Summer Undergraduate Research Experience I
and is replicated here as part of the SERC Pedagogic Service.
In this collaborative summer research experience, one or more faculty members work with a student to co-author an original research paper. The ultimate goal is to publish the paper in a disciplinary peer-reviewed journal. This example discusses a period of time that covers roughly 18 months: from the point of initial work on an internal proposal for summer funding through the point where the paper is first submitted to an appropriate professional peer-reviewed journal for publication.
- To critically read professional research;
- Advanced research methodology (e.g., graduate level applied econometrics related to project);
- Advanced computer skills (e.g., STATA, Matlab, etc.);
- To write highly technical, professional-quality papers;
- To present research findings to a professional audience;
- About the economics profession (i.e., what graduate school is like, what professional research is like, etc.)
Context for Use
Choosing the right student is crucial to the success of these kinds of collaborations. For the most part, I have worked with students who are planning to pursue graduate school in economics or related fields. These are typically rising seniors (i.e., they start the research in the summer between their junior and seniors years). The main reasons have to do with maturity, content knowledge and skills. I mostly do applied econometric research so I need students who have taken econometrics. Our econometrics course uses mostly SAS. However, to conduct best-practice applied work in economics these days, STATA is a virtual must. Thus, students will have to learn a good deal about STATA during the summer. Students also need to be able to read economics journal articles. One major criterion these days is good calculus training. Given the sequencing in mathematics, not all economics students at this stage will have seen constrained optimization, partial differentiation or comparative statics. This can be taught relatively quickly as long as they are good mathematicians. While I do not always put formal models into my papers, certainly the papers they read are likely to have them so they need to be able to understand straightforward economic models. I also find that constructing a simple constrained utility maximization problem and performing some comparative statics is a good way for undergraduates to come to understand the underlying theory upon which the ultimate econometric model will be built.
Mostly, students need to be motivated. For this reason, I usually wait for them to approach me first about their desire to work with me in the summer. In addition, students must be able to meet deadlines, keep copious research notes, write well, and be obsessive about data. This rules out the vast majority of undergraduates to be sure.
In reality, "summer" undergraduate research experience is a misnomer because there is no way a project can start and be completed in two months. The idea is to start the project in summer. However, it is important to make it clear to the student that they will continue to work throughout their senior year on the project to get the paper ready to submit to a journal. Given their own course loads (and my teaching), it is essential that we make sufficient progress during the summer so that the fall and spring semesters are used mainly for writing, sensitivity analysis, presenting the paper, and preparing the draft for submission. All projects differ, of course, and there is no way to know what roadblocks you will run into along the way. But having done a half-dozen of these over the years, I have found the following timetable to be pretty realistic.
- February-March: This time is devoted to collecting basic references from related literature, developing a competitive proposal to submit by mid-March (6 weeks).
*note: decisions are made in mid-April, but with final exams, etc., there is little more time to make progress on the project until summer begins the first week of June.
- June: We finish collection of literature (and typed annotated bibliography) (2 weeks), do some simple comparative statics exercises (2-3 days), and start initial construction of dataset (2 weeks).
- July: We begin econometric modeling, continue to construct the dataset, make appropriate transformations and scaling, etc. (2 weeks), and start getting some preliminary results, including descriptive statistics and at least a basic OLS regression (2 weeks).
- September-December: Students typically sign up for 1 credit hour of independent research (which means they work with me in the "lab" 4 hours a week), which gives us time to complete modeling, testing and sensitivity analysis by the end of the semester.
- February-May: With results from the fall, the spring is (ideally) all about writing and preparing for conference presentation. Again, students sign up for 1 credit hour of independent research. We write the paper together. Usually I have them take the first shot at all the sections (after all, this first and foremost a learning experience for them), with me coming in to smooth it out and deal with the more complex topics (e.g., if we have to write about econometric modeling issues such as describing our identification strategy in an IV Tobit model, I do that). We complete first draft and present paper at spring economics conference (e.g., eastern Economics Association meetings, Mid-West Economic Association meetings, etc.).
- June-August: After graduation, I usually take full control of the final draft to prepare it for submission. Via email, student collaborators can easily read drafts (they are especially good at catching confusing/bad writing). Realistically, though, if they are starting graduate school or jobs, they cannot be expected to do any more analysis on the paper. Moreover, at this point the value of getting the paper accepted is higher to me than it is to them! So all subsequent paper revisions, etc. are done by me, with them doing editing and sometimes helping with additional literature as needed.
*Note: obviously this timetable only covers up to the point of first submission to the targeted journal.
Teaching Notes and Tips
I usually work with them on Tuesdays, Wednesdays and Thursdays during the summer. We start with a morning meeting where I ask them to update me about whatever they were working on the previous day. After that, we set out the plan for the day. Typically we keep a detailed "to do" list written on the large whiteboard in the room. When I am not going to be in the office the next day, we always have a meeting at the end of the day to set out a plan for them to follow the following day. They always have my cell phone number, but I want to make sure they know what needs to be done. As the summer wears on, these meetings are more collaborative, with me asking them what they think needs to be done next. Particularly when they get into the data, they have a well-specified "to do list" so this becomes easy. When I am at the office, though, I am usually also working on other research projects (revise and resubmits, etc.). My office is 2 doors from the lab, so they can come get me when they have questions. Most questions are either about them not understanding a section in a paper they are reading or confusion regarding writing STATA code.
While I do not plan long vacations during the two months these programs run, I often go to conferences or the beach for 3-4 days. Usually I am out of the office for only 2 days a week. Again, the key is that time is not wasted. They have to be independent and hard-working, but it is my responsibility to make sure they know what to do. This is their first time doing this kind of research so I cannot assume they know what needs to be done, at least not in the early weeks. By July, however, they slowly start taking control and my job gets easier.
References and Resources
Link to forthcoming journal article in World Development based on the 2008 project