total assets).

In 2009, out of the Big Five banks, the lowest Tier 1 capital ratio was Bank of Nova Scotia, which still had 9.7%–still more than twice recommended Tier 1 capital ratio in Basel II.  The highest was CIBC at 11.5%

In fact, in 2009, the average Canadian bank maintained its Tier 1 Capital Ratio about 40% greater than  the OSFI regulations–and a startling 2.5 times the Basel II

Such excess reserves, given that Canadian Banks performed not only safely, but performed well.  Canadian banks were rated #1 in performance during the credit crisis–something that , avoiding the layoffs that were prevalent on Wall Street, seems to indicate that the reserve levels could be optimized.  In order to inform the decision of “how much is enough?:”we must take a step back and elucidate the underlying theory of bank capital structure and what factors most influence stability. Only when we understand the underlying theory can we make changes with  predictable nehaviour.

Day 3 Lunch, Validity, Junk English 030 (800x600) North Park and Residency Checkin 7Jan2010 032 (800x600) Day 3 Lunch, Validity, Junk English 009 (800x600) North Park and Residency Checkin 7Jan2010 038 (600x800) 100_1660 (800x600) Day 3 Lunch, Validity, Junk English 007 (800x600) North Park and Residency Checkin 7Jan2010 033 (600x800) Day 3 Lunch, Validity, Junk English 037 (800x600) Walden DFW Residency Day 3 002 (800x600)

Cross-sectional and Quasi-Experimental Research Designs

Posted on by
This entry is part 2 of 2 in the series Scholarly Research

Spoilt Scientists

Modern scientific inquiry seeks to increase society’s knowledge base as accurately and completely as possible by maximizing validity, reproducibility, and ability to generalize findings of incremental research efforts.  In the natural sciences, research has traditionally been quantitative and experimental in design.  This is largely because the basic components of systems that the natural sciences seek to examine are more fundamental and the relationships between them more simple than those of the social sciences.  Since the components of systems studied in the natural sciences are more fundamental, they behave more like commodities: one atom of carbon, regardless of which isotope, will react chemically in the same way as any other atom of carbon.  Since these fundamentally-indistinguishable objects behave the same, validity can be maximized by obtaining two identical samples and observing the difference in behaviours between a sample that receives treatment and one that does not.

So, Shall Social Science Save Science?

Systems in the social sciences, however, are much more complex than those found in the natural sciences.  Just as physics represents the most basic study of the behaviours of fundamental, chemistry represents another order of complexity that aims to study interactions between these objects.  Yet more complex are biological systems, where macro structures of these objects are governed by factors that could not exist at lower orders of complexity; a protein, for example, can interact with its environment in many ways: it can exhibit chemical reactivity as a complex of amino acids.  It can exhibit intermolecular forces, such as hydrogen bonding.  It interacts based on its three-dimensional shape and configuration.  It also introduces synergistic effects, such as flexibility of interactions with other molecules based on shapes due to differences in strengths between different types of inter- and intra-molecular bonds.  Since the social sciences comprises complex systems of biological individuals, it introduces yet another level of complexity that is more than simply the sum of its parts; it can exhibit complex behaviours that are not simply linear combinations of the behaviours of its component atoms and molecules.  As such, maximizing validity in social sciences research requires novel approaches to research design that often attempts to extract true behaviours at the analytical phase rather than at the design phase.  Because an treatments applied to individuals can often irreversibly change future behaviours, it is often not possible to manipulate only one variable at a time, holding all the others constant and repeating the experiment to examine each variable.  Examining the effects of changing only one variable at a time can also fail to reveal complex interactions between variables.  To this end, social scientists have devised ways that can examine multiple concurrent effects, although at the expense of some validity.

Frankfort-Nachmias & Nachmias (2008) describe experimental designs, such as the classic experimental design and Solomon four group design, as well as cross-sectional and quasi-experimental designs that can be employed when experimental designs are not feasible.  The most appropriate design for a given study aims to maximize validity subject to certain restrictions, such as ethics or institutional review board guidelines.

