MS Thesis Guidelines & Checklist The post-secondary education of many students culminates in the MS thesis. In the ECE department the thesis committee consists of two members of the ECE department and a graduate representative assigned by the university. The committee must approve both an oral defense of the student's thesis work and the written thesis. Unfortunately, standards and expectations for thesis quality and scope vary among faculty. To maintain a greater degree of consistency, here is a checklist of criteria that I will rely on to determine whether an MS thesis is of acceptable quality or not. This is designed for a thesis that proposes something new, such as an algorithm or design, to solve a specific problem. It may not apply to theses that make theoretical advances. Introduction (First Chapter) Problem: Is the problem clearly stated and fully specified? Objectives: Are the objectives of the thesis clearly stated in the introduction chapter? Significance: Is the significance of the proposed solutions clearly explained? Literature Summary: Does the thesis summarize solutions proposed by others that have been published in conference proceedings and journal articles? If no solutions have been proposed, has a summary of the most closely related literature been summarized and it's relationship to the problem of interest fully explained? Current Practice: Does the thesis summarize the "best current practice", which is how people presently solve the problem in practice. What's Better?: Does the thesis explain the potential advantages of the new design as compared to other proposed solutions in the literature and the best current practice? Contributions: Does the first chapter clearly and concisely summarize the contributions of the thesis? Design Methodology Mastery of Knowledge: Does the thesis make it clear that the student has mastered the methodology necessary to solve the problem? For a thesis proposing new algorithms, this may be a demonstrated mastery of the knowledge of statistical estimation, signal processing, image processing, pattern recognition, nonlinear optimization, etc. Constraints: Are the design constraints and/or specifications clearly and quantitatively specified? Is relevant domain knowledge appropriately considered and leveraged to improve the design? Tradeoffs and Metrics: Are design tradeoffs recognized and discussed? In many cases this may be the dollar cost of design features or the computational cost of algorithm components. Are quantitative metrics appropriately defined to quantify the design tradeoffs? Design Rationale: Are design decisions appropriately justified based on the application and/or sound design principles, or is the design merely empirical? Optimality*: Are any optimality claims made? Is "optimal" defined? Are the claims justified? Assessment/Validation Methodology Benchmark*: Did the student choose an appropriate standard of comparison? Performance Metrics: Are the performance metrics appropriate? Is the choice of a performance metric sufficiently justified? Sufficiency: Once the assessment is completed, will it clearly demonstrate whether the proposed solution is better and/or meets the stated specifications? Prospective Assessment: Was the same data used to motivate and refine the design as was used to measure performance of the new technique? New data should be used. If it wasn't, the use of the same data for both development and validation must be justified and the potential bias in the results must be fully discusssed. Experimental Design: If new data was collected, was a testable hypothesis stated before it was collected? Was the methodology for testing the hypothesis fully developed before the new data was collected? Results Presentation: Are the results appropriately presented with tables and figures? Captions: Do the figure and table captions sufficiently summarize the content of the figures and tables? Can the figures and tables be understood from the caption alone? Are the axes clearly defined? Are units provided for all axes and table entries? Sufficient?: Are the results sufficient to judge the performance of the proposed solution? Better?: Do the results clearly indicate the improvement of performance of the proposed solution as compared to a reasonable standard of comparison? Discussion Anomalies: Does the discussion explain any discrepancies or anamolies in the results? Weaknesses: Does the discussion discuss pitfalls and weaknesses in the proposed solution? Is the solution always better, or only under certain conditions? If the latter, what are the conditions? Interpretation: Is the interpretation of the results accurate and reasonable? Or are the claims broader than what is supported by the results? Conclusion Contributions: Does the conclusion concisely summarize the contributions of the thesis (again)? Futher Research*: Does the conclusion discuss ideas for future research and further improvements in the proposed solution? Conclusion: Does the conclusion concisely state what can be concluded from this work? Organization Grammar: Is the grammar and prose of sufficient quality for an MS thesis? Organization: Is the thesis appropriately organized? Are the section headings appropriate and clear? * These categories may not apply to all theses. Revised 10.26.05