---
name: theoretical-and-conceptual-framework-development
description: >
  Selects and justifies a theoretical lens against alternatives, then builds a
  conceptual framework whose constructs, relationships and indicators connect to
  the research questions. Use for "theoretical framework", "conceptual
  framework", "what theory should I use", "difference between theoretical and
  conceptual framework", "my framework diagram", "operationalise my constructs",
  "how do I justify my theory choice", "framework for a qualitative study",
  "my study is theory-generating not theory-testing".
category: 15 Academic University Research
ref: "15.07"
tier: 3
inherits: [K2, K3, K4, K5]
---

# Theoretical and Conceptual Framework Development

## 1. One-line description

A method for choosing a theoretical lens with a defensible reason, and building from it a conceptual framework in which every construct is defined, every relationship is specified rather than merely drawn, and every element earns its place by connecting to a research question.

## 2. What this skill is used for

**The research problem it solves.** Framework chapters fail in a small number of predictable ways, and markers penalise all of them. The four terms are conflated, so a study claims a theoretical framework and presents a diagram of variables. A theory is adopted because it appeared in the first three papers read, with no account of what was rejected or why. Constructs are named but never defined, so the same word means different things in the literature review and the methodology. Relationships are drawn as arrows with no stated direction, sign, form or condition, which makes the diagram undisprovable and therefore useless. And the framework, once drawn, plays no further part: it does not shape the instrument, the analysis or the discussion, so the chapter is decorative. This skill forces each of those to be resolved explicitly, and it makes the framework do work downstream.

**Where it sits in the research lifecycle.** After the literature review has established what is known and what is contested, and before the methodology chapter, because the framework determines what must be measured or observed and therefore constrains the design. It is revisited in the discussion, where the findings are read back against it.

**Typical use cases.**
- Choosing between two or three plausible theoretical lenses and writing the justification.
- Turning a literature review into a framework rather than a summary.
- Defining constructs so that the instrument measures what the framework claims.
- Building a diagram that a supervisor cannot dismiss as boxes and arrows.
- Using a framework in qualitative work without letting it dictate the codes.
- Handling a study that generates theory rather than testing it, where the conventional framework advice does not apply.
- Preparing to defend the theory choice in a viva against the question "why not X instead".

**Who uses it.** Masters candidates writing a framework chapter or section; supervisors reviewing a framework before the methodology is written; students in management, education, psychology, information systems, health services research, sociology and public policy, where an explicit framework is expected.

## 3. When to use it

- The dissertation structure requires a theoretical or conceptual framework chapter or section.
- You have a literature review and no clear account of what theory the study is using.
- A supervisor has asked "what is your theoretical framework" and the honest answer is not yet available.
- Two or more theories could plausibly frame the study and the choice has not been made or argued.
- You have a diagram with arrows and cannot say what any arrow claims.
- The instrument is being designed and it is unclear which constructs it must capture.
- The study is qualitative and you need a framework that orients without prescribing the analysis.
- The study is exploratory or theory-generating and you need to know what belongs in the framework section instead.

## 4. When NOT to use it

- **The discipline does not use explicit frameworks in this way.** Some fields expect a literature-driven rationale and treat a formal framework chapter as an imported convention. Engineering, some areas of computer science, some historical and legal scholarship, and some strands of applied statistics do not require one. Check the departmental handbook and recent accepted theses in the department before manufacturing a framework nobody asked for. A framework bolted on because a template had a slot for it is visible and is marked down.
- **The study is a systematic review with no primary data.** The framework work in a review is different: it concerns how the review classifies and synthesises evidence, not how a study observes the world. Use **15.06 Systematic Literature Review**, which handles the classification structure directly.
- **The research question has not stabilised.** A framework built against a moving question changes with it and wastes the effort twice. Settle the question first with **15.01 Academic Research Topic Selection** or, where the problem is that the question is not yet answerable, **01.02 Business Problem to Research Question**.
- **The purpose is generating testable propositions rather than framing an empirical study.** Developing theory as the contribution, with construct definitions, propositions, mechanisms and boundary conditions, is a doctoral-level task and is **15.15 Theory Building and Conceptual Contribution**. This skill covers theory-informed and theory-generating masters work up to and including an emergent framework as an outcome. It does not cover formal theory construction as the primary contribution.
- **Where applying a framework would distort a genuinely inductive design.** In approaches that require the researcher to hold prior conceptualisation at arm's length, imposing a framework before fieldwork breaks the method's own logic. Step 12 handles this case, and the correct output is a stated orienting position plus an explicit account of what was deliberately not pre-specified, not a framework diagram.
- **The framework is being reverse-engineered to fit results already obtained.** Selecting a theory after analysis because it explains what you found, and presenting it as though it guided the study, misrepresents the research process. It is also easy for an examiner to detect, because the instrument will not match the framework. If a theory only became relevant after the findings, that is a discussion-chapter contribution and belongs in **15.11 Discussion and Contribution Development**, described honestly as post hoc.
- **Your institution prohibits AI assistance for this task.** See §12 item 1.

