Basics

Every SDM student (indeed, every student completing an MS degree at MIT) must write a thesis. You will need to come up with a topic and find an advisor who shares your interest in the topic. Some students prefer to find a cool advisor and then get an idea from that advisor instead of coming up with one on their own: this is OK. Yet other students, especially those sponsored by or working with some organization, integrate their work and thesis together into an in-depth project in some area of interest to the organization (even in this case, though, you must still find an MIT advisor interested in the work).

Start thinking about your thesis early. Especially if you are a 13-month student, you have very little time for thesis research and writing. Don't panic, a lot of people get nothing done until the fall. But you can make your life easier if you start early.

The Thesis Seminar is a small but required course during your first Spring semester. In it, SDM program staff will present ideas and methods for approaching the entire thesis process. They will also bring in professors who have ideas for theses and want to serve as advisors. This is a good way to hook up an advisor early.

Online Resources

The MIT libraries maintain a list of thesis resources, including specifications for writing and submitting the thesis.

Finding an Advisor

Popular advisors routinely have more students who want to work with them than they can fit in. This definitely includes a number of the senior Sloan professors, such as Cusumano, Henderson, Utterback, von Hippel, and Roberts. Don't be offended if they don't want to be your thesis advisor: it's not personal. They often recommend another faculty member, usually a more junior or non-tenured one, to be your advisor instead. Engineering professors tend to be easier to hook up with, and ESD professors tend to be easiest of all.

Tips

1. Limit the Scope.

2. Solidify your idea as early as possible.

3. If data is required, get it as early as possible so it doesn't hold you up at the end.

4. Limit the scope.

5. Have at least one advisor from engineering. Note: few people agree on this.

6. Plan your work and work your plan.

7. Limit the scope.

8. Pay attention to the criteria for the Exec. Sum approval - the actual expectations may be the opposite.

9. Don't spend too much time doing literature search.

10. Recognize only a 1/3 to 1/4 of the first thesis semester will actually be applied to thesis work. 2nd semester use only 1/4 on research, 1/4 structuring, 1/4 writing and 1/4 approval cycle.

11. Limit the scope.

12. Don't let advisor goals (extensions of knowledge, publications, etc.) supercede your goals graduation,corporate issues, etc.).

13. Select Advisor early.

Lessons Learned

1. The threshold for an acceptable thesis is slowly being raised towards that of a full-time standard discipline MSc thesis even though they have a lesser course load. To avoid getting caught out by this read the ‘engineering’, ‘management’, and ‘systems’ expectations very carefully and make sure you can tick the right points off.

2. Be aware of your advisors commitments (busy/popular).

3. Ask in detail about how they manage the process and what their expectations are.

4. If you can, don't let it (thesis) evolve.

5. Don't rely on outside data sources.

6. Define your analysis method early.

7. Delivery in complete "chunks".

8. Limit the number of data sources-the more data sources the longer it take to analyze, interpret and write about.

9. If you are self-sponsored you probably have poor access to data sources so think more about a model-based thesis and/or a conceptual thesis. These require good co-ordination with faculty so select your faculty early.

10. If you are on the 13 month track then you have no time to write a thesis – so start in January by reading previous theses and noting the subjects, the standards, and the advisers. Then select a combination you can cope with, and discuss your constraints early and openly.

11. Be careful about having two advisors as this normally means extra work.

  • No labels