“Conceptual framework” is the section most students write twice: once as a list of definitions, and then again after a supervisor explains that a list of definitions is not a framework. This page skips the first version. It sets out what a conceptual framework has to contain, how it differs from a theoretical framework in practice rather than in principle, and four frameworks built out in full — the concepts, the diagram described in words, the propositions that link them, and the paragraph that defends each choice.
The frameworks below are shapes, not claims. They contain no real study and no proposition attributed to a real author. [cite] marks a source you must supply.
The difference that actually matters
The textbook distinction is that a theoretical framework applies an existing theory and a conceptual framework assembles one. True, and not useful at 11pm. Here is the operational version.
| Theoretical framework | Conceptual framework | |
|---|---|---|
| Where it comes from | One named theory, adopted whole | Concepts drawn from several literatures and assembled by you |
| What you must defend | Why this theory rather than a rival theory | Why these concepts, and why these links between them |
| How it fails | The theory does not fit the data you collected | A concept appears in the diagram and never again in the thesis |
| What it produces | Hypotheses derived from the theory’s propositions | Variables, relationships and the boundaries of the study |
| Test of adequacy | A reader of that theory recognises your use of it | Every arrow corresponds to something you will actually measure or observe |
The last row is the one to remember. A conceptual framework is adequate when every element of it reappears in your methods chapter. If a box has no measure and an arrow has no analysis, that box and that arrow are decoration, and an examiner will ask about exactly those. Where your study does adopt a named theory whole, the treatment is different; the site’s theoretical framework guide for social work works through that case.
The four parts every framework needs
- The concepts, each defined in one sentence, with a citation for the definition you adopted rather than one you wrote.
- The relationships, stated as propositions in plain sentences: A affects B, B is stronger when C is present.
- The diagram, which must be readable without the text and must contain nothing the propositions do not state.
- The boundary: what is deliberately outside the framework and why.
Write them in that order. Students who start with the diagram usually end up defending a picture rather than an argument.
Framework 1: nursing — discharge preparedness and follow-up
Concepts. Discharge preparedness (the patient’s self-assessed readiness to manage care at home [cite]); primary-care attachment (whether the patient has a regular provider to return to); symptom burden (count and severity of concurrent conditions); help-seeking (contact with any health service within 30 days).
Propositions. (1) Higher discharge preparedness is associated with earlier and more appropriate help-seeking. (2) The association is moderated by primary-care attachment: where there is no attached provider, preparedness has less to act on. (3) Symptom burden affects both preparedness and help-seeking directly, and is therefore a confounder rather than a mediator.
Diagram in words. Preparedness sits left, help-seeking right, with a direct arrow between them. Attachment sits above that arrow with a line meeting it at right angles — the moderator convention. Symptom burden sits below with arrows to both boxes.
Boundary. The framework excludes the quality of the discharge conversation itself, which would require observation the study cannot fund. It also excludes cost, which is measured in this system but not attributable at the patient level.
Why it holds up. Every box becomes a measure: a preparedness instrument, a yes/no registry field, a comorbidity count, and an administrative contact record. Proposition 3 does real work, because deciding that symptom burden is a confounder rather than a mediator determines whether you control for it or model a path through it — a decision your examiner will test.
Framework 2: education — assessment policy and classroom practice
Concepts. Policy prescription (what the ministry or board documents require); teacher interpretation (what a teacher understands the requirement to demand); competing accountability (the presence of an external examination weighting); enacted practice (the assessment activities actually used).
Propositions. (1) Policy prescription influences enacted practice only through teacher interpretation; there is no direct path. (2) Competing accountability shapes interpretation rather than practice directly. (3) Enacted practice feeds back into interpretation over time, so the relationship is recursive rather than linear.
Diagram in words. A left-to-right chain: prescription → interpretation → practice, with competing accountability entering the interpretation box from above, and a return arrow from practice back to interpretation.
Boundary. Student outcomes are outside the framework. This is a study of how a policy becomes a practice, not of whether the practice works.
Why it holds up. Proposition 1 is a genuine commitment: it says a purely quantitative design correlating policy exposure with practice would be mis-specified, which justifies a qualitative or mixed design. A framework that constrains your method is doing its job.
Framework 3: engineering — infrastructure renewal decisions
Concepts. Asset condition (measured or modelled state of a pipe segment); environmental loading (freeze-thaw cycles and soil-moisture variation over the observation window); break history (recorded failures per segment); renewal priority (the rank a segment receives in the municipal plan).
Propositions. (1) Break history is a function of asset condition and environmental loading jointly, not of condition alone. (2) Renewal priority is currently a function of break history and segment attributes only. (3) The unexplained variance in (2) is therefore partly attributable to the omission of environmental loading from (1).
Diagram in words. Condition and loading both feed break history; break history and attributes feed priority; a dashed arrow runs from loading directly to priority, labelled “omitted”.
Boundary. Budget constraints and political factors in the renewal decision are acknowledged and excluded; the study models the technical input to the decision, not the decision.
