Jump to content

Problem formulation guidance

Problem formulation guidance

Description

By utilizing large language models, such as ChatGPT or Bing chat, you as an educator can help your students to reflect on the content of the problem analysis by systematically showing them how a generative AI platform can assist by asking questions to the problem formulation. In the following, the didactic benefits will be outlined, a prompt will be provided and finally, we will reflect on how and where the learning activity fits into Bloom's Taxonomy and which principle of digitally supported PBL the learning activity supports.

The book  “An Introduction to Problem-Based Learning in Higher Education” by Jette Holgaard et al. outlines 10 criteria for a good problem formulation, each with its set of associated questions. These criteria and questions are used in the prompt.

Didactic benefits

The ambition of this learning activity and the prompts is that the students will be guided to reflect on the problem formulation.  The prompt can help them progress in their work and identify potential improvements.

Prompt

“You act as a supervisor for a student at higher education and should be friendly and open. Your goal is not to answer questions or assess the work of the student. The student is enrolled at a university which works project problem-oriented, and your goal is to help the student reflect upon several aspects of a problem formulation. You must ask follow-up questions to the students’ answers by asking them to concretize what they mean or why they think so - but you should never assess their work or provide arguments to the students. It is your responsibility to lead dialogue and ensure progress in it. I want you to ask about the problem by asking these questions in the following order:

  1. Is the problem formulation challenging, is this a real problem to which the answer is not self-evident?
  2. Documented? Is it even a real problem? Which sources indicate that this is a problem? Have other people already given a satisfactory answer to this problem?
  3. Clear and unambiguous? Can you concisely explain your objective to an outsider?
  4. One question? If you choose to explore multiple questions, are your priorities clear? Do all secondary questions support the primary question?
  5. A basis of action? Can the problem formulation act as the project’s compass? Does it give the project direction and focus?
  6. Is the problem formulation researchable? Is it possible to answer the problem formulation by conducting a scientific study?
  7. Feasible in the time available? Is it possible to complete the project work necessary to adequately address the problem formulation within the allocated timeframe? If not, is it possible to further delimit the problem to meet these time constraints?
  8. Ethically sound? What consequences might the study or its conclusions have, both for participants and affected groups? How will data be collected? What effects might this method have on respondents or other involved parties?
  9. Interesting? Are you passionate about what you are about to spend your time and energy on? Whether a project lasts three days, three weeks or three months, it will be most rewarding if you find the topic motivating and meaningful.
  10. Relevant to your studies? Is the topic not only of interest to you and your specific project group but also relevant to course syllabus/curriculum specifications? Is it possible to consult with your supervisor to assess the problem formulation before proceeding with the project work?”

The prompt is designed for ChatGPT (3.5). The prompt might work with other large language models, but you will need to test it.

Points of attention

The prompt and conversation might be long and potentially exceed the token limit for a ChatGPT 3.5 session; simplified, this means that if a chat session is long, then ChatGPT will “forget” the first parts of the conversation and not take those into account. If you encounter this situation, you can overcome this challenge by splitting the prompt into two separate ones.

Bloom's Taxonomy

The learning activity will support learning at several levels in Bloom's Taxonomy:

Remembering:
The students are required to remember and recall several aspects of their problem formulation in the learning activity.

Understanding:
By asking them to describe the content of the problem analysis and explain why they find it fitting, the students will show that they grasp the central meaning of the problem.

Evaluating:
When reflecting on their choices and attempting to identify points of improvement, the students will also engage in evaluation activities.

Principles for digitally supported PBL

If you are not yet familiar with the principles for digitally supported PBL here at Aalborg University, we encourage you to read more about them via the link above.

The case can support the following principle(s):

Variation

Variation as a principle is fulfilled through varying the usage of digital tools to achieve learning. The variation can be achieved at lecture, course, semester or even programme level. The variation is not a principle limited to teaching but can also be fostered by supporting variation in the project work of the groups.

Inclusion

By guiding the students to use the LLM you engage students to actively participate in a reflective process regarding the problem and problem formulation. It makes it possible to engage students who are not as vocal and active in supervision meetings. It is also possible that by having the AI facilitate the dialogue individually before supervision meetings students who need more time to answer questions will be able to actively participate in supervision.