UG dissertations

This page provides an overview of some things that I think are useful if you consider writing a dissertation with me.

General things:

Some general tips about writing a dissertations can be found here: Writing tips for dissertations

A list of past topics for inspiration and to get a general idea about what people have been doing: Past topics (undergraduate)

Picking a topic and preparing for the first meeting with me:

If you feel intrigued by one of the broad topics that I offer in a year, you should think briefly about whether there is something that particularly interests you. It is perfectly fine if this interest is still relatively broad: Say you're interested in the economics of crime. One thing you might want to think about before meeting with me in this case is whether you're more interested in the determinants or the costs of crime. Obviously, if you have a very specific idea - even better. Just don't think that you have to come up with a perfect question and read lots of literature before meeting with me. Essentially, try to meet with me within the first two weeks after signing up with me.

One thing to be aware of is that every dissertation with me will involve an element of econometric analysis of some data. This doesn't have to be super-sophisticated econometrics - I had a dissertation marked at 80 that just used stage 2 econometrics (essentially OLS). If you can find something that let's you do some difference-in-differences or some instrumental variables, e.g., some policy changes, however, you should seriously consider this, as most of these dissertations turn out to be really nice.

My aim for the first meeting is usually twofold:

1. First, we should try to agree on a relatively specific topic/research question, i.e., after the meeting you should have a relatively clear idea what you're going to do over the next year. This will likely involve some trade-off between what interests you and what is realistically doable as a UG dissertation (or what is doable at all). Having a (specific) question early on is in my experience beneficial as it makes it easier for you to look for literature, theories etc. Occasionally, this might not be possible for numerous reasons: I once had a case where someone wanted to write about the impact of the Webster-ruling on footballers' wages. In this case we weren't sure that the necessary data exists and agreed on some time frame after which we would reconsider whether this topic would be possible (it was, but with a slight modification - to look at transfer fees instead - and turned out to be a really nice dissertation). Delays like this are generally not a big problem, as long as both of us have a clear timeframe and know what we're doing.
2. Second, in most cases as soon as we have a question, I will have an idea about an existing dataset that you can use. Having a good idea about what data to use early on is beneficial as it prevents us from getting into potentially catastrophic situations later such as having a nice theory and a nice literature review, but having no way to actually test any of this because there simply is no data. Plus, in an empirical thesis the data somewhat dictates what can be done. Experience shows that most of my dissertation students use existing survey or government datasets. However, if the topic calls for it you can also collect you own data. Just be aware that collecting data will costs lots of time - in the football case above the student collected data on all premier league transfers (including player statistics, etc.) for several seasons (took him a couple of days). If you want to do questionnaires, interviews or experiments, we might also need to go through a bit of a hassle with ethical approval etc.

Afterwards:

Most of the things that come after the first meeting, will become clear during that and subsequent meetings. There are few general rules as things will depend on your topic and your working style.