## 5. Required inputs

**Required.**
- **The research questions or objectives, in their current wording.** The framework's job is to connect to these. Without them there is no test of whether a construct belongs.
- **The literature the framework will be built from, actually read.** Framework work is a reading of specific sources. A framework assembled from general knowledge of a field will misattribute constructs and cite theories for claims they do not make.
- **The study's basic design intent: broadly quantitative, qualitative or mixed, and broadly confirmatory or exploratory.** This determines whether the framework is a set of testable relationships, a sensitising structure, or an anticipated output.

**Optional, and what each one adds.**
- **The departmental handbook's definition of the required sections.** Departments use "theoretical framework" and "conceptual framework" inconsistently, and matching the local usage removes an unnecessary argument with a marker.
- **Two or three candidate theories the supervisor has mentioned.** Gives the comparison something to start from, and reveals the supervisor's prior, which the chapter will be read against.
- **Existing validated instruments for the constructs.** Shows how the field has operationalised each construct already, which is the fastest route to a defensible indicator set and to comparability with prior work.
- **A prior study using the same lens in a different context.** Provides a model for how the theory has been applied, and the contextual difference is often the study's contribution.
- **The analysis plan, even in draft.** Lets you check that the framework's relationships are actually testable or observable by the analysis proposed, which is where frameworks most often turn out to be decorative.
- **The marking rubric.** Tells you what proportion of the marks the framework carries, which should determine how much of the word count it gets.

## 6. Questions to ask before starting

1. **Is this study testing a theory, using a theory to look, or building one?** This is the single most consequential question and it determines the whole shape of the chapter. Default if unanswered: infer from the design intent, state the inference explicitly at the top of the chapter, and use the theory-informed middle case, which is the most common at masters level.
2. **What does your department mean by "theoretical framework" and "conceptual framework"?** Usage varies and both terms are used for both things in different handbooks. Default: define both terms explicitly in your own text, state which sense you are using, and proceed. A stated definition cannot be marked wrong for being the other convention.
3. **Which candidate theories has the literature already applied to this phenomenon?** Determines what the alternatives are and therefore what your justification must argue against. Default: identify at least two alternatives from the literature you have read, and compare against them rather than against nothing.
4. **What will the framework have to do later: shape an instrument, shape a coding structure, or be an output?** Determines the level of specification required. A framework feeding a survey needs indicators. Default: assume it must shape the instrument, which is the more demanding case.
5. **Is there a level-of-analysis mismatch between your theory and your data?** A theory about organisations tested with data from individuals is a common and heavily penalised error. Default: state the level of every construct explicitly and check it against the unit of data collection.
6. **What would this theory say is surprising?** If you cannot answer, the theory is being used as a label rather than a lens. Default: work it out before writing, because an examiner will ask.
7. **What is the framework's boundary: to whom, where and when does it claim to apply?** Determines the generalisation claim in the discussion. Default: state the boundary as your study's context, and treat any wider claim as requiring an argument.

