The paper titled “Principles for Conducting Critical Realist Case Study Research in Information Systems” by Donald Wynn Jr. and Clay K. Williams introduces a powerful yet complex way to do research in Information Systems (IS) using Critical Realism (CR). This article expands and simplifies their key ideas and contributions for anyone curious about applying this approach to real-world research.
What is Critical Realism (CR)?
I would love to break this definition down so you can understand it deeply – this is not another research paper so there’s really no need to use ambiguous words. Critical realism is a special way of doing research that says: There’s a real world out there – which we can’t fully grasp or even measure. What we think we see or observe when we do a positive research or interpretive research (numbers, interviews or behaviours) is just a tip of the iceberg – to truly understand what is going on we need to really dig deeper and look for hidden causes, systems, and reasons behind what we see. Let me give you a simple analogy. If a plant is not growing well, you don’t just look at the leaves, instead you check the roots, the soil, the water and the sunlight. Critical Realism helps researchers to conduct these kinds of deeper thinking in their studies. According to Critical Realism, things that we can’t see cause the things we can see
In information Systems CR is great for studying complex things like how people use technology at work and why certain digital systems succeed or fail. It’s not just about what is happening – it is about why it is happening. To truly understand why things happen, we need to explore the deeper layers of reality, like systems, forces, and hidden mechanisms.
Donald Wynn Jr – who is currently the Sherman-Standard Register Associate Professor of MIS at University of Dayton – (I actually have the habit of checking out the authors of the papers I read online – most times I add them on LinkedIn to keep tabs on them) argued that critical realism (CR) is a promising alternative paradigm to the dominant positivist and interpretivist approaches in information systems (IS) research. This is quite interesting because earlier before now I always believed that positivist and interpretivist were the only paradigms that existed.
With CR we get to see that there’s a reality that exists independent of our knowledge, but our understanding of it is inevitably theory-laden and fallible. Under CR, explaining a phenomenon means identifying the underlying structural entities and contextual conditions that interact to generate observable events
In Critical Realism, causality (what causes what) isn’t just about seeing that two things happen together (like “when X happens, Y also happens”). Instead, CR says: Let’s find out why Y happens. It’s not enough to say X and Y are linked—we want to understand the hidden forces or mechanisms that cause Y to happen, especially in real-life situations where many things are happening at once Example: Imagine people stop using a new app at work. A simple study might say,“People stopped using the app because it was confusing.”
But Critical Realism would ask: “What deeper things caused that confusion? Was it poor training? Was the app badly designed? Did people feel unsafe using it?” So instead of just saying “A caused B,” Critical Realism digs into the underlying reasons—like peeling an onion to see what’s underneath.
What did Wynn and Williams do?
They helped researchers who want to use Critical Realism (CR) in case studies—especially in Information Systems (IS) research—by creating five practical rules to follow – which I think was a genius idea because before this, CR was mostly a big idea (a philosophy). But it wasn’t clear how to actually use it when doing real-world research.
So, Wynn and Williams said: “Let’s take the deep ideas from CR—like the belief that reality has layers and isn’t always visible—and turn them into clear steps researchers can use in case studies.”
What are the key ideas behind their rules? Reality exists even if we don’t see it directly. What we observe (like people using tech) is just the surface layer. Real understanding means looking for the hidden causes and structures underneath. Knowledge is always partial and shaped by how we study it.
What did they actually do?
They explained each of their five principles using real examples from IS case studies that already used CR, whether the authors said so or not. By doing this, they made CR easy to use for researchers who want to move beyond just describing things or making predictions—and instead want to build theories that explain why and how things happen.
Why does this even matter?
Explains CR philosophy in a way that makes sense for IS research. Gives five simple principles that show how to actually do CR-style case studies. Shows real examples so it’s not just theory—it’s practical. Wynn and Williams gave IS researchers a toolkit to do deeper, more meaningful research using Critical Realism. Their work is now widely used by scholars who want to go beyond the surface and explain complex tech and organizational issues.
How Critical Realism (CR) connects to ontology and epistemology:
First: Let’s understand the basics – which are:
- Ontology = What is real? What exists?
- Epistemology = How do we know what we know?
Now: How does Critical Realism relate to these?
✅Ontology in Critical Realism
CR believes that:
There is a real world out there, whether or not we can see or measure it.
