thesis. Some evidence from experiments, design, empirical data, or analysis is needed to help substantiate
the thesis.
2.5 Straying Far from Accepted Systems Engineering
Only the best students with strong convictions about their intended topic should even think about straying
far from systems engineering accepted knowledge, methods, processes, and techniques. In most cases,
students who stray from the larger body of knowledge end up with a poor thesis, and in some cases they
are unable to complete the thesis (perhaps because nobody in the department is willing to sign off on
the thesis). Remember a thesis must be substantiated with evidence derived from one of the research
methods: experimental, empirical, design, or analysis. As an example, it is likely risk management models
and practices are not perfect, but if your thesis idea is a complete departure from current risk management
methods, then it is unlikely anybody will sign off on it unless you are able to provide substantive and
compelling evidence to support it.
2.6 Statistical Thinking or Data Analysis
Most research methods generate data, and just having the data is insufficient. The thesis should use
appropriate experimental design methods, statistical methods, and data analysis methods on the data.
Experimental, empirical, and analysis types of thesis benefit from proper statistical analysis. In a design
thesis, you need to actually build some prototype to test the design and data might come from the
prototype. Again data analysis would be necessary.
2.7 Validation
In a thesis, you make a claim, you perform engineering research to gather data to prove it, and you validate
your method, data, and results. A thesis that makes a claim and offers no validation of the data is a
rather weak thesis. Validation is not a yes or no type of proposition, but instead should be viewed as the
degree of confidence a person has in the research results. For example, if you use simulation experiments,
then the validation question is how well do the simulation results match the data you expect from actual
systems? The most convincing validation technique would be a statistical test such as the student t-test
comparing simulation results to known actual results. However, if actual data is unavailable, there are
other validation techniques such as face validity via subject matter experts, boundary testing, and so
forth. The bottom line is every thesis needs some validation via data and other evidence to support
the claims and evidence. All the thesis research methods have valiation. Experiments generate data for
validation, empirical studies produce data for validation, design thesis create prototypes for validation,
and analysis thesis generate results for validation.
2.8 Lack of Technical Rigor
A thesis should have a rigor at least equal to the level of the analysis done in the coursework. Actually,
if the thesis topic is scoped properly, then the level of detail in the analysis should be greater than what
is done in class. The reason is you have one to two quarters to do the analysis, whereas in class you
probably had only one to two weeks to do something equivalent. A thesis is supposed to be a focused
study so it is expected to have sufficient depth to fully explore a narrowly scoped topic.
2.9 Literature Review
The literature review is an important component of your research. The literature review critically assesses
the scholarly peer-reviewed, archival journal articles, conference proceedings, books, and other publica-
tions in order to establish your knowledge of the research topic, identify the state-of-the-art in the field,
and to put your research in context of the larger body of work. A good literature review is more than a
5