In cases where it is impractical or unethical to implement an experimental design, social sciences sometimes resort to studying the effects between different individuals.  For example, a researcher wishing to study whether being shot affects attitudes towards gun control for both ethical and practical reasons cannot first survey a sample of individuals about their attitudes towards gun control, shoot a sample of these individuals, and then survey them again afterwards.  In these cases, social sciences may instead survey individuals who have not been shot as well as a corresponding group of individuals who had already been shot.  This type of design attempts to extrapolate a causal relationship without applying an intervention and so loses some validity in establishing time order, one of the vital components of establishing causality (Frankfort-Nachmias & Nachmias, 2008).  Frankfort-Nachmias & Nachmias (2008) point out that cross sectional designs are “perhaps the most dominant design employed in the social sciences” (p. 116).  Rodchua (2009) employed a cross-sectional survey-based research design to establish that Small and Medium-sized Enterprises (SMEs) had significantly different quality costs than large organizations when compared on the basis of prevention costs, appraisal costs, internal failure costs, and external failure costs.  Because the researcher was not able to survey initial quality costs and then change the size of an organization from small or medium to large or vice versa, a cross-sectional approach was employed to study whether a sampling of SMEs had significantly different quality costs to a sample of large organizations.  Because the effect of organization size is compared between organizations rather than within them, differences between organizations such as organizational culture, vision and values, strategies, and industries can threaten the assumption that SMEs have similar quality costs or that large organizations share the same quality cost characteristics.  Using the underlying assumption that the samples of SMEs and large organizations are representative, however, Rodchua (2009) could approximate the effect of changing organization size through statistical techniques such as cross-tabulation instead of experimental intervention.  Maximizing validity of this type of design, therefore, depends significantly on ensuring that the samples are representative of the groups being studied.  Construct validity of the survey itself is also an important factor to consider in this type of design.  Rodchua (2009) stated that the survey instrument was pilot-tested first on a set of 14 individuals to establish reliability; this does not, however, establish that the survey questions themselves measured the desired attributes.

Instruction is another example of an intervention that cannot be reversed.  Cheng (2009) sought to establish that web-based collaborative learning was a more effective method to teach accounting than the traditional classroom-based individual method.  Because the material, once taught, cannot be untaught and taught again to the same individual, Cheng (2009) employed a quasi-experimental design.  While the cross-sectional design examines all subjects only once, it cannot establish definitively whether the intervention being studied caused the difference between groups or whether the difference was caused by other factors that members of each group happened to share in common.  Quasi-experimental design attempts to re-establish some measure of time order by administering both a pre- and post-test; however, it retains the cross-sectional characteristic of examining the effects of intervention on a different group than the untreated group.  Again, validity is adversely affected because the group receiving the intervention is a different group from the control untreated group.  This again introduces the possibility that the members of the group receiving the intervention react differently to the members that did not.

References

Cheng, K. W. K. (2009). The effect of web-based collaborative liearning methods to the accounting courses in technical educatinoe. College Student Journal. 43(3):755-65.

Frankfort-Nachmias, C., & Nachmias, D. (2008). Research Methods in the Social Sciences.  7ed. New York: Worth Publishers.

Rodschua, S. (2009). Comparative analysis of quality costs and organization sizes in the manufacturing environment. Quality Management Journal. 16(2): 34-43.

Walden DFW Residency Day 3 019 (800x600) Day 3 Lunch, Validity, Junk English 009 - 640x480 (640x480) Walden DFW Residency Day 3 038 (800x600) Day 3 Lunch, Validity, Junk English 007 (800x600) Day 3 Lunch, Validity, Junk English 018 (800x600) Day 3 Lunch, Validity, Junk English 030 (800x600) 100_1662 (800x600) Day 3 Lunch, Validity, Junk English 012 (800x600) Day 3 Lunch, Validity, Junk English 009 (800x600)