Choose an Accounting project topic by checking what the observation represents, how each measure is defined and which evidence you can actually obtain. A title promising an effect over ten years is not a research plan until the entities, reporting periods, variables and design are justified.
UniDraft's free topic generator can help build a starting shortlist. Test each suggestion against the evidence checks below before taking it to your supervisor.
Your topic should lead to a document you can explain from source to conclusion. UniDraft helps organise your confirmed Accounting question, actual analysis, checked sources and supervisor instructions into an editable Word project. Start your Accounting project document, bringing the evidence behind the figures and reviewing every calculation and claim yourself. See the project document workflow for what to supply and what you receive.
Choose an evidence route before a title
These are original planning directions, not approved titles or completed studies.
| Route | What must be available | A limit to keep visible |
|---|---|---|
| Public-report investigation | Relevant statements and notes, a defined entity set, reporting periods and measure definitions | Published figures do not automatically identify a causal effect |
| Disclosure comparison | Defined documents, inclusion rules and explicit criteria for examining disclosures | A disclosure checklist does not by itself establish fraud or overall reporting quality |
| Authorised process investigation | Permission and suitable evidence about a specific accounting process | Public annual reports cannot substitute for private transaction records |
| Participant-based investigation | An agreed design, appropriate respondents, instrument and permissions | Reported practice is not automatically observed performance |
For the broader brief, access and deadline checks, use the general project-topic selection guide. A source-reading seminar has a different task; compare the Accounting seminar guide rather than relabelling its discussion as a completed project investigation.
For example, a disclosure comparison could ask how selected reports describe a defined accounting estimate, while a participant study could ask how an authorised group describes a specific recordkeeping practice. These are provisional question shapes: select the documents or participants only after checking access, criteria and your brief. They are not findings or ready-to-submit titles.
Open a real report before promising a dataset
The official Access Bank interim financial statements for June 2025 illustrate the metadata check: the document distinguishes Group and Bank columns, contains an interim reporting period and specifies presentation units. Inspect each relevant statement and its notes before extraction. One report does not establish historical coverage across banks or years.
Build an original source register with the exact entity name, document title, reporting end date, publication/version details, statement or note reference and retrieval date. Record whether the figure describes a position at a date or activity over a period. Check the currency and scale printed beside the actual table.
For example, a closing balance and a period's income have different time meanings. Do not silently turn an interim figure into a full-year observation or mix consolidated and separate accounts. Keep your proposed observation rule consistent, and record how changes or restatements would be handled with your supervisor.
Fictional example: a ten-year title with one report
This is an invented planning exercise. No bank dataset has been assembled, calculations performed or findings supplied.
A student proposes “The effect of liquidity on profitability in every Nigerian bank over ten years.” They have opened one interim report. Neither “liquidity” nor “profitability” has an agreed operational definition.
The next step is a feasibility note, not a regression table. The student must establish the intended entities, periods and measures, then check the reports needed to obtain them. “Every bank” and “ten years” remain unverified coverage promises. The word “effect” also needs a design capable of supporting the proposed causal claim; an observed relationship alone is insufficient.
| Decision | Status in the fictional exercise | Required next step |
|---|---|---|
| Observation | Bank-year proposed | Specify the exact entity, scope and period rule |
| Coverage | One interim report available | List required reports and identify actual gaps |
| Liquidity measure | Undefined | Justify the construct and its source fields |
| Profitability measure | Undefined | Define the measure and reporting-period basis |
| Comparability | Not checked | Examine scope, units, definitions and changes |
| Design | Causal wording unsupported | Agree a question and design the evidence can support |
| Decision | Candidate only | Discuss access and revise before commitment |
A provisional question about an association between defined measures may be worth investigating. It is not automatically feasible either: the entity set, coverage, calculation basis and method still need justification. Use the research-design guide for the relationship between the question and evidence.
Make a variable dictionary and coverage register
For each proposed variable, record its meaning, source statement or note, exact field, unit, time basis, calculation rule where applicable and unresolved ambiguity. Do not select a proxy merely because its name resembles the concept in the title.
Keep a separate coverage register:
| Entity and period | Required source | Obtained? | Relevant fields found? | Unresolved issue |
|---|---|---|---|---|
| Fill from your proposed scope | Exact report/version | Record actual status | Record actual status | Missing page, changed definition or another verified issue |
This is a blank planning structure, not research data. Missing, not applicable and not disclosed must not become zero by default. A downloaded report is not necessarily a usable observation. If coverage fails, discuss a narrower scope or different question; do not manufacture the missing rows.
Take a decision note to your supervisor
Bring the proposed question, evidence route, observation definition, source register, variable dictionary and unresolved gaps. Explain which claim the evidence could support and which claim would need additional evidence or another design.
If the assignment concerns your placement instead, use the Accounting and business SIWES guide. Do not present a public-report exercise as workplace activity you performed.
Once the question and method are agreed, keep the analysis and document connected. Prepare your project with UniDraft using your confirmed work and instructions, then check every source, figure and interpretation. The software prepares documentation; it does not supply private records, recruit respondents, execute the analysis or guarantee approval.
Source
- Access Bank Plc: June 2025 interim financial statements, used only to illustrate reporting metadata. No financial figures or findings are reproduced here.