Choose a data collection method by identifying what you need to know and what record would let you answer it. A questionnaire can collect reported answers; it cannot automatically measure actual system speed or explain every reason behind a response. Write the collection plan before recruiting people or opening an online form.
Statistics Canada's collection guide discusses interview and self-completion modes, observation and measurement, and factors affecting the choice. Use those distinctions as background. Your supervisor and institutional process determine what is appropriate for your study.
A collection plan is most useful when the eventual report can explain what you actually did and what the evidence establishes. UniDraft helps organise your confirmed method, records and results into an editable project document. Start your project documentation, then review the method and findings against your study records. It does not recruit participants, collect data, run tests or provide ethics approval.
Compare the evidence you would obtain
This original planning table gives questions to ask, not instructions to conduct a specialised procedure:
| Route | Record you might obtain | Question to settle before using it |
|---|---|---|
| Self-completed questionnaire | Answers to a defined instrument | Can participants understand the items without clarification? |
| Interview | Recorded or written accounts under an agreed procedure | What follow-up is permitted, and how will accounts be documented and analysed? |
| Observation | Notes or coded records of specified events | What is being observed, how consistently, and with whose permission? |
| Measurement or system test | Values or test outcomes under defined conditions | What does the instrument measure, and how are conditions recorded? |
| Existing records | Previously created data or documents | What did the original collection capture, and is reuse permitted? |
An interview may administer fixed questions, while another interview approach may explore experiences more openly. Explain what you will actually do. For the wider strategy, use the research-design guide; for existing records, check primary and secondary data.
For a bounded case study, the case protocol and evidence worksheet helps connect sources to one defined case and examine differences between accounts, records and observations. Collecting several kinds of material does not remove the need for that analysis plan.
Turn each objective into a collection decision
For a structured observation task, define the event and what counts as an opportunity before making boxes to tick. The observation-checklist worksheet separates not observed, not applicable and unable to determine. If interaction between participants is the evidence you need, the focus-group planning guide explains how a facilitated group differs from individual interviews.
If interviews fit the question, build a guide before recruitment. The interview-guide worksheet connects an objective to an opening question, neutral probes and the evidence an answer could provide. It includes wording checks and a pilot-change record; the fictional questions are not a validated instrument to copy into every study.
Fill in this original worksheet for every objective:
| Field | Your answer |
|---|---|
| Objective | What the project intends to establish |
| Required evidence | An answer, account, observation, measure or record |
| Source | Who or what can provide it |
| Instrument | The approved items, guide, record sheet or test specification |
| Selection | How the people, cases, records or materials are chosen |
| Procedure | What happens, in what order and under which conditions |
| Permission and privacy | Required review, access, consent and safe handling |
| Analysis link | How this record will address the objective |
If a row has no analysis link, collecting more information may simply create work and privacy risks. If a key objective has no evidence source, revise the plan before collection.
If the plan combines quantitative and qualitative evidence, use the mixed-methods joint-display exercise to work out how the parts will connect. Collecting a questionnaire and interviews is not enough unless their relationship and combined interpretation are justified.
A fictional example: reported use versus tested function
Imagine a student is developing a room-allocation prototype. An objective concerns whether a duplicate allocation is prevented under specified conditions.
A satisfaction questionnaire would not directly test that rule. A defined test case would be needed, with expected behaviour, actual behaviour and retained evidence. Conversely, a passed functional test would not establish whether real users understand the interface.
These are different evaluation questions. If both are part of the approved project, give each its own justified procedure. No test, survey or user finding is being asserted in this example.
Test the process before the main collection
Use the pilot-study decision worksheet to identify the uncertainty being tested, the record to keep and the change it could justify. Instrument wording is only one possible issue; recruitment, recording and export procedures may also need a permitted trial.
Before collecting real study records, check what your institution permits for piloting or pretesting. A small test can reveal confusing instructions, missing options or delivery problems. It does not automatically establish that an instrument is valid or reliable.
Useful checks include:
- Do people understand the wording as intended?
- Does the record capture the needed unit and measurement?
- Can the chosen audience actually access the collection mode?
- Are required information and consent steps clear?
- Can records be exported, stored and interpreted without losing their meaning?
- Is there a procedure for missing, inconsistent or withdrawn records?
Keep trial and real collection clearly separated. The questionnaire design guide, Google Forms delivery guide and validity/reliability guide cover different parts of that preparation.
Keep a collection log you can explain
Record the approved instrument version, collection period, selection procedure, relevant conditions, interruptions and deviations. Use participant codes where appropriate and store identifying information according to the approved plan. Do not put names or private responses into a public appendix merely to prove collection happened.
If access fails or the method changes, discuss the change and record what was actually done. Never invent respondents, replace missing observations with imagined values or describe planned collection as completed work. Carry limitations into the report and defence.
UniDraft helps structure your project report around the collection and results you confirm. It does not recruit respondents, perform tests or supply ethics approval.