## 7. Step-by-step methodology

**Step 1. Fix the four terms, in your own text, before using any of them.**
The conflation is the commonest framework error and the easiest to fix. A **theory** is a general explanatory account: a set of concepts and stated relationships that explains a class of phenomena beyond any one study, developed and tested by others. A **theoretical framework** is your study's use of one or more existing theories: which theory, why this one, what it directs you to look at, and how it applies to your context. A **conceptual framework** is the structure of your particular study: the constructs it involves, how they are defined here, and the relationships proposed between them, whether or not they come from a single theory. A **model** is a formal representation of relationships, typically specifying form and often quantified, which can be tested and can fail. The practical distinction that matters: a theoretical framework is borrowed and justified; a conceptual framework is built and defended; a model makes a commitment precise enough to be wrong. Write one paragraph defining each as you will use them, and use them consistently for the rest of the dissertation. *Correct result: four definitions in your own words, and a stated position on which of the four your chapter presents.*

**Step 2. Establish what the study needs a framework for.**
Three cases, and everything downstream differs. **Theory-testing**: an existing theory generates hypotheses or expectations that this study evaluates, and the framework must be specified tightly enough to be disconfirmed. **Theory-informed** (the most common at masters level): a theory directs attention, defines constructs and organises the analysis, but the study is not designed to test the theory itself, and the framework's job is to justify what is looked at and what is ignored. **Theory-generating**: the study aims to produce a conceptual account, and the pre-fieldwork framework section states the orienting position and the sensitising concepts rather than a structure to be filled in. State which case you are in, in one sentence, in the chapter. Ambiguity here is what produces a chapter that claims to test a theory and then reports themes. *Correct result: a stated case with a one-line justification tied to the research questions.*

**Step 3. Identify the candidate lenses from the literature, not from recall.**
List the theories that the literature you have actually read applies to this phenomenon, with the source for each. Where a theory is well known but you have not read a primary or authoritative source for it, mark it as a candidate to be read, not as a candidate to be used. Three to five candidates is enough. Include at least one from an adjacent discipline if the phenomenon sits at a boundary, because cross-disciplinary lenses are a common source of masters-level contribution. Then, for each candidate, write the theory's core claim in one sentence in your own words. If you cannot, you do not yet know the theory well enough to reject it or to adopt it. *Correct result: a candidate table with theory, core claim in one sentence, the source you read, and the discipline it comes from.*

**Step 4. Interrogate each candidate on four questions, and let the answers do the selecting.**
For each candidate lens, answer: (a) **What does it tell you to look at?** Every theory foregrounds some things and backgrounds others; a theory of institutional pressure directs attention to external legitimacy and away from individual motivation. (b) **What counts as an explanation under it?** Some theories explain by mechanism, some by function, some by meaning, some by structural position, and the answer determines what your findings chapter will have to contain to count as an answer. (c) **What would surprise it?** A theory that cannot be surprised explains everything and constrains nothing, which is a reason to reject it as a lens even when it is true. (d) **What are its boundary conditions?** The populations, settings and periods in which it has been supported, and whether your context sits inside or outside them. Answering (c) is the step students skip and the step that most improves the chapter, because it converts a label into an analytical commitment. *Correct result: a four-column comparison across candidates, in which the differences between them are visible.*

**Step 5. Select, and write the justification as an argument against the alternatives.**
The selection is justified by fit to the research question, not by the theory's prominence. State the choice, then state, for each rejected candidate, the specific reason: it operates at the wrong level of analysis, its boundary conditions exclude your context, it explains a different part of the phenomenon, it cannot be observed with the data you can obtain, or it has been applied extensively here already and adds nothing. "It was the most appropriate" is not a reason. Where two lenses are genuinely complementary, combining them is legitimate, but a combination requires an account of how the two fit together and where they conflict, which is more work than choosing one and is only worth it if the study needs both. `RESEARCHER REVIEW RECOMMENDED` on any combination, since the compatibility judgement rests on knowledge of the field's debates (K5 §2.2). *Correct result: a justification paragraph naming the alternatives and the reason each was set aside, which is also the answer to the viva question.*

**Step 6. Extract the constructs and define each one twice.**
List every construct the framework contains. For each, write a **conceptual definition** (what it means, in the sense the framework uses, cited to where that sense comes from) and a **working definition for this study** (how it is bounded here, and what falls outside it). Two definitions are necessary because the literature's sense and your study's use rarely coincide exactly, and the gap between them is where construct drift starts. Then check every construct against three tests: is it distinct from the others in the framework, or does it overlap; is it at a stated level of analysis (individual, team, organisation, system); and is it named consistently, since a construct that is "engagement" in the framework and "commitment" in the instrument will be caught. *Correct result: a construct table with conceptual definition, study definition, level, source, and a note of what it is deliberately not.*