This means: Reality exists independently of us (even if we don’t fully understand it). This reality has layers:
- What we can observe (events, actions)
- What causes those things (underlying structures or mechanisms)
For Example: You see a worker resisting a new technology. That’s the event. But CR says: Let’s look deeper—maybe it’s due to hidden structures like poor training, fear of job loss, or organizational politics.
✅Epistemology in Critical Realism
CR believes:
Our knowledge of reality is limited and imperfect, and shaped by our tools, theories, and perspective. This means: We can’t directly know everything about reality, but we can get closer to the truth by using good research methods—especially ones that dig deep (like case studies, interviews, theory building). Knowledge is always mediated (influenced by context and interpretation).
For example: You do interviews to understand why a system failed. People give different views. CR says that’s okay—put those views together to infer the hidden causes behind the failure. CR says, The world is layered:
- The real layer contains all the things that exist, whether we see them or not.
- The actual layer includes the things that actually happen.
- The empirical layer is what we experience or observe.
There are structures and systems—like rules, cultures, or processes—that have power to influence outcomes, even if we can’t directly see them. Our knowledge is always partial and filtered by our perspectives, tools, and experiences. Research should aim to explain things by uncovering what can’t be directly observed—not just report what we see or guess what might happen.
📊 Why Case Studies Work Well with Critical Realism (CR)
Case studies are like deep dives into real situations. CR is about understanding what’s really going on beneath the surface. Together, they make a powerful team. Here’s why:
Case studies let you Look at real events over time, not just snapshots. Dive into the messy details of real life and Explore how people, technology, and the environment interact
🎯 What CR Case Studies Try to Do:
Instead of saying: “X causes Y all the time (because of stats). They say: “In this specific case, here’s what caused Y—and this helps us build a theory that explains how things work in complex situations.” Let’s take a real world example – Imagine a company failed to adopt a new system.
- A simple study might say: “It failed because the system was too complex.”
- A CR case study would ask:
What was happening in the company culture?
How did leadership behave?
Were people trained?
Were there hidden fears or power struggles?
By exploring all layers, it explains not just what happened, but why—and what that means for other companies too. CR + Case Studies = Rich, deep, real-world understanding that helps build better theories, not just stats.
Methodological Principles for Critical Realist Case Studies
Wynn and Williams (2012) propose five methodological principles for doing critical realist case study research in IS. Table 2 of their paper provides a summary (the principles and their rationale), and each principle is derived from core CR tenets
Below is a structured outline of each principle along with its meaning:
Explication of Events:
Clearly identify and describe what actually happened in the case. These are the key events or outcomes the researcher wants to explain—like a system failure, a successful change, resistance to technology, or a shift in behavior. Before you can explain why something happened, you need to be really clear about what happened.
Wynn and Williams say:
- Start with the real-life events—what people actually experienced.
- Use interviews, documents, or observations to describe the event.
This makes sure the research stays grounded in reality, not just in abstract theory.
Example: Let’s say a new software was introduced, but employees didn’t use it. Explication of Events would involve:
- Describing what happened: “Employees rejected the new software.”
- Using real data: “Based on interviews, logs, and reports…”
- Asking: “This happened—now let’s figure out why.”
Explication of Structure and Context:
Once you’ve explained what happened (the event), CR says the next step is to look deeper and ask: What structures and context made this event possible? This means identifying – structures: Things inside the organization that shape behavior. Example: Hierarchy, decision-making processes, reward systems
- Technology itself
Example: Features of a tool or system (like limited access or confusing interface)
- Social norms or rules
Example: Expectations about who should use tech, or fear of change
🔸 Contextual Conditions: External or environmental factors
- Industry pressure
- Company culture
- Government policy
- Historical background
Why is this important?
In Critical Realism, structures and context don’t directly cause things to happen, but they create the conditions that make events possible. They act like affordances (opportunities) or constraints (limits).
Example: Let’s say a digital transformation fails in a company.CR wouldn’t stop at: “Employees resisted the change.” It would dig deeper and ask:
- Was there a rigid hierarchy that blocked feedback?
- Did the company culture value tradition over innovation?
- Was the software hard to use due to poor design?
- Were there industry regulations that created fear or confusion?
These structures and contexts shape what can and cannot happen.
⚙️ CR’s Ontology in Action:
CR sees reality as layered.
- What we observe (the event) is just one layer.
- Structures and context are deeper layers that have causal power—they shape outcomes, even if we can’t see them directly.
