Introduction

In earlier sections of this chapter, we examined the process of writing, refining, and evaluating a research question. We outlined key strategies for narrowing the scope of inquiry with an emphasis on creating questions that are not only relevant and original but also generative insofar as they can drive sustained, meaningful research. In the section below, we demonstrate how these principles work in action.

Prompt Engineering Your Research Question

Let’s start with a very basic question and evaluate it. First example:

     Does Baldwin’s story, “Sonny’s Blues,” criticize racism?

Upon evaluation, we determined that this question is too general and that it lacks original or generative potential because it can easily be answered with a simple yes or no. To address these shortcomings, we refine the question to be more specific and open-ended by removing the possibility of a binary response.

Here is our second attempt:

     Is Baldwin’s critique of racism in “Sonny’s Blues” different or more effective than critiques of racism in works by other African American authors?

While this version is an improvement over the original, it is still overly simplistic, vague, and open-ended. It needs further refining, but how should we proceed? This is where AI can serve as a valuable tool. Let’s start by creating a well-crafted prompt. As noted in our introduction to AI, and in our search strategies section, an effective prompt should be specific, avoid ambiguity, provide relevant background and context, articulate a clear objective, and establish defined boundaries.

Here’s a sample prompt for CoPilot AI that includes a question which addresses all the above criteria:

Before we input this prompt into CoPilot, remember that you’ll also need to craft a thesis statement that directly responds to your question. As you use CoPilot to help refine your research question, consider how your question might shape and support a compelling argument. We’ll delve deeper into thesis statements later in the course. For now, let’s see how CoPilot can help improve our proposed research question.

CoPilot yields several thematic options to choose from:

The option under Relational Identity is close to what Jada was working toward, but it could use more refining, so let’s ask CoPilot to add some tension and create a question that challenges conventional readings and interpretations. Here’s how it responded:

Note that CoPilot shows its strategy for refining the question before composing a new question with more nuance and provocation and breaking down why it works. These are all good insights into strategies for crafting and evaluating effective research questions.

The notion of provocation is interesting in this context. As you develop a question that will be provocative to readers, think also about what provokes you and use that to refine your question. For example, I’m interested in the idea that jazz is both expressive and elusive as noted in the “Why This Works” section. How can we integrate that idea into what we’ve started. Let’s see what CoPilot does.

New prompt:

In the “Why This Works” section, you noted that jazz is both expressive and elusive. How does that work within the context of the story?

Response:

This response provides something close to a workable question, but we want to add one more prompt to make it more provocative in a way that adds to the scholarly conversation by sparking further discussion and interpretations. Here’s the new version:

A note about connecting your thesis statement to your research question: you can continue using Copilot to transition directly from a research question to a thesis statement. However, we recommend using this research question to guide your searching and deriving your thesis from your results. You can then use CoPilot to test the strength of that statement, a process we will cover in Chapter 11 under Positing a Thesis Statement.

Exercises

Assignment 1: Using AI to Avoid Common Pitfalls with Research Questions

AI can help us avoid common pitfalls in our research questions:

• Too broad or vague
• Too narrow or specific
• Too speculative / not researchable
• Lacks a clear focus
• Can be answered with a simple yes/no
• Too leading/biased
• Too anecdotal or rooted in personal opinion

Also remember that your thesis statement (which we go over in a later chapter) will ultimately need to connect to your research question. AI can also help to assess how well we have made that connection once we have put the two together.

Activity: Draft a research question and run it through CoPilot, asking it to assess your question in terms of the criteria listed above.

Assignment 2: Using AI to Strengthen Your Research Question

Keeping in mind the advice about research questions we gained from Bruce Janz, let’s look at ways to use AI to strengthen our research question.

• Does your question take a risk?
• Is there a “that’s weird” aspect of your question?
• Are there secondary or tertiary questions that could be asked?
• Are there presumptions at work in your question that need to be addressed?

Put your draft research question into CoPilot and ask it to give you feedback based on the criteria in the list above.

For reflection: Did this activity change your perspective on your research and/or on your personal assumptions about the world? If so, explain how.