**Step 7. Specify every relationship, and delete any arrow you cannot specify.**
An arrow in a framework diagram is a claim, and it must be readable as one. Specify five properties for each: **direction** (which construct is proposed to affect or precede which, and on what basis, since a plausible reverse direction that is not addressed is a standing viva question); **sign or expected pattern** (positive, negative, curvilinear, or a described qualitative pattern); **form** (direct, mediated through a named construct, moderated by a named construct, or reciprocal); **conditionality** (under what circumstances the relationship is expected to hold); and **grounding** (the theory or the empirical evidence the proposal rests on, cited). Write each relationship as a numbered proposition in a sentence: "P3: Perceived autonomy is positively associated with reported engagement, and this association is stronger where task interdependence is low." Then apply the deletion rule: if you cannot write the sentence, remove the arrow. Frameworks improve dramatically under this rule, because most decorative frameworks are decorative in precisely the places where the author had nothing to say. *Correct result: a numbered proposition list in which every arrow in the diagram appears exactly once, and no proposition is a restatement of a definition.*

**Step 8. Operationalise, and make the construct-indicator gap visible.**
Work down the ladder for every construct that will be measured or observed: construct → dimensions (if it has them) → indicators → items, questions or observable events. Then state the gap explicitly, because this is where frameworks quietly become untestable. An indicator is never the construct: a self-report of intention is not behaviour, a count of logins is not engagement, an exam mark is not learning. For each construct, write one sentence naming what the indicator captures well and one naming what it misses. Where the field has validated instruments, prefer them and say why, since comparability with prior work is usually worth more than a bespoke item set, and adopting an instrument requires its own justification (see **15.08 Research Design and Methodology Chapter**, and **02.07 Scale and Measurement Selection** for the selection criteria). Where a construct in your framework has no available indicator and none can be built, that is a finding about the design: either the construct comes out of the framework or the study cannot address it. *Correct result: an operationalisation table running from construct to indicator, with a stated gap statement per construct, and any unmeasurable construct explicitly resolved.*

**Step 9. Build the traceability matrix that ties the framework to the study.**
One table, and it is the single most useful artefact this skill produces. Columns: research question, construct or relationship it involves, where it comes from in the framework, how it will be measured or observed, and where it will be analysed. Every research question must appear. Every construct must appear against at least one question. Anything in the framework that does not appear in the matrix is decoration and should be cut. Anything in the research questions that has no framework element beneath it is unframed, and either the framework is incomplete or the question is out of scope. This matrix reappears in the methodology chapter, the analysis chapter's structure and the discussion, and building it once saves rebuilding it three times. *Correct result: a matrix with no empty cells and no orphan constructs.*

**Step 10. Draw the diagram, then test it against four criteria.**
A meaningful framework diagram satisfies four tests. **Every element is labelled with a construct that appears in the construct table.** **Every arrow corresponds to a numbered proposition**, and the diagram carries the numbers so a reader can look them up. **The notation is explained**, so that different arrow types, box types and groupings mean something stated in a key rather than something the reader must guess. **The diagram would change if the study's claims changed**, which is the test that separates a framework from an illustration: if you could delete a construct without altering any proposition, the construct is ornamental. Keep it to what fits legibly on one page; a framework that needs a foldout has too many constructs for a masters study. *Correct result: a numbered, keyed, one-page diagram in which nothing appears that is not in the proposition list.*

**Step 11. In qualitative work, use the framework as a lens rather than a coding structure.**
The risk in qualitative studies is real and specific: a pre-specified framework becomes the code frame, the analysis finds the framework's categories because it was looking for them, and the study reports its own assumptions back as findings. Prevent it with three practices. First, use the framework's constructs as **sensitising concepts** that orient attention, stated as such in the methodology, rather than as codes applied at first pass. Second, code inductively at least once before mapping to the framework, and keep both the inductive codes and the mapping, so the reader can see what the framework captured and what it did not. Third, and most importantly, **report the residual**: the material that did not fit the framework. That residual is usually where the study's contribution is, because it is the part the existing theory does not account for. A qualitative study using a framework and reporting no residual has either an unusually complete theory or an analysis that stopped looking. *Correct result: a stated position on how the framework was used in analysis, an inductive-to-framework mapping, and an explicit account of what did not fit.*

