Overview of the needs assessment process

One goal of RadGrad is to be tailorable to different disciplines. While RadGrad started as a way to help students learn about and prepare for the variety of applications and professional pursuits related to computer science, we believe that this approach can be beneficial to other domains as well.

Let's say you've come across RadGrad and think it might be a useful way to help students learn about the diversity of applications and professional pursuits of your discipline that might not be obvious to an undergraduate. For example, you might suspect that undergraduates at your institution tend to have narrow, preconceived notions about what it means to pursue mechanical engineering (cars!), or biology (med school!), or geography (printed maps!).

A good way forward is to perform a "needs assessment", in which you try to figure out whether or not RadGrad can actually provide useful value to your student demographic. The Needs Assessment process for RadGrad has the following steps:

(1) Develop a domain model for your domain. This involves specifying the Interests, Career Goals, Opportunities, Courses, and (potentially) Internships for your domain. You might discover through this process that your discipline actually doesn't need RadGrad---for example, you can only come up with a small number of each of these entities. In that case, maybe you don't need RadGrad: maybe you just need a brochure. Developing the domain model generally takes 5-10 hours of work, and works best with review by students, faculty, and advisors in your discipline.

(2) Implement a pilot instance. Assuming your domain model looks promising, the next step is to implement a new RadGrad instance with that domain model. Implementation will require you to "flesh out" your domain model by providing high quality descriptions of your entities, external links, teaser videos, images, course descriptions, and so forth. It will generally take several weeks to months to build a pilot RadGrad instance with a high quality, non-trivial domain model, depending upon your software engineering background and experience with the RadGrad tech stack.

(3) Perform a pilot study. The last step is to perform a pilot study in which you walk through your pilot RadGrad instance with selected students, advisors, and faculty. In this step, you can get useful feedback on your domain model and a sense for whether a full deployment would add value. Note that the pilot study won't illustrate some of the value that comes with a full deployment, such as reviews of courses and opportunities or communities of practice. You can perform a pilot study with a few students, faculty, and advisors in a week or two.

Let's look at each of these steps in turn.