Section 7: Creating Quizzes

Section 7: Creating Quizzes

Using ChatGPT to Assess Without the Stress

Quizzes are like the Brussels sprouts of education—some students dread them, but when cooked (or written) well, they’re incredibly nourishing. In the context of your AI-assisted textbook, quizzes are more than just evaluation tools—they’re opportunities to reinforce learning, promote engagement, and reflect the unique focus of your course. And luckily, ChatGPT can help you generate quiz questions tailored to your content, learning outcomes, and tone—without spending hours agonizing over each multiple-choice option.

This section walks you through how to create effective quizzes with ChatGPT, from aligning them with your learning outcomes to formatting them for your course platform. Whether you’re quizzing students on international relations theories or the physics of time travel in film, this chapter will help you create assessments that are both useful and aligned with your textbook.


Part 1: Aligning Quizzes with Learning Outcomes

Good quiz questions don’t just test if students were paying attention—they assess whether students are achieving the learning outcomes you laid out in your textbook. So before asking ChatGPT to generate anything, revisit the learning outcomes from the chapter.

Example:

  • Learning Outcome: Students will be able to compare realism and liberalism in international relations.

  • Matching Question: Match the theory to the belief:

    • A. States act in self-interest in an anarchic world → Realism

    • B. Cooperation between states is possible through institutions → Liberalism

Start by telling ChatGPT what your learning outcome is and asking for questions that assess it directly. This helps ensure that your quiz isn’t just checking for memorized facts—it’s checking for real understanding.


Part 2: Choosing the Right Types of Questions

ChatGPT can generate a wide variety of question types. Consider what kind of thinking or recall you want to assess:

Question Type Best For
Multiple Choice Knowledge recall, concept distinctions
True/False Quick checks for conceptual understanding
Short Answer Application, synthesis, or personalized reflection
Matching Terminology, definitions, or theory comparisons
Fill-in-the-Blank Specific recall of terms or concepts

Example Prompt for a Multiple Choice Question:
“Write a multiple choice question that tests whether students understand the core principle of constructivism in international relations.”

ChatGPT Output:
Which of the following best describes constructivism in international relations?
A. States act primarily out of self-interest in a competitive environment
B. State behavior is shaped by social norms, identities, and ideas
C. International institutions create peace through trade
D. The distribution of power determines global order
Answer: B


Part 3: Prompting ChatGPT to Generate Questions

Once you’ve chosen your question type and tied it to a learning outcome, it’s time to prompt ChatGPT. But first, we need to “chunk” your textbook. Why?

Generative AI tools like ChatGPT work by processing content in tokens, which are small units of text—often a few characters, syllables, or whole words. For example, a sentence like “States act in their own self-interest” may contain 7–10 tokens depending on how the words are split. Each version of ChatGPT has a maximum token limit, which includes both the user’s prompt and the model’s output. For GPT-4, that limit can range up to 128,000 tokens, but practical usage often stays well below that to avoid delays, truncation, or performance issues. A full-length textbook could contain 80,000+ tokens, far exceeding what the model can handle in a single interaction.

To work effectively within these limits, it’s best to divide your textbook into chapters or sections before submitting content to ChatGPT. This modular approach not only avoids token overflow but also leads to better, more focused quiz questions that align tightly with the material. When prompting the AI, copy and paste one chapter—or even a subsection—along with your learning outcome and desired question format. This ensures the model processes a manageable amount of content, allowing it to extract key ideas, terms, and examples without missing important context. Download a Word document or PDF of each individual chapter, then upload that file to your generative AI before asking it to create quiz questions.

Generic Prompt:

“Based on the following textbook content, write five multiple choice questions aligned with the learning outcome: ‘Students will be able to define and apply the theory of realism.’ Include the correct answer and three distractors.”

IR-Specific Prompt Example:

“Using the content from my chapter on Realism and the film World War Z, generate three multiple choice questions for college freshmen. Focus on how Realism explains international responses to global crises.”

Output Example:
In World War Z, the rapid closure of national borders and unilateral state actions reflect which key idea of realism?
A. Collective security
B. Constructed identities
C. State self-interest and survival
D. Global governance
Answer: C


Part 4: Reviewing and Editing AI-Generated Questions

Let’s be honest—sometimes AI-generated questions are… a little off. They might be too vague, oddly phrased, or just not quite on target. That’s okay! AI gives you a strong first draft.

Tips for Reviewing Questions:

  • Read them aloud—do they sound like something you’d ask your students?

  • Watch out for trick questions or double negatives—these often confuse more than they challenge.

  • Replace or revise any culturally specific examples that might not resonate with your students.

Prompt for Refinement:
“Rewrite this question to be clearer and more concise: ‘Is it true or not that the states sometimes act in ways that suggest cooperation over conflict even when realist theory says otherwise?’”
Improved Version:
“True or False: States cooperating in international organizations challenges the assumptions of realist theory.”


Part 5: Creating Answer Keys and Explanations

Why stop at just quiz questions? Ask ChatGPT to provide rationales or explanations for the correct answers—this is great for your own reference or even as feedback for students.

Prompt Example:
“Create an answer key with one-sentence explanations for each answer.”

ChatGPT Output:

  1. C – Realism focuses on state self-interest and survival, which explains the closure of borders during a global crisis.

  2. B – Constructivism highlights how norms and ideas shape state behavior, not just power.

  3. D – Global governance challenges traditional realism by emphasizing institutional cooperation.

Want to add student feedback? Try this:
“Give correct answers and include short feedback comments for students who got the wrong answer.”


Part 6: Formatting and Exporting Quiz Content

Once you’ve got a solid quiz, you’ll likely want to use it in your LMS or quiz platform. You can prompt ChatGPT to format the content for different platforms or layouts.

Prompt Examples:

  • “Format the following quiz for Canvas in a numbered list with answer choices labeled A–D.”

  • “Convert this quiz into a table format for easy copy/paste into Word.”

  • “Create a QTI-compatible format version of this quiz with correct answers indicated.”

Bonus Tip: Ask ChatGPT to bold correct answers, italicize questions, or organize content by chapter—it’s surprisingly good at formatting once you ask!


Conclusion: Quizzing with Confidence

Creating quizzes with ChatGPT doesn’t just save time—it helps you build assessments that reflect your course’s unique focus, style, and structure. From generating multiple choice questions to formatting answer keys, the AI can handle the busywork while you stay focused on the bigger picture: helping students think critically, make connections, and actually learn something.

Just remember: you’re still the instructor here. Use your judgment, revise as needed, and don’t be afraid to ask ChatGPT to take another swing when something’s not quite right. After all, the real quiz isn’t whether AI can write a test—it’s whether we can keep our students awake long enough to take it.