**Step 12. Where the study is theory-generating, write the section the design actually calls for.**
Theory-generating studies still need this chapter, but its content is different. It contains: the orienting theoretical position (what tradition of explanation the study works within, since no study observes without one); the sensitising concepts drawn from the literature and the reason for each; an explicit statement of what has deliberately not been pre-specified, and why the design requires that; and the criteria by which the emergent conceptual account will be judged adequate. The framework diagram then appears at the end of the dissertation rather than the beginning, as an output, and the discussion compares it to existing theory. State this structure explicitly early, because a marker expecting a framework in chapter three and finding one in chapter six will read it as an omission unless told. *Correct result: an orienting-position section with sensitising concepts and non-specification justified, plus a stated plan for where the emergent framework appears.*

## 8. Analytical framework

Two structures, one vertical and one horizontal, and the chapter is built where they meet.

**The specification ladder**, applied to every construct:

    Theory (general explanatory account, borrowed)
      → Construct (defined here, at a stated level of analysis)
        → Dimension (where the construct has parts)
          → Indicator (what will stand for it)
            → Item or observable event (what is actually collected)

Applying it: each descent loses something, and the loss must be stated rather than assumed away. The framework's claims live at the construct level; the study's evidence lives at the item level; and the distance between them is the study's measurement risk. A framework chapter that never descends below "construct" cannot support an instrument, and a methodology chapter that never ascends above "item" cannot support a conclusion.

**The linkage chain**, applied to the framework as a whole:

    Research question → Construct → Proposition → Indicator →
    Analysis → Finding → Reading back against the framework

Applying it: every link must be traversable in both directions. Forwards, it says the framework determines what the study does. Backwards, it says every finding can be located in the framework, which is what makes the discussion chapter writeable. The traceability matrix in Step 9 is this chain in table form, and the commonest structural failure in a dissertation is a break in it: a research question with no construct, a construct with no indicator, or a finding with nowhere in the framework to land.

## 9. Output format

**1. Definitions.** Theory, theoretical framework, conceptual framework and model, as used in this dissertation, in the author's own words.

**2. The study's relationship to theory.** Theory-testing, theory-informed or theory-generating, stated and justified against the research questions.

**3. Candidate lenses considered.**

| Theory | Core claim (one sentence) | What it directs attention to | What counts as explanation | What would surprise it | Boundary conditions | Source read |
|---|---|---|---|---|---|---|

**4. Selection and justification.** The chosen lens, and the specific reason each alternative was set aside.

**5. Construct definitions.**

| Construct | Conceptual definition (with source) | Working definition in this study | Level of analysis | Deliberately excludes |
|---|---|---|---|---|

**6. Propositions.** Numbered, one sentence each, specifying direction, sign, form, conditionality and grounding.

**7. Operationalisation.**

| Construct | Dimensions | Indicators | Items or observations | What the indicator captures | What it misses |
|---|---|---|---|---|---|

**8. Framework diagram.** One page, numbered to the propositions, with a notation key.

**9. Traceability matrix.** Research question against construct, proposition, measurement and analysis location.

**10. Framework boundary and limitations.** To whom, where and when the framework claims to apply, and what it does not attempt to explain.

**When the theory base is thin, the format must not force fabrication (K4 §1).** Where a phenomenon has little theoretical treatment, the honest output is a smaller framework built from empirical regularities in the literature, labelled as such, with the absence of an established lens stated as a finding about the field and a reason for the study. Do not import a well-known theory that does not fit in order to fill the section, and do not cite a theory you have not read. A three-construct framework that is fully specified is stronger than a nine-construct one where six arrows mean nothing.

## 10. Quality checks

Run before the chapter goes to a supervisor. These sit on top of K4 §8.

