How to Collect Survey Data for Your Dissertation

Why Surveys Are a Strong Choice for Dissertation Research

Surveys are one of the most practical primary research methods available to students at undergraduate, master's, and doctoral level. They let you gather quantitative or mixed-method data from a defined population, at relatively low cost, within a manageable timeframe.

That said, surveys are not the right method for every research question. They work well when you need to:

• Measure attitudes, perceptions, or behaviours across a sample
• Compare responses across different groups or demographics
• Gather data at scale that would be impractical through interviews alone

They are less suited to questions that require deep contextual understanding, longitudinal observation, or access to sensitive records. Before committing to a survey-based approach, confirm with your supervisor that a questionnaire is the right instrument for your specific research question.

Step 1: Define Your Research Question and Method

The foundation of any survey project is a well-defined research question. Before you design a single question, you need clarity on:

• What you are trying to find out — a specific, answerable question, not a broad topic
• Why a survey is the right method — what your questionnaire can and cannot answer
• What the survey will not answer — being honest about limitations strengthens your methodology chapter

A useful discipline is to write out your hypothesis or research objective in one sentence, then test every survey question against it. If a question does not contribute to answering your research objective, cut it.

Your research design choice — whether cross-sectional, longitudinal, experimental, or descriptive — will shape how you structure your survey and how you later analyse the data. Confirm the design with your supervisor before moving to the next step.

Step 2: Get Ethics Approval

In most universities, any research involving human participants requires institutional ethics approval before data collection begins. This applies regardless of whether your institution calls it an IRB (Institutional Review Board), ethics committee, or research governance process.

Key requirements typically include:

• Informed consent — participants must understand what the study involves, how their data will be used, and that participation is voluntary. Most online surveys include a consent statement on the opening screen.
• Anonymity and confidentiality — your ethics application should specify whether responses will be anonymous (no identifiable information collected) or confidential (identifiable information collected but protected). Know the difference.
• Data storage — ethics committees will ask where data is stored, for how long, and who has access. You need a clear answer. Using a platform with SSL encryption and secure data storage makes this answer straightforward.
• Participant risk — any survey touching sensitive topics (health, identity, trauma, discrimination) may require additional ethical safeguards.

Do not begin distributing your survey until you have written approval in hand. Collecting data without ethics approval is a serious breach of research integrity and may result in your data being unusable.

Step 3: Design Your Questionnaire

Questionnaire design is where most dissertation surveys run into problems. Poor question design produces data that is difficult to analyse and easy to challenge in a viva.

Choose the right question types

• Likert scales — useful for measuring attitudes and agreement (e.g. Strongly Agree to Strongly Disagree). Use an odd-numbered scale (5 or 7 points) unless you have a specific reason not to.
• Close-ended questions — multiple choice, ranking, or binary (yes/no). Produce quantitative data that is straightforward to analyse.
• Open-ended questions — allow free text responses. Useful for capturing nuance but require thematic or qualitative analysis. Use sparingly — too many open-ended questions reduce completion rates.

Principles of good question design

• Ask one thing per question — double-barrelled questions produce ambiguous data
• Avoid leading questions that suggest a desired answer
• Use plain language — avoid jargon, acronyms, or assumptions about background knowledge
• Order questions logically — move from general to specific, and avoid priming effects by placing sensitive questions towards the end
• Include routing or skip logic where necessary — not every respondent should see every question

Pilot your instrument

Always pilot your questionnaire with a small group (5–10 people) before full distribution. A pilot reveals unclear wording, technical issues, unexpected question order effects, and problems with the response scale. Fix these before you launch.

Step 4: Choose Your Sampling Method

Your sampling method determines who participates in your study and how you generalise your findings. Dissertation research typically uses one of two broad approaches:

Probability sampling

Every member of the target population has a known, non-zero chance of selection. This is the gold standard for generalisable findings but requires a defined, accessible sampling frame — often difficult to achieve in student research.

Common types: simple random sampling, stratified random sampling, systematic sampling.

Non-probability sampling

Participants are selected through non-random means. Less rigorous for generalisation but far more practical for most dissertations.

• Convenience sampling — recruiting whoever is readily accessible (e.g. fellow students, social media followers). Easy to execute, but generalisation is limited and must be acknowledged.
• Purposive sampling — selecting participants who meet specific criteria relevant to your research question. Common in qualitative and mixed-methods work.
• Snowball sampling — existing participants recruit further participants. Useful for hard-to-reach populations.

