Scholarship for All in Computer Science

This past year, the Grinnell College faculty passed a resolution that we would provide every student with the opportunity to pursue a research experience. The final resolution read as follows:

Student research is focused inquiry supervised by a faculty member that seeks to answer or explore an intellectual and/or creative problem or question. Grinnell College seeks to provide every student an opportunity to pursue research and share the product publicly. [1]

We tend to refer to this as the research for all resolution, even though it doesn’t really guarantee research for all. In case it’s not clear from the text of the resolution, our discussion made it clear that (a) not every student has to have a research experience, since not every student will want to have a research experience; (b) research is interpreted fairly broadly, in that we don’t expect students to necessarily develop new knowledge; (c) our Dean would prefer that these research experiences build upon prior work, so the significant research experiences that students have in, say, BIO 150 [2] won’t count [3]; (d) every department is expected to come up with its own plan for providing these research experiences; and (e) we’re supposed to add these research experiences without any more faculty resources. Oh, and there’s a significant stick - faculty are not eligible for our new expanded sabbatical plan unless (a) their department has a plan for providing research opportunities and (b) they contribute to research in the department.

This essay is my attempt to think through how we will approach these expectations. Some text is taken and modified from too many memos to the Dean that were collaboratively developed last year.

Let’s start by considering what we mean by research. The Computer Science Department has consistently taken a broad view of the concept of research, preferring instead the term scholarship [4,5]. Following the work of Ernest Boyer in Scholarship Reconsidered, we note that scholarship includes at least four types of activities, the scholarship of discovery; the scholarship of integration; the scholarship of application; and the scholarship of teaching [6]. Each of these areas can involve pushing current knowledge in new, creative and innovative ways. The scholarship of discovery is what folks traditionally call research; it’s the creation of new ideas in the discipline. Boyer notes that it is as important to bring together ideas from multiple disciplines (integration), to apply these ideas to new projects (application), and to develop ways for others to learn these ideas (teaching).

The department strongly believes that all four mechanisms can be particularly appropriate for scholarship in computer science. In fact, we spent a lot of time putting together a departmental statement on scholarship that affirms these four principles [7,8].

For students, there is also a fifth possible kind of scholarship, which I’ll call the scholarship of replication. Students learn an enormous amount when they try to replicate a research project that they’ve only read about in the literature. In addition to learning a lot about the research process itself, they learn about the hidden subtleties that don’t get mentioned in papers. They also learn to be very close readers.

In addition to promoting all five kinds of scholarship, we believe that different students will benefit from different kinds of scholarly experiences. Some will be best served by spending a full summer generating new knowledge. Some will be best served by developing teaching materials. Some will be best served by working on a large software project, exploring the scholarship of application in the process. Some will be best served by trying to synthesize the literature, most commonly by writing a long review paper. And, as noted in the prior paragraph, some will benefit from replicating an existing project.

That’s the background. Let’s move on to a possible statement of research for all in my discipline. This version is a very rough draft. I’ve written it myself, rather than with my colleagues, although I think I’ve discussed most of the issues with my colleagues. We’ll talk about it soon.

The Grinnell College Department of Computer Science encourages every student to participate in a faculty-supported scholarly experience appropriate to their skills and interests. To accommodate a variety of student interests, we provide a variety of opportunities for students. We hope that each student will avail themselves of one or more of the following opportunities for scholarly work.

• An extensive research project, most typically through the College’s Mentored Advanced Project (MAP) or Mentored Introductory Project (MIP) programs. Because of the particular benefits associated with concentrated effort on a project, we prefer to offer such projects through full-time summer MAPs and MIPs. However, we will, on occasion, offer academic year research projects. While many MAPs and MIPs focus on developing new knowledge, others will emphasize the creation of teaching materials that will support others learning important topics in computing. We would also accept a summer REU (Research Experiences for Undergraduates) experience under this model.
• An extended on-campus software development experience. All students are expected to contribute to a large project through CSC 322, Team Software Development for Community Organizations. However, for those with particular interest in software development, a second semester of CSC 322 in which they continue on the same project would provide an appropriate scholarly experience.
• A faculty-supervised, for-credit internship. A for-credit internship is particularly appropriate for students interested in better understanding how CS principles can be applied to real problems. By doing the internship under faculty supervision, the student is able to consider what they are doing from a broader context.
• A course designated as having a significant scholarly component. Many of our upper-level courses, particularly our 300-level special topics courses, include 3- or 4-week modules in which students either develop something new or replicate an existing scholarly project.
• A series of public essays about computing topics, written with faculty supervision. Public scholarship is particularly suited to many liberal arts students. These kinds of scholarly experiences will most frequently take the form of 2-credit individual studies.

That’s my draft departmental statement. I’m sure that my colleagues will have fun both wordsmithing it and debating the five categories. We should also consider whether we should make do at least one of these things a requirement for honors.

But, you know what? The Dean’s office wants me to answer five questions instead of just writing a statement about scholarly expectations. So, let’s look at those questions.

1. Do your students currently have a research requirement? What is the product of this research?

2. Do students have the skills they need when they undertake this research experience? Where are students acquiring the skills to undertake research?

3. Are your department members satisfied with your current approach?

4. Do you want to make modifications to your curriculum? In what ways?

5. Where will students present work publicly? How will they be prepared to do so?

Well, my answer to the first question is no. I think that means that the answers to the second half of 1, all of 2, and 5 are n/a. But perhaps I should be a bit less snarky. So, here goes.

Do your students currently have a research requirement?

Our department does not currently have a research requirement. However, based on the recent research for all resolution, we are exploring recommendations for ways in which all students might have the opportunity to have a scholarly experience within the discipline. Right now, our primary research experiences for students fall within the MAP and MIP programs, although we also give students scholarly experiences in some upper-level courses and other venues.

