How to Draw a Conceptual Framework for a Dissertation, Step by Step (With a Business Example)

You have read the literature, you know which variables matter, and your supervisor has asked for a conceptual framework. This guide gives you seven numbered steps to draw one, the output you should have after each step, and a fully worked business example. Check your module handbook first, because some departments want the framework in the literature review and others in the methodology.

What is a conceptual framework in a dissertation?

A conceptual framework is your own model of the study. It shows the main concepts or variables you will investigate and the relationships you expect between them, usually as a diagram with a short written explanation. It is built from the literature, but it is not a copy of any one source: it is the argument of your literature review compressed into one picture.

It is different from a theoretical framework. A theoretical framework names an established theory, such as the theory of planned behaviour or Herzberg’s two-factor theory, that you apply. A conceptual framework is what you build, and it may draw on one theory, several theories or none. Imenda (2014) discusses whether the two terms really differ, and your own handbook will say which your department expects. If you need to choose a theory first, see our comparison of theoretical frameworks for education and our guide to theoretical frameworks for marketing.

Step 1: Write your research question and list the concepts in it

1. Write your main research question in one sentence.
2. Underline every noun phrase that names something you will measure or explore.
3. Write those concepts as a list.

Expected output: a list of between two and six concepts. If you have more than six, your question is probably two questions, and your framework will be unreadable. Cut or split before you draw anything. If your question is still vague, our guide to the research gap and rationale will help you narrow it.

Step 2: Give each concept a role

Assign every concept one of these roles, because the role decides where it sits on the page and how the arrows run.

• Independent variable (or predictor): the factor you expect to influence something else.
• Dependent variable (or outcome): what you expect to change.
• Mediator: a variable that explains how or why the independent variable affects the outcome.
• Moderator: a variable that changes the strength or direction of the relationship.
• Control variable: something you hold constant or measure so it does not distort the result.

Expected output: each concept labelled with one role. In a qualitative study you may not have independent and dependent variables. In that case label concepts as the central phenomenon, influencing factors and contexts.

Step 3: Define each concept and say how you will measure it

For each box, write a one-line definition from a cited source and one line on how you will measure it. This is what links the framework to your methodology and stops markers asking how a box in your diagram turns into a questionnaire item. The operationalisation table is where this work gets finished.

Expected output: a three-column table of concept, definition with citation and measure. A good test is that someone reading only your table could tell which scale or interview question belongs to each box.

Step 4: Decide the direction and type of each relationship

For every pair of concepts that you believe are linked, decide three things.

1. Direction: does A influence B, or do they simply vary together?
2. Sign: do you expect a positive or negative relationship?
3. Evidence: which source in your literature review supports it?

If you cannot name the evidence for an arrow, either find it or remove the arrow. An unsupported arrow is the most common reason a framework is criticised as speculative. If the evidence is mixed, say so in your explanation and word the hypothesis neutrally.

Expected output: a list of numbered relationships, each with a direction, a sign and at least one citation.

Step 5: Draw it

1. Put the independent variable on the left, the outcome on the right and any mediator between them.
2. Place moderators above or below the arrow they affect, pointing at the arrow rather than at a box.
3. Use boxes for measured variables, ovals for latent concepts if your department follows that convention, and one consistent arrow style.
4. Label each arrow with the hypothesis number (H1, H2 and so on).
5. Add a figure number and a caption, and cite any source you adapted: ‘Figure 2.1 Conceptual framework (adapted from Author, Year)’.

Expected output: one clean diagram that fits on a page, has no crossing arrows and could be understood without the surrounding text. Any drawing tool works, so choose the one you already use, because clarity matters more than software.

Step 6: Write the paragraph that explains it

A diagram on its own is never enough. Write one to three paragraphs that do four jobs: introduce the model, explain each relationship with its supporting evidence, say which parts your study tests and which it treats as given, and link the model to your research questions. Refer to the figure by number in the first sentence, and write the explanation in the order the hypotheses appear.

Here is an illustrative opening you can adapt: “Figure 2.1 presents the conceptual framework for this study. It proposes that flexible working (independent variable) is associated with higher job satisfaction (mediator), which in turn is associated with a stronger intention to stay (outcome), with tenure controlled. Each relationship is developed below with reference to the literature reviewed in Sections 2.2 to 2.4.” Notice that it names the figure, the roles, the direction and where the evidence lives.

Expected output: a short passage, roughly 250 to 400 words, in which every arrow in the figure is mentioned and every hypothesis is introduced.

Step 7: Check it against your research questions and your method

Run this checklist before you submit the chapter:

• Does every research question appear in the framework?
• Does every box have an instrument or a coding approach in your methodology?
• Does every arrow have a hypothesis, or in a qualitative study a guiding question?
• Will your discussion chapter return to the framework and say what the data showed for each arrow?

