If you are a PhD student or researcher in Information Systems, you have heard this phrase countless times. It is the gatekeeper to top journals and the currency of academic contribution. Yet despite how often it is invoked, there is surprisingly little agreement on what theory actually is.
- Is a taxonomy a theory
- Is a predictive algorithm a theory
- What about a design framework for a new software application
In her seminal 2006 paper The Nature of Theory in Information Systems, Shirley Gregor addressed this confusion directly. She argued that the field spends too much time debating how to build theory, that is epistemology, and not enough time defining what exactly we are building, that is ontology.
To make sense of this, it helps to view theory construction through the lens of architecture. Just as an architect relies on different types of plans, from site surveys to detailed schematics, researchers need different types of theory to build knowledge.
Gregor’s framework gives us five distinct blueprints for Information Systems theory.
The Raw Materials of Theory
Before drawing any blueprint, an architect must understand their materials. In research, our raw materials are:
Causality
The relationship between cause and event. In Information Systems, this often takes a teleological form, meaning actions taken by human agents to achieve goals.
Explanation
The why. Explanation is a communicative act intended to create understanding in the reader.
Prediction
The what will be. Prediction allows theories to be tested against future outcomes.
Gregor combines these elements to classify theory into five types.
Blueprint I
The Site Plan
Theory for Analysis
Core purpose
Says what is.
Trying to build a house without surveying the land would be reckless. You need to know the boundaries, terrain, and conditions.
Theory for Analysis does exactly this. It classifies dimensions, characteristics, and categories. It organizes reality into a structured map. It does not explain why things happen or predict what will happen next. It simply defines the landscape.
Key components
Taxonomies, frameworks, schemas, and ontologies.
Why it matters
Data alone are not theory. Analytic theory turns raw data into meaningful constructs. Without it, the field lacks the vocabulary needed for more advanced research.
Classic example
Gorry and Scott Morton’s framework for management information systems. It did not predict outcomes, but it gave researchers a shared language.
Blueprint II
The Narrative Rendering
Theory for Explanation
Core purpose
Says what is, how, why, when, and where.
An architectural rendering helps you imagine what it feels like to inhabit a space. It tells a story.
Theory for Explanation focuses on deep understanding. It explains how and why phenomena occur, often grounded in a particular worldview. These theories act as sensitizing devices. They do not always predict outcomes, but they profoundly shape how we see and interpret reality.
Key components
Rich causal explanations, often centered on human action and intention.
Why it matters
Sometimes prediction is impossible or irrelevant. Understanding alone is the contribution, especially in complex contexts like large-scale IT failures.
Classic example
Orlikowski’s Structurational Model of Technology, which explains the reciprocal relationship between human action and technological structures.
Blueprint III
The Performance Model
Theory for Prediction
Core purpose
Says what is and what will be.
Sometimes you do not need to know why the wind blows. You just need to know whether the building will withstand it.
Theory for Prediction focuses on forecasting outcomes from a set of variables, even when the underlying causal mechanisms remain opaque. Explanatory depth is secondary to predictive accuracy.
Key components
Testable propositions and measurable variables.
Why it matters
This type of theory offers high precision but limited explanation. It is extremely useful for decision making, such as forecasting system loads or market behavior.
Classic example
Moore’s Law, which accurately predicted growth in computing power without explaining the underlying technological or economic mechanisms.
Blueprint IV
The Detailed Schematic
Theory for Explanation and Prediction
Core purpose
Says what is, how, why, when, where, and what will be.
This is often considered the gold standard in top journals. Theory for Explanation and Prediction combines deep causal understanding with empirical testing.
It represents a mature state of knowledge where researchers both understand the mechanisms at work and can model them precisely enough to generate predictions.
Key components
Clearly defined constructs, scope conditions, causal explanations, and testable hypotheses.
Why it matters
This theory satisfies the scientific ideal of understanding and verification.
Classic example
The Technology Acceptance Model, which explains why users adopt technology and predicts usage behavior.
Blueprint V
The Construction Guide
Theory for Design and Action
Core purpose
Says how to do something.
Information Systems is not only a social science. It is also a design science. We build artifacts.
Theory for Design and Action provides prescriptive guidance. It specifies principles, methods, and design requirements for building systems that achieve particular goals. The focus is not on explaining the world, but on improving it.
Key components
Prescriptive statements linking design features to desired outcomes.
Why it matters
This legitimizes building as a theoretical contribution. It moves beyond implementation into theory about the artifact itself.
Classic example
Markus and colleagues’ design theory for knowledge management systems, which offered a blueprint for systems that support emergent knowledge processes.
Building a Robust Discipline
When Gregor reviewed fifty articles in top Information Systems journals, she found that Theory for Explanation and Prediction dominated the field. While powerful, a healthy discipline requires diversity.
We need site plans to map new frontiers such as quantum computing.
We need narrative renderings to understand issues like ethical AI adoption.
And we need construction guides to design the next generation of compassionate digital systems.
As researchers, we are architects of knowledge.
The real question is this:
Which blueprint are you drawing today?
Leave a Reply