%202.png)
Theory Building and Conceptual Contribution
Turn a conceptual model into a real theory: defined constructs, specified relationships, an actual mechanism, and boundaries that make it testable.
Free. Works on Claude, ChatGPT, Gemini or any assistant that accepts a skill file.
What this skill does
The method, encoded.
A great deal of work described as theoretical is not. Boxes joined by unlabelled arrows are a picture of a hypothesis space. A list of factors that matter is a checklist. A framework that organises a literature is a filing system. All three are useful, and none of them explains anything, because explanation means saying why one thing produces another and under what conditions it stops.
This skill supplies the elements that make a theory a theory. It forces construct definitions that exclude specific things and are differentiated from the neighbouring constructs they keep being confused with. It replaces arrows with specifications: direction and what establishes it, functional form, expected sign, mediation or conditionality, and timescale. It treats mechanism as the heart of the contribution, written as a sequence of things that happen, tested by whether it generates any prediction a bare association would not. It separates propositions from hypotheses and makes the derivation between them visible. It sets boundary conditions, on the principle that a theory claiming to apply everywhere explains nothing, and it asks the author to name a case their theory should fail to explain.
It then runs the test that catches the commonest false contribution: translate the theory into existing vocabulary and see what survives. Whatever is left is the contribution, and it is usually smaller, more precise and far more defensible than what was claimed.
Best used for
- Turning a boxes-and-arrows model into a specified theory
- Defining a construct precisely enough to be measured and distinguished
- Establishing boundary conditions for an over-general claim
- Building theory from qualitative data with a named approach
- Testing whether a claimed theoretical contribution is a relabelling
Typical inputs
What you give it.
A stated phenomenon or puzzle that existing theory does not explain, Existing theory in the space, read in full, A current conceptual diagram or model, Qualitative data or an intended evidence base, Existing measures of the constructs, Competing explanations for the same phenomenon
Typical outputs
What you get back.
Explanandum statement and account of why existing explanations fail, Construct definitions with exclusion sets and differentiation from neighbours, Relationship specification table with direction, form, sign and timescale, Stepwise mechanism statement with interruption points and secondary predictions, Numbered propositions and derived hypotheses with explicit mapping, Boundary conditions on population, setting, time and assumptions, Contribution classification with the evidence its burden requires, Relabelling test result and differential predictions against rival theories, Refutation conditions per proposition
Method coverage
What the skill works through.
- What makes something a theory rather than a framework or a list
- Defining a construct so that it excludes things
- Specifying relationships instead of drawing arrows
- Mechanism: the heart of a theoretical contribution
- Propositions and hypotheses, and why they are not the same
- Boundary conditions and why universality is a weakness
- Extending, refining, challenging or building, and the burden of each
- The relabelling test
- Parsimony against completeness
- Testability and stating what would refute you
- Building theory from qualitative data
- Academic integrity and AI use
Download
Free skill. One file.
Enter your email once. Every skill you download after that takes a single click.
How to install
Add the skill file and the five kernel protocols to a Claude Project, a ChatGPT Project, a Gemini Gem, or paste them at the top of any assistant conversation. Then give it your real research material, not a description of it.
Download skillQuestions
Common questions.
What is the difference between a theory and a conceptual framework?
A conceptual framework organises constructs and shows how they are expected to relate. A theory adds the mechanism, which explains why the relationships exist, and the boundary conditions, which say where the explanation holds. A framework helps you design a study; a theory makes predictions that could turn out to be wrong.
What is the difference between a proposition and a hypothesis?
A proposition states a relationship between constructs at the conceptual level and is not directly testable. A hypothesis states a relationship between operationalised variables in a specified population and is testable. One proposition should generate several hypotheses; if it generates only one, it is probably a hypothesis in formal dress.
Why do boundary conditions matter?
Because a theory that claims to apply everywhere explains nothing and cannot be refuted. Specifying the population, the setting, the timescale and the assumptions the mechanism depends on makes the theory more useful and more credible, and naming a case the theory should fail to explain is one of the strongest signals of a well-developed contribution.
How do I know if my theoretical contribution is real or just new terminology?
Rewrite your theory using only vocabulary from the existing literature, substituting established terms wherever a plausible equivalent exists. Whatever survives the translation is your contribution. Then ask what your theory predicts that the two nearest existing theories do not. If there is no differential prediction, the contribution is presentational rather than theoretical.
Can you build theory from qualitative data?
Yes, and there are several established approaches that differ in ways that matter: classic grounded theory, its systematised variant, constructivist grounded theory, multiple-case theory building, and abductive or interpretive theorising. They disagree about whether categories emerge from data or are constructed, so the choice commits you to a position and must be stated. Whichever is used, the trail from data to construct to relationship must be inspectable, including the negative cases that changed the account.
What counts as a mechanism?
A sequence of things that happen, each independently plausible and in principle observable, connecting the antecedent to the outcome. The test is whether it generates a prediction that a bare statement of association would not. A mechanism that predicts exactly what "A is associated with B" predicts is doing no work.
The skill chain
Works well with.
Research where people already are.
Analyse it where you already work.
Yazi helps researchers conduct surveys, AI interviews and longitudinal research directly through WhatsApp.