To explain why something happened, you need to analyze the hidden systems and the broader environment. These background factors make certain events more likely or less likely to occur.
Retroduction:
Retroduction is a type of reasoning where you ask: “What hidden mechanism or power must exist to explain these observed events?” Unlike simple cause-and-effect or statistical association, retroduction digs into why something happened by inferring underlying factors that aren’t immediately visible.
How Does It Work?
- Start with Observed Events:
Look at the events or outcomes that have been clearly identified in your case study. - Identify the Structures and Context:
Understand the broader setting—what organizational structures, technologies, social norms, or environmental conditions are in play? - Ask “What Must Be True?”:
Based on what you see, ask:- “What hidden mechanism or force must be present for these events to have occurred?”
- For example, if a digital tool isn’t being used, what underlying factors (like a lack of training, organizational resistance, or poor design) must exist for this to happen?
- “What hidden mechanism or force must be present for these events to have occurred?”
- Propose Plausible Causal Mechanisms:
Use creativity and your theoretical knowledge to formulate hypotheses about these hidden drivers. This might involve suggesting that:- A rigid hierarchy limits the free flow of information.
- Poor system design creates obstacles that discourage use.
- Cultural values inhibit innovation.
- Provide Logical and Analytical Support:
It’s not enough to just guess—your proposed mechanisms should be backed by logical reasoning and, where possible, evidence from your case. Explain how each mechanism could realistically generate the outcomes observed.
Why is Retroduction Important?
- Goes Beyond the Surface:
It helps you move past simple descriptions of what happened and instead build a theory about why it happened. - Builds Causal Explanations:
By inferring hidden mechanisms, you create explanations that are more robust and potentially applicable to other similar situations—even if they are specific to the case studied. - Encourages Creativity and Insight:
Retroduction is both a logical and creative process. It challenges researchers to think deeply and innovatively about the forces shaping the observed events.
Empirical Corroboration:
Once you’ve come up with ideas about the hidden causes (mechanisms) that might explain what happened in your case, you need to go back to your data and ask:
“Is there any evidence that supports this explanation?”
Since these hidden mechanisms can’t be seen directly (you can’t observe “fear of job loss” or “lack of trust” like you can see a login screen), you look for clues or patterns that those things are likely happening.
What does this look like in practice?
- You propose a mechanism:
“I think employees avoided using the new system because they feared being monitored.” - You check the data:
Are there interview quotes, survey results, or behavioral patterns that suggest people were worried about being watched? - You look for consequences:
If that fear is real, you might expect:- Low system usage,
- People sticking to old methods,
- Comments like “I don’t want them tracking me.”
- Compare with other explanations:
Maybe someone says it’s not fear—it’s just that the system is slow.
You ask: “Which explanation fits the data better?”
Triangulation and Multiple Methods:
Use multiple data sources, types, or methods to examine the phenomenon from different angles, thereby enhancing validity and mitigating bias. Reality is complex, and no single method gives the full picture. So, to understand what’s really going on, we need to look at the case from different angles—just like you’d check a sculpture from every side, not just one.
What does this mean for research?
It means using more than one source of information or method to study the same thing. For example:
- Different data types → Interviews, documents, system logs, surveys
- Different people → Managers, employees, tech support, users
- Different methods → Observing behavior, analyzing documents, conducting interviews
- Even different cases, if you’re comparing more than one situation
Why is this important?
- It helps avoid bias or one-sided conclusions.
- It makes your explanations stronger and more believable.
- It allows you to see both:
- Subjective experiences (what people feel or say)
- Objective facts (what actually happened, based on records or observations)
Example:
Let’s say you’re studying why a new IT system failed.
- You interview employees and they say they weren’t trained properly.
- You check training documents and see that training sessions were canceled.
- You observe system logs and notice people tried logging in but gave up quickly.
All these different sources point to the same mechanism: lack of proper training.
That gives you stronger evidence than just relying on interviews alone.
These five principles form an integrated approach: the researcher begins by defining what happened (events), looks at what might have caused it (structures, context, and inferred mechanisms), and then rigorously checks those causes against evidence, using multiple sources to ensure credibility. Wynn and Williams intended these principles to guide both conducting a CR-informed case study and evaluating its quality
.In their paper, each principle is illustrated with examples from prior IS case studies (which the authors reinterpret through a CR lens), demonstrating concretely how an investigator might implement the principle in practice
Leave a Reply