Primary data is collected specifically for your study; secondary data already exists and is reused for your question. The important decision is whether the evidence fits the research question, not which label sounds more impressive. A project can use either or both when its design and permissions support that choice.
Statistics Canada explains that it gathers information directly and also uses existing sources. That is an example of different evidence routes, not permission for a student to access confidential records or a rule for Nigerian project chapters.
Keep the source of your data clear from the methodology to the conclusion. UniDraft helps organise your confirmed sources, actual methods and findings into an editable project document. Start your project documentation with your verified evidence and instructions, then check that the draft describes what you actually collected or reused. Explore the project workflow; the software does not obtain private records or run your analysis.
Classify the source by what happened
These are fictional teaching examples, not completed studies:
| Planned evidence | How to describe its source |
|---|---|
| Responses collected using your approved questionnaire for this study | Primary data, with the actual collection and selection process explained |
| Measurements you take under an approved project procedure | Primary data, with instruments, conditions and procedure documented |
| A published dataset downloaded from its original provider | Secondary data; identify the provider, version and original collection context |
| Existing organisational records released for an approved analysis | Secondary data, with permission, definitions and access restrictions checked |
Your own notes about a placement are not automatically a research dataset. A literature review also does not become an empirical dataset merely because it cites several papers. Explain the role each source plays in the actual assignment.
Check an existing dataset before choosing it
Use this original suitability worksheet:
| Check | What you need to find |
|---|---|
| Provenance | Who collected or produced it, and where the original documentation is |
| Purpose | Why it was collected and what the records represent |
| Unit | A person, visit, transaction, school, device, year or something else |
| Variables | Definitions, categories, measurement units and coding notes |
| Coverage | Population, place, period and exclusions |
| Quality | Known missingness, revisions and collection limitations |
| Access | Conditions of use, permissions and restrictions on sharing |
| Fit | Which part of your question these records can actually answer |
If documentation is missing, record the uncertainty instead of supplying an invented definition. Search for a data dictionary, methodology note or provider contact. A spreadsheet's column name alone may not tell you how a measure was obtained.
Work through a fictional mismatch
Suppose a proposed project asks about the experiences of current students using a registration service. The available file contains annual totals of completed registrations.
Those totals might describe recorded activity. They would not reveal why individual students encountered difficulties or what they thought of the service. You would need a different question or a justified additional evidence source.
Now suppose you also plan interviews. Do not claim the interviews “validate” the totals merely because both concern registration. Explain the different question each source addresses. The research-design worksheet helps connect the question, evidence and conclusion.
Primary does not mean automatically accurate
New collection can still have unclear questions, poor measurement, selection problems or incomplete responses. An existing dataset can be valuable while still having limits for a new question. Evaluate the actual evidence rather than ranking quality by its label.
For your own collection, keep the approved instrument, dated collection record and a clear account of changes. For reuse, retain the original file and documentation alongside a working copy. Record transformations in a cleaning log; do not overwrite the only original.
Describe your actual contribution
For a discipline-specific planning example, the Economics seminar topic guide connects questions to official price and labour-market documentation. It shows why finding a published table is only the start: its definitions, coverage and period still need to fit the question.
Use this original writing checklist in your methodology chapter:
- State whether you collected new observations, reused existing records or did both.
- Identify who originally collected each source and for what purpose.
- Explain selection, dates, definitions and access arrangements.
- Describe what you changed, excluded, recoded or analysed.
- State what the source cannot establish for your question.
Fictional wording pattern: “This project proposes to analyse [identified existing dataset/version] to examine [bounded question]. The provider's documentation defines [relevant unit or measure]. The planned analysis excludes [justified exclusion] and cannot address [missing information].”
Replace the brackets with verified details. Do not invent a dataset, collection date, sample size or permission. Cite both the dataset and relevant methodological publications according to your required style; see the references guide.
If your seminar uses published financial statements, the Accounting topic and report-source worksheet helps record the entity, year, units and supporting notes. It includes a verified filing example without treating company disclosures as student-collected fieldwork.
UniDraft helps organise your written project around confirmed sources, methods and results. It does not collect observations, grant access to private records or turn an unsuitable dataset into evidence for a different question.
If you are choosing a spatial or environmental question, the Geography seminar topic guide provides a worksheet for checking map dates, scale, quality information and what a source can establish.
For a question about firms or business decisions, the Business Administration seminar topic guide shows why loan status, reported obstacles and business performance need distinct evidence.
For an environmental question, the Environmental Science seminar topic guide shows why questionnaire responses about collection services cannot substitute for waste-mass measurements or a facility assessment.
For a social question, the Sociology seminar topic guide illustrates why household definitions and survey exclusions matter before applying findings to university students.
For election or participation evidence, the Political Science seminar topic guide shows why a national aggregate, a campus observation and a policy document answer different questions, with a feasibility worksheet.
If your assignment concerns historical records rather than a newly collected dataset, consult the History seminar source-planning guide. It explains how to check a catalogue lead, access status and the limits of the material actually consulted.
