A conceptual framework for a public health research report is a diagram, plus a short explanatory paragraph, that shows the variables in your study and how you expect them to relate. Draw it in six steps: list your variables, choose a published model as the backbone, map the model to your variables, draw it, write it up, and check that everything in it is measured.
What is a conceptual framework in a public health research report?
In public health the conceptual framework is usually a modified version of an established behavioural or social model, adapted to your outcome and your population. It is not the same thing as the theoretical framework, although South African faculties sometimes use the two headings loosely. The theoretical framework names the theory or model you draw on. The conceptual framework is your diagram: which factors you expect to influence your outcome, which you measure, and the direction of each relationship.
The site’s guide with conceptual framework examples across South African disciplines shows six general cases. This guide goes one level deeper for public health, where students tend to use three families of models, each suited to a different kind of question.
Step 1: List your variables from the research questions
Output: a list with one outcome, the explanatory factors and any moderators or confounders.
Start from your research questions, not from a model. Write the outcome first (for example, attendance at a clinic for a screening test), then every factor your questions say may influence it, then any characteristics you will control for, such as age, sex, education or distance to the facility. Only once you have this list do you look for a model that organises it. A framework chosen first tends to force your variables into boxes that do not fit them.
Step 2: Choose a published model as the backbone
Output: one named model, with its original source, that fits the kind of question you are asking.
Three families cover most public health research reports. Always cite the original source of the model, not a textbook summary of it.
| Model | Best for questions about | Core structure | Original source |
|---|---|---|---|
| Health Belief Model | Why individuals do or do not take a preventive or screening action | Perceived susceptibility, perceived severity, perceived benefits and perceived barriers, with cues to action; self-efficacy was added in later versions | Rosenstock (1974); Janz and Becker (1984) |
| Andersen’s behavioural model of health services use | Who uses health services and why, including access | Population characteristics (predisposing, enabling and need factors), health behaviour (service use) and outcomes | Andersen (1995) |
| Socio-ecological model | Factors at several levels, from the individual to policy | Intrapersonal, interpersonal, institutional, community and public policy levels | McLeroy et al. (1988) |
Pick the model whose structure matches your question. A study of why women do not attend cervical screening may fit the Health Belief Model. A study of why rural households use or avoid a clinic fits Andersen. A study of adolescent health that includes school, family and policy fits a socio-ecological structure.
Step 3: Map the model to your variables
Output: a table that matches each model construct to a variable, a data source and an instrument item.
This is the step that turns a borrowed model into your own framework. Illustrative example, not a real study. Suppose your question is why adults at a peri-urban clinic in a South African province do not return for TB screening results, and you build on the Health Belief Model.
| Model construct | Your variable | How it is measured |
|---|---|---|
| Perceived susceptibility | Belief that one could have TB | Three questionnaire items on a five-point agreement scale |
| Perceived severity | Belief that untreated TB is serious | Two items on the same scale |
| Perceived benefits | Belief that returning for results helps | Three items |
| Perceived barriers | Cost of transport, time off work, stigma | Checklist of barriers plus a rank question |
| Cues to action | Reminder call or SMS from the clinic | Yes or no, from the clinic register |
| Control variables | Age, sex, education, distance to clinic | Demographic section |
Notice that every row names a measurable variable, not a concept only. If a construct has no row here, it does not belong in your figure. The site’s guide to the operationalisation of variables table covers how to complete the third column properly.
Step 4: Draw the diagram
Output: a clean figure with boxes, directional arrows and a label on every arrow.
Follow these drawing rules, which examiners check one by one:
- Read left to right. Explanatory factors on the left, the outcome on the right.
- One box per measured variable, with the exact wording used in your instrument.
- One arrow per hypothesis, labelled H1, H2 and so on, with a plus or minus sign if you predict a direction.
- Show moderators and confounders as arrows pointing at an arrow, or as boxes set apart, depending on your design.
- Do not cluster unmeasured ideas. If you did not measure it, leave it out of the figure and mention it in the limitations.
- Number the figure and give it a caption that names the model you adapted and the source.
A standard drawing tool is enough. Legibility matters more than polish. Keep the font readable once the figure is printed at the width of a page.
Step 5: Write the paragraph that explains the figure
Output: one or two paragraphs naming the model, your adaptation, and what each arrow predicts.
Examiners will not accept a figure that stands alone. Write a short passage that does four things: names the model and its source, says which constructs you kept and which you added or dropped, states each hypothesis the arrows represent, and says why the model fits your population. An illustrative opening for the TB example above reads: “This study adapts the Health Belief Model (Janz and Becker, 1984) to explain failure to return for TB screening results among adults at a peri-urban clinic. Perceived susceptibility, severity, benefits and barriers are retained, cues to action are operationalised as a clinic reminder, and distance to the facility is added as a context variable because transport is a documented barrier in the study setting.” Every statement in that passage should be backed by your own literature review.