Expected output: a framework that your methodology, results and discussion all refer back to. If your proposal already had a draft framework, revise it to match what you now know rather than leaving the old version.

A worked example (illustrative): flexible working and staff retention

This example is invented for teaching and is not based on real data. Suppose your research question is: Does flexible working affect intention to stay among administrative staff in a UK university, and does job satisfaction explain the effect?

1. Concepts: flexible working, job satisfaction, intention to stay, and tenure as a control.
2. Roles: flexible working is the independent variable, job satisfaction is the mediator, intention to stay is the outcome and tenure is the control.
3. Definitions and measures: flexible working (hours or location choice, measured by a short self-report scale you cite), job satisfaction (a validated scale such as the one in our job satisfaction dissertation guide) and intention to stay (a short published scale).
4. Relationships: H1, flexible working is positively related to job satisfaction; H2, job satisfaction is positively related to intention to stay; H3, the direct effect of flexible working on intention to stay becomes smaller when satisfaction is included, which is the mediation hypothesis.
5. Drawing: flexible working on the left, an arrow labelled H1 to job satisfaction in the middle, an arrow labelled H2 to intention to stay on the right, and a curved arrow labelled H3 from flexible working directly to the outcome. Tenure sits below the outcome with a dotted line.
6. Explanation: the paragraph introduces Figure 2.1, cites the literature for each arrow, says that the study tests all three hypotheses and states that tenure is controlled but not hypothesised.
7. Check: three hypotheses, three measures and one mediation analysis, all of which appear in the methodology.

A Harvard reference for a source you adapt would follow this pattern, but check your department’s referencing style in your module handbook, as it varies: Imenda, S. (2014) ‘Is there a conceptual difference between theoretical and conceptual frameworks?’, Journal of Social Sciences, 38(2), pp. 185–195.

How does the framework change by subject? (illustrative sketches)

The seven steps stay the same, but the boxes change with the discipline. These sketches are invented to show the pattern, not to prescribe a model.

• Nursing: staffing levels (independent), nurse burnout (mediator) and patient-reported experience (outcome), with ward type as a control.
• Marketing: social media engagement (independent), brand trust (mediator) and purchase intention (outcome), with age as a moderator.
• Education: a tutoring programme (independent), pupil confidence (mediator) and attainment (outcome), with prior attainment as a control.
• Qualitative health study: the central phenomenon is the experience of discharge, surrounded by the influencing factors of family support, communication with staff and housing, drawn as a map rather than as a chain of hypotheses.

In each case the framework tells the reader what you will measure, how the pieces fit together and what you expect to find, which is exactly what a marker looks for in the first few pages of the chapter.

What mistakes make a conceptual framework fail?

• Too many boxes. A crowded diagram signals that the study has no focus.
• Arrows with no evidence. Every relationship needs a citation or a clear reason.
• No link to the method. A box that cannot be measured is a decoration.
• Drawing it and never using it. The results and discussion should return to it, arrow by arrow.
• Confusing it with the theoretical framework. Use both terms only if your department does, and define each.

Tesify helps you build the framework step by step alongside your literature review and methodology, so the model, hypotheses and measures stay consistent while you write every word yourself. Over 9,000 students have used it across more than 15,000 chapters. Build your conceptual framework with Tesify.

Conceptual framework FAQ

What is a conceptual framework in a dissertation?

It is your own diagram and short explanation of the main variables in your study and the relationships you expect between them, built from the literature.

What is the difference between a conceptual framework and a theoretical framework?

A theoretical framework names the established theory you use. A conceptual framework is your own model of your variables and how they relate, which may draw on one or more theories.

Do I need a conceptual framework for a qualitative dissertation?

Often yes, but it can be looser, showing the concepts that guide your questions rather than testable hypotheses. Check your module handbook.

Where does the conceptual framework go in a dissertation?

Usually at the end of the literature review, so it summarises the argument, or at the start of the methodology. Follow your department’s template.

How many variables should a conceptual framework have?

As few as your research questions need. A framework with three to five variables is usually easier to measure and defend than a crowded one.

Should every arrow have a hypothesis?

In a quantitative study each arrow you test should map to a hypothesis, and the label on the arrow should match the hypothesis number.

Can I copy a published conceptual framework?

You can adapt one and cite it, but you must explain how you changed it for your context. Copying a figure without citation is plagiarism.

What software can I use to draw it?

Any tool that makes clean boxes and arrows, including PowerPoint, Word shapes or a dedicated diagram tool. Clarity matters more than the software.

How long should the explanation of the framework be?

Usually one to three paragraphs: introduce the model, explain each relationship with evidence, and state which parts your study tests.