DOC PREVIEW
Characteristics of Good Dissertation Research

This preview shows page 1-2-3-4 out of 12 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

3 Characteristics of Good Dissertation Research• In your dissertation you have to justify that your research contains new contributionsand motivate the importance of your research (i) in statistics as a discipline or (ii) interms of practical applications (if any).• As a PhD student you want to find a good research topic in statistics for a dissertation.Although there is no guaranteed approach to finding a good research topic, the followingare characteristics of good research topics: originality, relevance, manageability,and clarity.3.1 Originality and relevance in statistics.• The topic must lead to original research in statistics. That is, you must present ma-terial that no one else has ever seen. This may include significant modifications andimprovements to existing theory or methods, developing algorithms to solve a compu-tational problem, developing new statistical techniques, and establishing new theoremswith rigorous proofs.• The topic should also be relevant research in statistics. That is, the research you arepresenting should be of interest to academia, and hopefully to practitioners of statisticsin the research area. For example,– Research in response surface methodology will be relevant to engineers and in-dustrial statisticians.– Research in sampling will be relevant to statisticians who design, collect, andanalyze data from surveys.– Research in spatial statistics will be relevant to ecologists, geologists, and otherenvironmental scientists.• After you have a research topic, you need to think about the dissertation in termsof original unsolved problems. These problems determine what you plan to discoverabout the research topic.My personal experiences• The following are examples of the originality and relevance of the research from threedissertations.1. Regarding the dissertation of Prof. John Borkowski• As an improvement to the Taguchi system of designs, this dissertation presentedand evaluated a new class of response surface designs (called minimum aberrationmixed resolution designs) which are based on a single factor array.– Originality: Creating a new class of designs.– Relevance: Improvements to the Taguchi system of designs.23• The mixed resolution designs were extended to another new class of responsesurface designs (composite mixed resolution designs) that allow estimation ofquadratic model effects. This will then lead to improvements in the design ofexperiments for achieving a robust process.– Originality: Extension to another new class of designs.– Relevance: Improved designs for a robust process.• In the evaluation of mixed resolution designs, the problem of finding globallyoptimum designs for achieving a robustness process was addressed and solved.D− and G−optimal designs for the mixed resolution model were found and theiroptimality properties were studied. The D− and G−efficiencies of the mixedresolution designs were then calculated.– Originality: Studying optimality properties of designs with respect to themixed resolution model.– Relevance: Finding theoretically D− and G-optimal designs.2. Regarding the dissertation of Dr. Philip Turk (a former advisee who is now an AssistantProfessor at West Virginia University)• Multiple factors known to influence the efficiency of adaptive cluster sampling(ACS) were studied using a designed computer experiment and application ofresponse surface methods to computer-simulated data. This approach allows sig-nificant interaction and quadratic effects to be studies in contrast to ignoring themas had been done in the literature.– Originality: Using a computer experiment to study factors related to ACS.– Relevance: Understanding interaction and quadratic effects that had not beenstudied.• The efficiency for the Horvitz-Thompson estimator of the population total τ wasstudied relative to simple random sampling without replacement. Recommenda-tions were proposed that can be applied in practice.– Originality: Studying the relationship b etween efficiency and ACS study fac-tors. Determining how and which study factors affect ACS efficiencies.– Relevance: Providing recommendations for practical use.• Current research is focused on comparing different simulation approaches to gen-erate populations following a Neyman-Scott point process. It was shown thatthe validity of conclusions regarding ACS will depend on the simulation approachused to generate the population data.– Originality: Investigating different simulation approaches.– Relevance: Showing that conclusions depend on the choice of simulationmethod.243. Regarding the dissertation of Dr. Julia Sharp (a former advisee who is now an AssistantProfessor at Clemson University)• Her dissertation research was centered on applications of statistics in the field ofproteomics which is the study of protein compositions in biological substances.Likelihood ratio tests and exact tests were developed to analyze data from pro-teomics laboratories.– Originality: Developing statistical methods to analyze proteomics data.– Relevance: The identification of proteins as well as establishing the presenceor absence of certain proteins in biological substances has far-reaching appli-cations in the biological sciences, especially in medicine.3.2 Manageability with respect to a time frame.• You must be able to do all of the work required within the time frame set by the Ph.D.program at your university. This is obvious, but most students do not think enoughabout the details. Suppose, for example:– You are scheduled to present your Ph.D. proposal on November 1. When shouldyour committee receive a copy of the proposal? Suppose they need 10 days. Thenyou have to have it ready to submit on October 22. So can you just type theproposal and hand it to the committee on October 22? No.– You should not give the proposal to your committee before your advisor reviewsit. Therefore, you will need to answer the question “When should my advisor seethe proposal?”. The answer to this will depend on other questions such as:(i) How much time will my advisor need to review it? Suppose the advisor needs10 days.(ii) How much time will I need to make the changes recommended by the advisor?Suppose you need 5 days.(iii) After making the revisions, does my advisor want to read the proposal again?Suppose the answer is yes. This will take another 4 days.(iv) How much time will I need to make the changes recommended by the advisor?Suppose you need 2


Characteristics of Good Dissertation Research

Download Characteristics of Good Dissertation Research
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Characteristics of Good Dissertation Research and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Characteristics of Good Dissertation Research 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?