What do the other two model families look like in practice?
Illustrative example, not a real study. A rural clinic-use question built on Andersen’s model might read as follows. Predisposing factors on the left: age, sex, education, household composition. Enabling factors: household income, distance to the nearest clinic, whether a child-support grant is received. Need factors: self-reported illness in the past month. These feed one behaviour box, “used the clinic in the past three months”, which leads to an outcome such as satisfaction or perceived health, if your design measures it. The hypotheses are then of the form “greater distance is associated with lower clinic use”, one per arrow.
Illustrative example, not a real study. A socio-ecological framework for adolescent substance use would put the learner at the centre and surround them with levels: intrapersonal (attitudes), interpersonal (peers and family), institutional (school rules), community (availability of outlets) and public policy (regulation). Here the diagram is a set of nested rings rather than a left-to-right chain, and each ring is paired with the measures you use at that level. Because nested diagrams invite vague boxes, give every ring only the variables you actually measure.
If your design is a case study or qualitative, the framework sensitises you to concepts to explore rather than hypotheses to test, so label the arrows with the questions the framework raises. The site’s guide to the Chapter 1 of a public health dissertation shows where the framework sits in the introduction, and the guide to researching rural communities in public health shows how population context shapes the variables you add.
Step 6: Check alignment with your instrument and analysis
Output: a one-page check showing every box measured and every arrow tested.
A framework earns its place when the rest of the dissertation follows it. Check that every box appears in your instrument or data source, that every arrow corresponds to a research question or hypothesis, that your analysis plan tests each arrow, and that your results chapter reports in the same order as the figure. Any box you cannot measure should be removed or listed as a limitation. If your framework is built on models that already come with validated scales, say which instrument you use and report its reliability in your own sample.
How do you present the framework to your supervisor before you submit?
Bring a one-page version to your next supervision meeting: the figure, the mapping table from Step 3 and the explanatory paragraph from Step 5. Ask your supervisor three specific questions. Is this the right model for my question? Is any box missing that the literature on my outcome treats as essential? Can I defend each arrow with a source? Changes made at this stage cost an afternoon, while changes made after data collection can cost a chapter. Keep a dated copy of each version, so that you can show the examiner how the framework developed and why you changed it.
Where do students go wrong?
- Copying a published figure unchanged with no link to their own variables.
- Boxes without arrows, or arrows without hypotheses.
- Confusing the conceptual framework with the theoretical framework and presenting only a theory description.
- Constructs that no instrument item measures.
- Citing a textbook for the model instead of its original source.
- A figure that cannot be read when printed.
How can Tesify help with your framework?
Turning a variable list into a mapping table, a labelled diagram description and an explanatory paragraph is organising work where it is easy to lose consistency between figure, instrument and analysis. 9,000+ students have written 15,000+ chapters with Tesify, which helps you structure and organise your dissertation, and the text stays 100% written by you. Organise your framework chapter with Tesify.
Frequently asked questions
What is the difference between a conceptual and a theoretical framework in public health?
A theoretical framework is an existing theory or model you borrow. A conceptual framework is your own diagram of the variables in your study and how you expect them to relate, usually adapted from one or more models.
Can I use the Health Belief Model unchanged?
You can use it as the backbone, but you should adapt it to your outcome and population, and show which constructs you measure and how.
How many boxes should a conceptual framework have?
As many as you measure and no more. Every box should map to an instrument item or a data source.
Do qualitative public health studies need a conceptual framework?
Many do. For a qualitative study the framework sensitises you to concepts to explore, and it is usually held more loosely than a quantitative framework.
Should the framework appear in Chapter 1 or Chapter 2?
Follow your faculty template. It is commonly placed after the literature review or within the theoretical or conceptual framework section.
Can I draw the diagram in Word or PowerPoint?
Yes. A clear diagram in a standard drawing tool is acceptable as long as it is legible, labelled and explained in the text.
Where do I find an existing public health framework to adapt?
Look in the introduction of recent studies on your outcome in your databases, then go back to the original source of the model and cite that source.
References cited
- Andersen, R.M. (1995) Revisiting the behavioral model and access to medical care: does it matter? Journal of Health and Social Behavior, 36(1), pp. 1–10.
- Janz, N.K. and Becker, M.H. (1984) The Health Belief Model: a decade later. Health Education Quarterly, 11(1), pp. 1–47.
- McLeroy, K.R., Bibeau, D., Steckler, A. and Glanz, K. (1988) An ecological perspective on health promotion programs. Health Education Quarterly, 15(4), pp. 351–377.
- Rosenstock, I.M. (1974) Historical origins of the Health Belief Model. Health Education Monographs, 2(4), pp. 328–335.