1. Are theory, theoretical framework, conceptual framework and model each defined and used consistently throughout the dissertation?
2. Is the study's relationship to theory stated as testing, informing or generating, and does the design match that statement?
3. Does the justification name specific alternatives and give a specific reason for setting each aside?
4. Has every theory cited been read in a primary or authoritative source, rather than characterised from secondary description?
5. Does every construct have both a conceptual and a study-specific definition, and a stated level of analysis?
6. Is any construct measured at a different level from the one the theory claims?
7. Does every arrow in the diagram correspond to a numbered proposition, and every proposition to an arrow?
8. Can every proposition be stated as a sentence specifying direction, sign, form and conditionality?
9. Does every measured construct have indicators, and is the gap between construct and indicator stated?
10. Does the traceability matrix contain every research question and every construct, with no orphans?
11. Does the diagram carry a notation key, and would it change if the study's claims changed?
12. In qualitative work, is the framework's role in analysis stated, and is the material that did not fit reported?
13. Does the framework reappear in the discussion chapter, or was it abandoned after chapter three?
14. Is the framework's boundary of application stated, and does any later claim exceed it?

## 11. Common failure modes

| Failure | How to recognise it | How to prevent it |
|---|---|---|
| **Term conflation** | "Theoretical framework" heading over a diagram of study variables | Define all four terms in your own text first (Step 1) |
| **First-theory adoption** | The lens is whatever appeared in the earliest papers read | Compare three to five candidates on four questions (Steps 3 and 4) |
| **Theory as label** | The theory is named in chapter two and never used again | Traceability matrix, and read findings back against it (Steps 9, 12) |
| **Meaningless arrows** | Arrows with no stated direction, sign or condition | The deletion rule: no proposition, no arrow (Step 7) |
| **Undefined constructs** | The same word used differently in chapters two, three and five | Two definitions per construct, checked against the instrument (Step 6) |
| **Level mismatch** | An organisational theory tested on individual respondents | State the level for every construct and check against the data unit |
| **Construct-indicator collapse** | Logins reported as engagement without qualification | Write the captures and misses statement per construct (Step 8) |
| **Coding straitjacket** | Qualitative themes that exactly reproduce the framework | Inductive pass first, and report the residual (Step 11) |
| **Reverse-engineered theory** | The lens explains the findings suspiciously well, instrument does not match | If it emerged post hoc, report it in the discussion as post hoc (§4) |
| **Framework inflation** | Eleven constructs in a study with one instrument and 90 respondents | Cut anything absent from the traceability matrix (Step 9) |
| **AI-attributed theory** | A theory summarised confidently from model knowledge, subtly wrong | Only theories read in a real source enter the chapter (§12) |
| **Invented framework citation** | A named framework attributed to plausible authors and a plausible year | Never supply a citation from recall (§12) |
| **Borrowed diagram** | A framework diagram reproduced from a source without attribution | Attribute, or build your own and say what it adapts |

## 12. AI guardrails

Skill-specific. The universal prohibitions in K4 apply in full and are not repeated. The human in the loop for every K5 marker in this skill is the candidate together with their supervisor.

1. **Academic integrity is a condition of use, not a footnote.** These skills assist a researcher's thinking, structure and rigour. They do not produce work to be submitted as the student's own unaided output. The user must comply with their institution's AI use policy and its declaration requirements, which vary by institution and by assessment. Where an institution prohibits AI assistance for a task, this skill must not be used for it. The skill never writes a passage for submission as though the student wrote it; it interrogates, structures, critiques and teaches. Operationally in this skill: it will supply the comparison structure, the interrogation questions, the construct and proposition templates, the traceability matrix and a critique of the student's draft framework. It will not write the framework chapter's argument, and it will not supply the theory the student should have chosen.

2. **Never attribute a theory, a construct or a framework to an author, a year or a source from recall.** This is the single highest-risk behaviour in framework work, because theory attributions are exactly the kind of thing that feels certain and is frequently wrong in the details that matter: the year, the co-author, the specific paper, the exact formulation. Name the theory, describe its claim, and say that the student must cite the source they have read. Never render an attribution in reference format.

3. **Never characterise a theory the student has not read as though it had been read.** A summary of a theory from model knowledge will be plausible and will subtly misstate its scope, its boundary conditions or its mechanism, and the error will propagate into the operationalisation. Offer the description as a starting characterisation to be checked against a primary source, labelled as such.

4. **Never invent a proposition the literature is said to support.** A proposition is either grounded in a source the student has read, grounded in their own reasoning and labelled as such, or absent. "The literature suggests" without a source is a fabrication with a hedge in front of it.

5. **Never fill a framework to make the diagram look complete.** If a construct has no grounding, no indicator and no research question, it does not go in. Format is not evidence (K4 §1). A sparse, fully specified framework is a better output than a full, partly meaningless one.