What is the product of this research?

It depends on the project. Some scholarly projects result in papers about new results, intended for a scholarly audience, most frequently in the form of refereed conference papers. Some scholarly projects result in new pieces of software that are used by others. Some scholarly projects result in literature review papers that can be shared with others in various ways. Some scholarly projects result in educational materials to help others learn more about the discipline. Some scholarly projects result in what we might deem public scholarship; useful writings for a general audience.

Do students have the skills they need when they undertake this research experience?

It depends on the student and the experience. It is rare that a student has all the background they need for an advanced project. They will often need to read the literature related to that project and develop some new skills. However, every student comes to their research experiences with a reasonable understanding of algorithm development and analysis, an ability to build things in code, and some mathematical skills.

Where are students acquiring the skills to undertake research?

Once again, it depends on the student and the experience. We usually build on a core set of knowledge and skills developed through our introductory sequence, CSC 151 (Functional Problem Solving), CSC 161 (Imperative Problem Solving and Data Structures), and CSC 207 (Algorithms and Object-Oriented Design). For some projects, we draw upon other courses. For example, projects in computer vision often draw upon skills that students develop in either CSC 261 (Artificial Intelligence) or CSC 262 (Computer Vision). Students also acquire many skills relevant to work in their Mathematics and Statistics courses, particularly MAT 208 (Discrete Structures), MAT 209 (Applied Statistics), MAT 215 (Linear Algebra), MAT 218 (Combinatorics), and MAT 210 (Experimental Design).

Are your department members satisfied with your current approach?

Mostly.

Do you want to make modifications to your curriculum? In what ways?

We’d love to have the staffing to be able to offer more courses, but I don’t think that’s really what you are asking here. We do not think we need to make many significant modifications to our curriculum in order to provide students with opportunities to do scholarly work, provided we continue to embrace an appropriately broad version of scholarship. We are exploring potential changes to some of our elective and special topics courses so that we can have a moderate-length (3- to 4-week) project as part of the course, In doing so, we might increase prerequisites for some courses and/or move them to the 300-level. In thinking about the ways in which students might do scholarly work, some of us have found value in considering a Writing about CS course in which students develop technical ’blogs or some equivalent series of essays.

Where will students present work publicly?

Many of our summer MAP students have their research accepted to peer-reviewed international conferences. Those students present their work at those conferences. Almost every student presents their research as part of the Computer Science Department’s weekly CS Extras series. Students who do large software projects in CSC 322 provide them in a public forum at the end of the semester in which we invite all students in the department as well as community members. Some present their work through public papers. Some have research posters which we post in our hallway.

We are also considering starting (or restarting) a departmental technical report series as a place for students to present preliminary work.

How will they be prepared to do so?

It depends on the student. Some students have clearly figured out how to develop a presentation or paper or poster. Others need individual faculty mentoring.

We also encourage them to attend Scholar’s Convocation so that they can see how successful scholars present [9].

That may be one of the longest essays I’ve written for this series, at least in terms of time spent [10]. Still, I found it useful to think about how I’d express these suggestions to students. More importantly, I had to have drafts of both a department statement and answers to the Dean’s questions, so it’s mostly writing I had to do anyway.

Of course, I’m not sure that I found the Dean’s questions all that useful [11], since they follow my normal experience with surveys from the Dean’s office - the questions generally don’t fit the same world view in which I work. But they did remind me that I wanted to consider a technical report series, so maybe the were useful to me. (And maybe the answers will be useful to the Dean.)

[1] Final wording taken from Minutes from Grinnell College Meeting of the Faculty, April 4, 2016, available in a rabbit hole somewhere in cyberspace.

[2] BIO 150 is a spectacular introductory biology course. About a decade ago, our Biology faculty looked at the expanding set of material that gets crammed into a traditional introductory biology textbook and asked themselves, What do we really care about our students knowing? They concluded that they want students who can understand the literature, design an experiment, conduct an experiment, and present the results. So our core Biology class doesn’t ask students to memorize any particular information. Rather, it trains them to be real scientists. Each section has a different research subject, often closely related to the faculty member’s only research interests. And all the sections get together at the end of the semester for a huge research poster fair. The course should be the envy of science departments everywhere.

[3] Well, during our initial discussions, the Dean said that BIO 150 wouldn’t count. More recent conversations have suggested that the Dean’s office is willing to consider a broader understanding of research.

[4] The College also uses the term scholarship rather than research in the faculty handbook.

[5] I did suggest that the research for all resolution be a scholarship for all resolution, but others seemed to think that the term scholarship set a higher bar.

[6] Earnest W. Boyer. Scholarship Reconsidered. p. 16.

[7] Grinnell College Computer Science Department. Scholarship Expectations. Online document available at <http://www.cs.grinnell.edu/drupal6/sites/default/files/scholarship- expectations_0.pdf>.

[8] Although we developed that document in response to a request from the Dean and the Personnel Committee, neither the Dean nor the Personnel Committee has signed off on this document.

[9] Well, we may be an overstatement. I encourage them all to attend convo. My colleagues are more mixed.

[10] Also in terms of number of words. I think this is slightly more than twice the average length of my essays. But that makes sense: The background is effectively one essay, the answers to the Dean’s questions are effectively a second essay, and the department statement adds a bit more text (and a lot more time).

[11] Sorry, Dean L. Although I approached the questions with a generosity of spirit, I don’t think I need the scaffolding. I’d rather just develop a draft policy. I also would have appreciated a question of the form What resources (other than more faculty) can we provide to help your department achieve these goals? Maybe I’ll think about that question in a future essay.

Version 1.0 of 2016-09-27.