The funnel approach: how a background chapter should narrow
A useful background chapter moves from the widest relevant evidence to the exact unresolved problem. The sequence is logical rather than ceremonial: every level introduces more specific evidence and explains why the next level deserves attention.
Global context
Explain what is happening with the phenomenon internationally. Establish its scale, development and significance using credible global evidence.
Regional or continental context
Narrow the discussion to the relevant region, such as Africa, Sub-Saharan Africa or emerging markets. Explain how regional institutions, conditions or evidence alter the wider problem.
National context
Present the situation in the study country. Use relevant policy, regulatory, economic, sectoral or institutional evidence rather than simply repeating the global discussion.
Sectoral or industry context
Move into the specific population, industry, institution or environment being investigated. Show how the wider national issue manifests within that setting.
Specific research problem
End with the unresolved contradiction, weakness, consequence or evidence gap that the study will investigate. At this point, the reader should understand why this particular study is necessary.
Defensible gap identification
The statement that no prior study combined a set of variables is not, by itself, a research gap. A defensible gap establishes that the unresolved issue matters and demonstrates how the proposed design can produce useful evidence.
| Gap type | What must be demonstrated | Common weakness |
|---|---|---|
| Evidence gap | Credible studies report conflicting or inconclusive findings that the new study can help clarify. | Listing different coefficients without examining design, context or measurement differences. |
| Contextual gap | Institutional, regulatory or market conditions make transfer from another setting uncertain. | Treating a new country as sufficient novelty without explaining why context changes the relationship. |
| Methodological gap | A recurring design limitation constrains prior conclusions and the proposed method addresses it. | Using a more complex estimator without showing what inferential problem it solves. |
| Measurement gap | A construct has been operationalised too narrowly or inconsistently, with consequences for interpretation. | Changing a proxy without theoretical or empirical justification. |
| Temporal gap | A material policy, economic or institutional change limits the relevance of earlier evidence. | Calling evidence old without identifying what changed. |
Common error
“Few studies have examined these variables together in Nigeria.”
Stronger direction
Explain the unresolved finding, why the Nigerian setting affects the relationship and how the selected design addresses the limitation.
Interpret trends before testing them
A trend discussion should identify the pattern, quantify the important movement, connect it to the study context and avoid claiming causation before the analytical model is applied.
Observe: state the direction and period. Quantify: report the material change. Contextualise: connect it to documented conditions. Reserve judgement: distinguish a plausible explanation from a tested relationship.
This produces a discussion that demonstrates analytical awareness, contextual understanding and preparation for subsequent diagnostic testing.
Use precise academic language
| Avoid | Prefer | Reason |
|---|---|---|
| “This proves that…” | “The result provides evidence consistent with…” | Most observational studies support bounded inference rather than proof. |
| “The result was insignificant.” | “The estimated association was not statistically significant at the specified level.” | Statistical significance is not the same as practical importance. |
| “All studies agree…” | “The reviewed evidence is broadly consistent…” | Absolute claims require exhaustive support. |
| “The researcher used random sampling.” | Name the procedure, frame and selection mechanism. | A label does not demonstrate that probability sampling occurred. |