6. **Never propose a causal relationship where the design cannot support one** (K4 §3.2). A framework may propose direction as a theoretical claim, and it must then say plainly that the study's design tests association rather than causation, in the same passage, so that the discussion chapter cannot quietly upgrade it.

7. **Never resolve a genuine theoretical debate on the student's behalf.** Where two schools disagree about a construct's nature, that disagreement belongs in the chapter, with a stated position and a reason. Presenting one side as settled is both wrong and easy for an examiner to expose.

8. **Never convert an inductive study into a deductive one by supplying a framework.** Where the student describes an exploratory or theory-generating design, the correct assistance is Step 12, not a set of propositions. Supplying hypotheses to a study not built to test them damages the design.

9. **Where the framework and the instrument do not match, say so rather than smoothing the description.** A mismatch between what the framework claims and what the study can observe is a design problem to be surfaced now, cheaply, rather than discovered by an examiner.

## 13. Best-practice principles

- **The theory you can be surprised by is the theory worth using.** A lens that accommodates any result organises nothing, and this test eliminates most weak candidates in a single question.
- **Justification is comparative or it is not justification.** "Appropriate for this study" says nothing. "Chosen over the alternative because that theory operates at the organisational level and the data are individual" says everything.
- **Define a construct by what it excludes as well as what it includes.** The exclusion is what stops it drifting into a neighbouring construct halfway through the analysis.
- **An arrow you cannot write as a sentence should not be drawn.** The deletion rule removes more bad framework material than any other single practice.
- **The gap between a construct and its indicator is not a flaw to hide, it is a limitation to state.** Every empirical study has it; only weak ones pretend otherwise.
- **Levels of analysis are where frameworks break silently.** Check the level of every construct against the unit at which data are collected, every time.
- **A framework that does not appear after chapter three was decoration.** Plan its reappearance in the analysis structure and the discussion before the chapter is finished.
- **In qualitative work, the interesting material is the residual.** What the framework failed to capture is where a masters-level contribution most often lives.
- **Borrowing a lens from an adjacent discipline is a legitimate and underused source of contribution**, provided the boundary conditions are checked rather than assumed to travel.
- **Fewer constructs, better specified, always.** A masters study can support perhaps three to six constructs and a handful of relationships. Ambition in a framework is paid for in the analysis chapter.
- **Write the viva answer while writing the chapter.** "Why this theory and not that one" is asked in almost every defence, and the answer is Step 5, already written.

## 14. Worked example

Generic fictional scenario, academic, in a public sector setting.

**INPUT**

A masters candidate in public administration is studying why a national digital services programme is adopted enthusiastically by some municipal offices and resisted in others. The literature review has surfaced three plausible lenses: a technology acceptance tradition centred on perceived usefulness and ease of use; an institutional tradition centred on legitimacy and conformity pressures; and a street-level bureaucracy tradition centred on frontline discretion. The design is qualitative, interviews with frontline staff and supervisors in six offices.

**PROCESS**

*Steps 1 and 2.* The candidate defines the four terms and states the case: theory-informed, not theory-testing, since six offices and 22 interviews cannot test a theory, and not theory-generating, since three well-developed traditions already address this phenomenon.

*Steps 3 and 4.* The four questions are answered for each candidate. The acceptance tradition directs attention to individual perceptions of the technology, explains by individual belief, would be surprised by adoption despite low perceived usefulness, and has boundary conditions established largely in voluntary-use settings. The institutional tradition directs attention to pressures from above and from peer organisations, explains by conformity and legitimacy-seeking, would be surprised by sustained visible divergence between similar offices under the same mandate, and operates at the organisational level. The street-level tradition directs attention to how frontline staff use discretion to manage the gap between policy and caseload, explains by the practical demands of the work, and would be surprised by staff adopting a tool that increases their per-case time with no offsetting relief.

*The judgement call.* The acceptance tradition is the obvious default and the one the supervisor mentioned. It is rejected, and the rejection is written out: use of the system is mandated, not voluntary, which sits outside the boundary conditions where that tradition's constructs have been supported, and its explanation would reduce a question about organisational variation to a question about individual attitude. The institutional tradition is rejected on level of analysis: the data are individual accounts from 22 people in six offices, which cannot support organisational-level claims about field-wide conformity, and pretending otherwise would be the level mismatch that gets caught in a viva. The street-level tradition is selected because its unit of explanation, the discretionary decision made by an individual under caseload pressure, is exactly the unit the interviews will capture.

