03 · What You Need to Know
Theory Extension Is About Improving Explanation
Theories are expected to develop
An established theory represents a scholarly explanation at a particular stage of development. Subsequent research may identify contexts in which its propositions behave differently, constructs that need greater precision, mechanisms that were underspecified, or empirical observations that the original explanation cannot adequately accommodate.
That is not necessarily evidence that the theory has failed completely. It may indicate that the explanation can be refined.
Fisher and Aguinis describe theory elaboration as an approach that begins with existing conceptual ideas or a preliminary model and develops new theoretical insights by contrasting, specifying, or restructuring constructs and relationships to account for empirical observations. Their framework illustrates an important point: theoretical advancement can occur between simply testing an unchanged theory and attempting to invent an entirely new one.
Applying a theory is different from extending it
Researchers sometimes describe a study as a theoretical extension merely because an established theory is being used with a new population, technology, country, or institution.
That is not automatically an extension.
If you apply the original constructs and relationships without changing the explanation, you are primarily applying or testing the theory in another setting. Such research can still be valuable, especially when the new context provides a meaningful test of generalizability or boundary conditions.
An extension makes a stronger contribution. Something about the theoretical explanation itself is being added, refined, distinguished, restructured, or bounded.
Theory application
Uses an existing theoretical explanation to understand or investigate a phenomenon without substantively changing that explanation.
Theory extension or elaboration
Develops the existing explanation by refining its constructs, relationships, mechanisms, sequence, scope, or boundary conditions.
Being precise about whether you test, apply, extend, or build theory prevents a modest contextual application from being presented as a larger theoretical contribution than it is.
Adding a new construct can extend a theory, but not automatically
Suppose an established theory proposes that A influences B. You believe C also matters, so you add C to your research model.
Have you extended the theory?
Perhaps, but the answer depends on what C contributes theoretically.
If C explains why A influences B, specifies when the relationship occurs, changes the meaning of an existing construct, or accounts for observations the theory could not previously explain, then the addition may contribute to theoretical elaboration.
If C is included only because previous empirical studies found that it predicts B, you may simply have expanded an empirical model.
The difference lies in explanatory reasoning. A variable becomes theoretically consequential when its inclusion changes what the theory allows us to understand.
You can refine the meaning of a construct
Theoretical development does not always require adding something.
A construct may be too broad, ambiguously defined, or treated as one dimension when empirical observations suggest meaningful distinctions. Fisher and Aguinis identify construct specification and construct splitting among the ways researchers can elaborate theory.
Imagine that an existing theory treats “institutional support” as a single construct. Further investigation might suggest that technical support and policy support operate through different mechanisms. Distinguishing them could improve the explanatory precision of the theory if the distinction is conceptually defensible and empirically meaningful.
Splitting constructs merely because a factor analysis produces several dimensions, however, is not enough by itself. The theoretical meaning of those distinctions still needs explanation.
You can specify a mechanism
Sometimes a theory predicts that two constructs are related but provides only a limited explanation of how that relationship occurs.
A researcher might propose an intervening mechanism that explains the process connecting them. In empirical models this is often represented through mediation, although statistical mediation and theoretical mechanism are not synonymous.
The theoretical contribution comes from explaining the process.
For example, suppose a theory predicts that institutional support increases instructors' technology use. A researcher might propose that support influences use partly because it increases perceived capability. The important theoretical question is why institutional support should alter perceived capability and why that change should affect behavior.
Simply inserting a mediator into a structural model does not answer those questions.
You can identify boundary conditions
A theory may explain a relationship well under some circumstances but poorly under others.
Identifying those circumstances can refine its scope.
A boundary condition specifies when, where, for whom, or under what circumstances a theoretical proposition is expected to hold. Empirically, such conditions are often investigated through moderators, subgroup differences, contexts, or temporal conditions, although boundary conditions are fundamentally theoretical rather than merely statistical.
For instance, a relationship between perceived usefulness and technology adoption might differ when technology use is mandatory rather than voluntary. If the original theory implicitly assumes volitional choice, studying mandatory use could expose an important boundary.
This is more theoretically informative than saying simply that “context matters.”
You can restructure relationships among existing constructs
Theory elaboration can also involve changing how existing constructs are connected.
A relationship assumed to be direct may operate through another process. A supposedly linear relationship may be nonlinear. Two constructs thought to operate independently may interact. Relationships may be reciprocal rather than unidirectional.
Fisher and Aguinis explicitly identify the structuring of specific, sequential, and recursive relationships as routes to theoretical elaboration.
These changes are consequential because they alter the explanatory architecture of the theory rather than merely expanding its list of variables.
You can clarify temporal sequence
Many theories imply processes that unfold over time even when empirical studies test them using cross-sectional data.
Research may reveal that the sequence of constructs matters. A causes B only after C occurs, or a relationship that appears reciprocal may actually unfold through alternating stages.
Clarifying temporal ordering can therefore constitute theoretical refinement.
It also creates methodological obligations. A proposed temporal process cannot be strongly evaluated using a design incapable of observing temporal change. Theory and method need to develop together.
Context can justify modification, but novelty of context is not enough
Researchers frequently propose adding a variable because their study occurs in a “unique context.”
Sometimes that is defensible. Context may expose assumptions that were invisible in the setting where the theory was developed. Cultural norms, regulatory structures, institutional arrangements, technological affordances, or resource constraints may alter how a theoretical mechanism operates.
But a new country, university, profession, or technology does not automatically require theoretical modification.
You need to explain why the contextual difference should alter a construct, relationship, mechanism, or boundary. Otherwise, the study may simply be an application of the existing theory in another setting.
Modification should begin with a limitation, not with available variables
A common workflow runs backward: researchers inspect their dataset, notice several variables that predict the outcome, and then add those variables to an established theory.
The resulting model may predict well, but prediction alone does not establish theoretical extension.
A stronger process begins by identifying an explanatory limitation. What does the theory currently fail to explain? Why might that limitation exist? What theoretical change would address it? What evidence would allow that proposed change to be evaluated?
Existing explanation Establish what the original theory actually claims.
Explanatory limitation Identify an unresolved observation, mechanism, boundary, construct, or relationship.
Proposed elaboration Explain how a modification addresses that limitation.
Empirical implication Specify what evidence should be observed if the proposed elaboration is useful.
Do not silently rewrite the original theory
If you modify a theory, readers should be able to distinguish the established theory from your contribution.
State which constructs and propositions come from the original formulation, which come from later theoretical development, and which modifications you are proposing.
This is particularly important when theories have already undergone multiple revisions. What researchers casually call “the original theory” may actually refer to a later version containing constructs or relationships absent from the foundational publication.
Watch Out
Do not redraw an established theory with your own constructs added and then present the resulting diagram under the original theory's name without qualification. Readers should be able to tell where the established theory ends and your proposed extension begins.
A modified theory requires evidence proportional to the claim
One study can provide evidence consistent with a proposed extension. That is not the same as permanently revising an established theory.
Theoretical claims generally become more credible through repeated investigation, conceptual scrutiny, comparison with alternatives, and evidence across relevant conditions.
A thesis or article might therefore state that findings “suggest a possible boundary condition,” “support the proposed extension,” or “indicate that the theoretical model may benefit from distinguishing these constructs.” Such language is often more defensible than announcing that the theory has now been definitively modified.
Theory is cumulative scholarship. A diagram does not become canonical merely because it survived one structural equation model.
