A Note to and Exercise for Prospective External Ph.D. Students

I previously described why it is difficult being an external Ph.D. student. One of the aspects is that you have to convince me that you can prevail, even though I have much less opportunity to advise you than I’d have with a “wissenschaftlicher Mitarbeiter” (TA/RA position funding Ph.D. studies) that I employ.

So, if you are thinking about approaching me to become your advisor, I recommend you go through the following mental exercise first: To layout the work that motivates you as a combination of research question, possible hypotheses, research approach and employed methods, and expected results. You should be able to write this up as a 100-200 word abstract.

Here are two examples (I wrote this blog post on a whim, so please don’t hold me to the quality of these examples):

This dissertation answers the question whether a pair of programmers is more productive than two individual “solo” programmers. We hypothesize that working as a pair, programmers develop less but higher-quality code than two solo programmers. We use a series of student and professional programmer experiments and find, independently of the level of professionalism, that pairs produce only about 60% of the code (mean) that solo programmers produce combined, but that for every bug the pair produces, the two solo programmers produce about 4.5 bugs. […]

The research question in this example is the question of pair programmer productivity, a specific (not always necessary) hypothesis is the inverse code amount / bug rate relationship, experiments are the chosen research method, and the results, well, they have been fabulated since we haven’t undertaken such a thesis (yet).

Here is another fictitious example:

This dissertation develops a theory of social presence in software development. We hypothesize that making developers aware of which other developers are working right now motivates them to stay longer and work harder. We derive a qualitative theory from a series of interviews with developers at Bunga Inc, an Internet startup. After using a sign-in/sign-off board provided positive preliminary results, we developed a software tool for monitoring developer presence and activity. In a series of case studies at companies of various sizes we find that awareness of social presence does indeed motivate developers at startups, while at large established corporations it has exactly the opposite effect. We conclude that judicious use of our theory and tool can improve performance at a subset of software companies. […]

So you are writing these abstracts as if you already finished your thesis. I find this a useful exercise to straighten out your thinking. If you can’t do it convincingly to yourself, well, maybe you have some more thinking to do as to your thesis idea.