More than just Methods

When new/inexperienced researchers start to do research, they jump right into research methods. While important, focusing directly on methods is a great way to make poor study design predictions. Methods alone don’t guarantee that you’ll conduct fruitful research.

Different research methods are effective at helping you collect different and important kinds of data. But the reasoning or logic behind each method is not commonly thought about. To explain this, let's use the diagram below.

Whether you know it or not, you automatically operate using the underlying reasoning that fuels a research method. If you jump straight into methods when designing a study, you miss out on critical thinking about what those methods are designed to do. Once you understand that every method falls into a larger category of knowledge and learning about the world, you can recognize how and when to use any research method.

Setting aside ontology and epistemology (see Handbook 2 and Handbook 3 in this collection for more on qualitative and quantitative research philosophies), which can be complex to digest, let’s focus on methodology and methods. All methods use either inductive or deductive reasoning.

A quick note: Common diagrams for inductive and deductive reasoning use different triangles. However, a quick search online shows how different sources flip the triangles but use them for the same type of reasoning (aka an upside triangle used to explain inductive and deductive reasoning). The diagrams below avoid triangles and focus on the number of cases, situations, possibilities, elements, or people within any single ring. The larger the circle, the more elements, instances, or ideas are possible or are being referred to.

Let’s start with inductive reasoning.

‍

Inductive Reasoning

Qualitative methodologies use inductive reasoning. You use specific observations (such as individual interviews, contextual visits, or diary studies) to make general claims or theories about other, unobserved situations.

For example, you might interview ten people to understand their motivations when booking an international flight to make general claims or assumptions about all fruits which might book a similar flight. This might sound abstract so let’s visualize this.

For inductive reasoning, imagine a small circle that contains all the qualitative data you’ve collected for a study. These are the quotes, pictures, videos, and behaviors you’ve recorded during your qualitative study. The circle is small because you can only collect so much data in your limited time.

Next, imagine a bigger circle surrounding the first one, but this time it’s labeled as “qualitative codes.” This circle represents all the way you’ve assigned meaning and descriptions to your raw qualitative data. The next circle is how you’ve grouped or clustered those codes (making qualitative categories).

The final circle represents all your qualitative themes or the underlying meaning or pattern you’ve found in your data. The last circle has dashed lines because themes connect your observed data, codes, and categories to the theoretical world (jump to this Handbook for for more on qualitative codes, categories, and themes) which is where the “truth” is.

With inductive reasoning, each circle gets bigger and bigger. When the circle gets bigger, it’s slowly increasing the number of unobserved situations or cases your research can apply to. Your codes might be specific to your participants, but your themes should hold for the larger population. You’re moving from specific data to more general ideas. You can read more about inductive reasoning here.

When forming themes and theories out of qualitative research, you might want to test them using a deductive reasoning approach.

‍

Deductive Reasoning

On the other side, deductive reasoning fuels quantitative methodologies. You start with a general claim or theory and use specific observations or data to validate if that theory holds any weight. You believe that new customers struggle with parking curbside pickup at the local grocery store, so you choose to run a survey to validate this general idea.

In the diagram, you’re moving from the abstract, wide world of theory to a decision about how “true” that theory is. In the diagram, theory exists outside of the core reasoning circles.

To test a theory, you write a hypothesis. This is an initial assumption or belief that you think is true. You use the hypothesis to plan a study to collect specific quantitative data. You’ll only create relevant and practical hypotheses to study, limiting the size of the circle. Jump to Collection 3, Handbook 3, Topic 4 for more on research hypotheses, or review the guide below.

Guide 06: A Research Hypothesis Checklist

Next, you’ll collect quantitative data to test your hypotheses. This circle is even smaller because you need specific data to test your specific hypotheses. You then could use a statistical method or significance test to analyze that data. Each method or test will produce a limited range of results, making the next circle even smaller (Jump to this topic for more on using significance tests).

Finally, at the very center is the decision you make after collecting, analyzing, and interpreting your quantitative data. It’s essentially a point because your goal is to either say “yes, this evidence supports the initial hypotheses,” or “no, the evidence contradicts the hypothesis.”  You can read more about deductive reasoning here.

There’s one other scientific reasoning approach you could take in your research studies, and it’s a combination of inductive and deductive reasoning.

‍

Abductive Reasoning

In practice, you’ll probably take an abductive reasoning approach in many of your studies. For example, mixed method study designs take an abductive reasoning approach when addressing research questions.

With abductive reasoning, your thinking oscillates between inductive and deductive reasoning. It’s a pragmatic or practical approach because your results are viewed as tentative or the best possible results based on limited data. You never really have all the data you need, nor is it always possible to verify how "correct" your findings are.

You, your stakeholders, and the entire business has gaps and contradictions in its understanding about the the product, the people that use it, and the general user experience. Everyone is already taking an abductive reasoning approach because everyone is already using incomplete data to propose the best, current, and tentative explanation for what's real about the world around them. You update your understanding as needed if newer, better data comes along. You can read more about abductive reasoning on this page.

When designing a study, be mindful of what type of reasoning you’re using. It’ll help you collect the most factual, relevant data. But what do you do if you need to study something abstract – like satisfaction, trustworthiness, or inclusion?