Thinking and Writing Analytically

One of the most important differences between high school and university writing is that university writing requires analytical thinking. Analytical thinking is a part of critical thinking–it is a systematic way of breaking down information or data to understand, evaluate, or solve problems. When you think analytically, you look for patterns and relationships and posit explanations about those relationships to discover something meaningful about the world. At university, you might be required to analyze a text, cultural artifacts like paintings or posters, or data collected by you or others.

If your instructor asks you to write an analytical text, you must think deeply about the phenomenon you have been asked to analyze. Not all of this analytical thinking will make it into your final draft, and often, the drafting process itself is part of understanding the phenomenon you are analyzing. You must spend time with this process–good analytical thinking requires several steps and some hard mental work.

Think of the process of analytical thinking as building depth in your thinking. Typically, you will move through three stages as you analyze a phenomenon. Each stage adds layers of depth and complexity to your thinking, which is one of the objectives of a university education. If your analytical writing stays at the lower layers of depth, you will not likely receive a good grade. For instance, if you write a lab report that only describes your data but doesn’t offer an analysis of that data and explore the implications of your findings, you will likely not get a good grade. The same is true in humanities. If you describe what happens in a text but don’t analyze patterns of meaning and the implications of this meaning, you likely won’t get a good grade on this paper either.

Step 1: Describe the Phenomenon

When you describe something, you must identify its parts and distinguish between the important and irrelevant parts.

Begin by gathering the materials or data you have been asked to analyze. Understand the context of these materials. For instance, if you have been asked to analyze a climate data set, find out where it originated. If you have been asked to analyze a text, consider who wrote the text and what type of authority or agenda they might bring to the text. This might be relevant to your later analysis.

Now, consider the material, text, or data you are analyzing. Jot down your initial thoughts and gut feelings. What surprised you? What bored you? What grabbed your attention? Although these gut feelings won’t make it into your analysis, they can be meaningful. These gut feelings can be the seeds of analysis that you don’t quite yet have the words for. But be careful here: gut feelings can also be the products of cognitive biases, which are common faulty thinking patterns. You can never stop at gut feelings if you would like to develop a strong analysis.

When analyzing a text, consider the central themes and key points. This part of analytical thinking is similar to summarizing, which we discuss in the chapter “Summarizing Sources.” In addition, consider how the information is organized in the text. Note how language is used. Note vocabulary, tone, and style.

If you have more than one example of the text or material you are analyzing, list similarities and differences. Consider creating a table that summarizes these elements.

For example, say your instructor asked you to analyze the differences and similarities between a human-written and AI-generated summary of a literary text. Create a table similar to Table 1 to organize your observations.

Table 1.

Table comparing a human-written and AI-generated summary with different criteria

Remember to keep examples of these observations from your text or data. You will need examples to support any of the claims you make in your analysis.

Step 2: Analyze the Patterns

You need to look for patterns in your observations to move from description to analysis.

Once you have a list of observations about your materials, text, or data, you can also ask yourself these questions:

• What elements go together?
• What elements are opposed to each other?
• What doesn’t fit at all? (Rosenwasser & Stephen, 2012)

Let’s continue with the example we presented in the previous step. Review the key observations of the differences between the human-written and AI-generated summaries in Table 1. Now, let’s answer the analysis prompts based on our observations. Table 2 provides some example answers to these prompts.

The next step in the analysis process is to develop hypotheses to explain what you have uncovered in your analysis. Ask yourself “how” and “why” questions to generate a hypothesis.

Here are some questions you could ask yourself about our example observations:

• Why does the human-written summary focus on the themes and characters of the original?
• Why does the AI-generated summary focus on plot points and a factual approach to the original text?
• How does the human writer choose which elements to include in the summary?
• How does the AI model choose which elements to include in the summary?
• How does the human-written summary balance analysis and empathy?
• Why does the AI-generated summary miss a key plot point?
• Why does the AI-generated summary have an upbeat tone?
• Why does the AI-generated summary use bullet points?

At this stage, you should generate as many questions as possible about the observed patterns. Play with potential hypotheses or claims about your observations. For instance, you could consider these potential interpretations of our example based on our observations:

• The human reader and writer can interpret the emotional nuances of the literary source text and summarize these aspects, whereas the LLM or AI model cannot. This may be due to biases in the LLM’s training model, which was primarily comprised of academic research texts that focus more on logic than emotion.
• The human writer can balance an analytical tone with empathy, but the AI model uses a positive tone, which is inappropriate given the subject matter of the literary source text. This reflects the AI-generated summary’s overall lack of emotional nuance.
• The AI model is not sensitive to differences between summaries suitable for public audiences and those written for academic purposes. This may be due to the nature of the prompts used to generate the AI summary.

Which hypothesis is the strongest? For which hypothesis do we have enough evidence? You must play around to see which angle will be the strongest for your paper. This part of the analytical process takes trial and error.

Your professor may want you to connect your hypothesis or explanation to other research. Review your assignment instructions or ask your instructor if this is a requirement.

Step 3: Explore the Implications

When you have developed a viable hypothesis with solid evidence, consider the big picture. Why does your hypothesis matter? So what? What are the implications? To answer these questions, you must identify your research's significance or ramifications. Answering these questions well is one of the most important parts of academic writing at every level, from undergraduate writing to the expert writing of your professors. They are so important that we often end our research papers by exploring these questions.

When your professors write research papers, they connect their analysis to the related conversation in their fields. For example, if I were to research AI-generated summaries of literary texts and find that these summaries consistently did not reflect the emotional nuances of the source text, I would connect my findings to other studies that looked at similar types of AI-generated writing. I would describe the significance of my findings for this larger conversation about AI-generated writing.

As students, you likely don’t know or understand the academic conversations that are taking place in your disciplines yet. You simply haven’t had the same time to read the research in these areas as your professors. However, you do know the conversations and the learning you have been doing in your course, and this is generally a good place to find the implications of your analysis. Review the learning objectives of your course to see if you can find a connection there. Review the key concepts and main themes of your readings and class activities. How does your analysis relate to these ideas?

For instance, if you were in an English literature class, you could explore the potential impact of your observation that the AI-generated summaries did not reflect the same emotional nuance as the human-written ones. For instance, you could speculate about what this meant for future learning. What if students rely on plot-based AI summaries that do not explore the themes and emotions of the literary source text? You could speculate that students may lose the ability to discern these emotional nuances in literary works, representing a significant learning loss. You have to walk a fine line here: You don’t want to overstate the implications, but you want to suggest why your analysis might matter in the big picture.

Remember that analytic thinking takes time–it is the hard work of university learning. This process takes practice, and your skills will develop throughout your university degree.

References

Rosenwasser, D., & Stephen, J. (2012). Writing analytically (6th ed). Wadsworth, Cengage Learning.

Attributions

"Thinking and Writing Analytically" by Nancy Bray, Introduction to Academic Writing, University of Alberta, is licensed under CC BY-NC 4.0