Why it holds up. The dashed “omitted” arrow is the thesis. A framework that visibly marks what current practice leaves out tells the reader what the contribution is before they reach the methods.
Framework 4: management — performance visibility and collaboration
Concepts. Indicator visibility (how prominently individual performance metrics are displayed); measured performance (the indicators themselves); discretionary collaboration (help given to colleagues that no indicator records); perceived fairness (whether staff regard the indicator set as a fair account of their work).
Propositions. (1) Indicator visibility raises measured performance. (2) Indicator visibility lowers discretionary collaboration. (3) Both effects are mediated by perceived fairness, which means the trade-off is not fixed and can in principle be managed.
Diagram in words. Visibility on the left with two arrows fanning right to the two outcomes, and perceived fairness sitting on both arrows as a mediator.
Boundary. Individual personality traits are excluded; the framework is organisational, and the study cannot administer personality instruments at the required scale.
Why it holds up. Proposition 3 converts a familiar observation into something actionable and testable, which is what separates a framework from a diagram of things you have read about.
The paragraph that defends the framework
Every framework needs one paragraph, usually at the end of the review chapter, that says why this one. The structure is fixed:
This framework draws its account of [concept 1] from [literature A]
[cite]and its account of [concept 2] from [literature B][cite]. The two are combined here because [the specific reason the phenomenon requires both]. An alternative would have been to adopt [named theory] whole; that was not done because [what it would have forced you to assume or ignore]. The framework excludes [excluded concept], which is relevant but [reason: not measurable here, outside scope, addressed elsewhere].
The third sentence is the one most students omit and every committee asks about. Naming the rival and saying why you did not take it is cheaper to write now than to improvise at a defence.
Five mistakes
- Definitions instead of relationships. A framework with four defined concepts and no propositions is a glossary. The arrows are the framework.
- Boxes that never reappear. If a concept has no measure in the methods chapter, delete it from the diagram or add the measure.
- Arrows with no direction. A line between two boxes says nothing. State whether A causes B, predicts B, or merely co-occurs with it.
- Confusing a moderator with a mediator. A moderator changes the strength of a relationship; a mediator carries it. They imply different analyses, and the diagram conventions differ.
- No boundary. A framework without exclusions claims to explain everything, which examiners read as not having thought about scope.
From framework to measurement
A finished framework is a to-do list for your methods chapter. Each concept needs an operational definition and an instrument; each relationship needs an analysis that can detect it. Turning a concept into something countable is its own task, worked through in the operationalisation guide, and choosing the analysis that matches a proposition is covered in which statistical test to use. Where a concept is measured with a published scale, its reliability evidence belongs with it — see what counts as an acceptable alpha and, for psychological constructs specifically, which scale to choose.
The framework also has to be consistent with the gap you stated. Read it against your problem statement: if the framework explains something your problem statement did not identify as unknown, one of the two needs rewriting.
Tesify drafts the framework section from your own reading, propositions included, so you start by arguing with a draft rather than staring at four boxes. Start your thesis with Tesify — 100% written by you.
Frequently asked questions
What is the difference between a conceptual and a theoretical framework?
A theoretical framework adopts one named theory whole and defends it against rivals. A conceptual framework assembles concepts from several literatures and defends the assembly. In practice the test is what you must justify: the choice of theory, or the choice of links.
Do I need both?
Rarely. Most theses need one. Some studies adopt a theory and then extend it with additional concepts, in which case you present the theory first and the extension as your conceptual contribution.
Does a qualitative study need a conceptual framework?
Usually yes, and it works differently: the concepts sensitise your analysis rather than specify variables. Some designs deliberately hold the framework loosely so it can be revised by the data, but that is a stated choice, not an absence.
Should the framework have a diagram?
In most fields, yes, and it must be readable on its own. The rule is that the diagram contains nothing the propositions do not state and omits nothing they do.
Where does the conceptual framework go?
Most commonly at the end of the literature review, after the review has established why these concepts. Some programmes place it in the introduction or give it its own short chapter.
How many concepts should it have?
Enough to state your propositions and no more. Four to six is typical for a master’s thesis. Every additional box is an additional measurement obligation.
Can I change the framework after collecting data?
In qualitative and exploratory designs, yes, and you report the revision as a finding. In a confirmatory quantitative design, changing the framework after seeing the results is a serious problem and must be disclosed if it happens.
What is the difference between a moderator and a mediator?
A moderator changes how strong a relationship is; a mediator is the mechanism through which it works. They are drawn differently and analysed differently, so decide which you mean before you draw.
Do I cite the definitions?
Yes. Adopt a published definition and cite it rather than writing your own, unless producing a new definition is part of your contribution — in which case say so explicitly.
What if my supervisor and my committee disagree about the framework?
Write the defence paragraph for both versions and take them to a meeting. The disagreement is usually about which rival was worth ruling out, and the paragraph makes that visible.