*Steps 6 and 7.* Constructs are defined: discretion (bounded here as the latitude a caseworker exercises in how a case is processed, excluding formal authority to vary the outcome), caseload pressure, perceived fit between the tool and the work, and coping behaviour. Four propositions are written, each as a sentence with direction and conditionality. One draft arrow, from "organisational culture" to "coping behaviour", cannot be written as a specifiable sentence at all and is deleted, which removes a construct that had no interview question behind it.

*Steps 8 and 11.* Operationalisation is qualitative: each construct is mapped to interview prompts and to the observable talk that would evidence it, and each carries a captures-and-misses statement (self-reported discretion misses the discretion staff do not recognise as discretion). The framework is stated as sensitising, the first coding pass is inductive, and the mapping is retained. The residual is planned for explicitly.

**OUTPUT**

A framework chapter with four defined constructs, four numbered propositions, a one-page keyed diagram, an operationalisation table linking each construct to interview prompts, and a traceability matrix connecting three research questions to constructs, prompts and planned analysis. The justification section names both rejected traditions and the specific reason for each. In the event, the inductive pass produces a strong residual theme about staff protecting a colleague who could not use the system, which the framework does not account for and which becomes the discussion chapter's contribution.

`RESEARCHER REVIEW RECOMMENDED` on the lens selection, since the judgement about whether the street-level tradition is currently well regarded in this subfield depends on knowledge of the field's debates that the framework work itself does not supply (K5 §2.2).

## 15. Advanced usage

**Framework as a design constraint, deliberately.** Build the framework, then use it to cut scope: any construct that cannot be measured with the access and time available is removed, and the removal is recorded as a scope decision with a reason. This turns the framework into the tool that keeps a masters study finishable, which is a use most students never discover.

**Competing frameworks as the design.** Where two traditions make different predictions about the same phenomenon, a study can be designed to see which better accounts for the data, and this is a strong masters contribution because it requires no novel method. It requires both frameworks specified to the same level of detail, and an honest account of the case where neither fits.

**Retrospective framework audit.** Take a finished draft and reconstruct the traceability matrix backwards from the findings. Every finding with no framework location, and every framework element with no finding, is a coherence defect a marker will see. This audit takes an hour and catches the most damaging structural problems in a dissertation. See **15.18 Thesis Architecture and Chapter Coherence** for the full-document version.

**Where the standard approach does not fit.** In interpretive traditions that treat pre-specified constructs as an imposition, the equivalent work is articulating the theoretical position and the assumptions the researcher brings, rather than building a structure. In highly applied or design-oriented work, the framework may be a set of requirements and constraints rather than propositions, and it is judged on whether it makes the design decisions traceable. In both cases, the requirement that survives is the traceability chain: something must connect the research question to what was done and to what was found.

## 16. Skill chain

**Recommended previous skills:**
- **15.06 Systematic Literature Review.** Hands over the constructs, relationships and contested claims the framework is built from, and the established-versus-assumed classification that shows which relationships are actually supported.
- **15.02 Academic Literature Search Strategy.** Hands over the sources in which the candidate lenses were found.
- **15.01 Academic Research Topic Selection.** Hands over the stabilised research questions the framework must connect to.

**Recommended next skills:**
- **15.08 Research Design and Methodology Chapter.** Takes the operationalisation table and the traceability matrix and turns them into a justified design.
- **02.07 Scale and Measurement Selection.** Takes the construct definitions and selects or justifies instruments for them.
- **15.10 Data Analysis Chapter Development.** Takes the traceability matrix as the analysis chapter's structure.
- **15.11 Discussion and Contribution Development.** Reads the findings back against the framework, including the residual.

**Runs well alongside:**
- **01.03 Hypothesis Development**, where the study is theory-testing and the propositions become hypotheses.
- **15.05 Academic Writing Structure and Argumentation**, for writing the justification as an argument rather than a description.
- **15.12 Thesis Defence and Viva Preparation**, which reuses the Step 5 justification directly.

---
A Yazi Supplied Skill and resource.