Sample size

There is no universal "correct" sample size for dissertation research. The appropriate number depends on your research design, the statistical tests you plan to run, and the level of precision you need. For most undergraduate and master's dissertations, samples of 30–100 are common for quantitative work. PhD research may require larger samples depending on the methodology.

Discuss sample size with your supervisor. For quantitative analysis, a power calculation is the methodologically rigorous way to determine a minimum sample — statistical software or online power calculators can assist.

Step 5: Choose Your Distribution Channel

How you distribute your survey directly affects response rate, sample quality, and the generalisability of your data.

Online survey link

Most dissertation surveys are distributed as a web link. The link can be shared via email, messaging apps, learning management systems (e.g. Moodle, Canvas), or posted to relevant online communities.

Email outreach

Direct email to a defined list of potential participants typically produces higher response rates than open social media posting, because the appeal is personalised. If your university provides access to staff or alumni email lists for research recruitment, check the conditions carefully.

University participant pools

Some universities operate participant pools or research panels that students can access for their studies. These are particularly common in psychology departments. Check whether your department has one — they can significantly speed up data collection.

Social media recruitment

Platforms such as LinkedIn (for professional samples), Reddit (for specialist communities), and Facebook groups can be effective for reaching specific populations. Be transparent about the research purpose in your recruitment post.

Research ethics note

Whatever channel you use, your recruitment approach must be consistent with your ethics approval. If your ethics application described a specific recruitment method, use that method. If circumstances change, update your ethics documentation.

Step 6: Collect and Monitor Responses

Once your survey is live, your job is not done. Active monitoring during data collection protects the quality and completeness of your dataset.

• Track your response rate — monitor how many responses you are receiving against your target sample size. If uptake is slow, send a reminder to non-respondents (typically after 5–7 days).
• Monitor completion rate — a high drop-off partway through the survey suggests a problem with length, question complexity, or a technical issue on a specific device. Fix these quickly.
• Watch for speeders — respondents who complete a 10-minute survey in 90 seconds are unlikely to have read the questions carefully. Most platforms allow you to flag or exclude these.
• Set a closing date — keep your survey open long enough to reach your target sample, but close it before your analysis deadline. Last-minute responses collected under pressure are harder to process carefully.

Keep a record of when you launched the survey, when you sent reminders, and when you closed it. This information belongs in your methodology chapter.

Step 7: Analyse Your Data

Analysis begins with cleaning your dataset — removing incomplete responses, checking for duplicate submissions, and flagging outliers.

Quantitative analysis

• Descriptive statistics — frequency counts, means, medians, and standard deviations give you a summary picture of your data before you run inferential tests.
• Correlation and regression — for testing relationships between variables. Requires an appropriate sample size and careful attention to assumptions.
• Cross-tabulation — useful for comparing responses across subgroups (e.g. age groups, departments, professional backgrounds).

Most academic researchers export their survey data to a statistical package such as SPSS, R, or Stata for analysis. Ensure your survey platform supports CSV or SPSS-format data export.

Qualitative and mixed-methods analysis

Open-ended responses require thematic analysis — identifying recurring patterns, categories, and themes across the text. This is time-consuming but provides depth that close-ended questions cannot.

AI-powered insights

Modern survey platforms can generate a preliminary narrative summary of your results — identifying key findings, patterns, and areas of divergence across your dataset. This is a useful starting point for your analysis, but it does not replace your own methodological interpretation. AI-generated summaries should be treated as a first-pass orientation, not a final analysis.

Common Dissertation Survey Mistakes

• Leading questions — questions that nudge respondents toward a particular answer contaminate your data and undermine the validity of your findings.
• Under-sized samples — collecting 15 responses when your analysis requires 60 produces unreliable results. Agree your target sample size before you launch.
• Poor ethics documentation — vague or incomplete ethics applications slow approval and may require resubmission. Be specific about what data you are collecting, how it will be stored, and how long it will be retained.
• Not piloting — skipping the pilot means avoidable errors reach your full dataset. Pilot testing takes a few days and can save weeks of remediation.
• Launching too late — survey distribution takes longer than students expect. Allow for slow early uptake, reminder cycles, and the possibility of needing to extend your recruitment period.

How onlinesurvey.ai Supports Dissertation Research

onlinesurvey.ai is built for researchers who need more than a basic form builder. Key features relevant to dissertation work:

• AI question generation — describe your research goal in plain language and the platform generates a structured, methodologically appropriate questionnaire. Useful for getting a first draft quickly, which you then refine with your supervisor.
• Skip logic and branching — route respondents to different questions based on their answers. Available from the Basic plan (3 rules) and unlimited on Pro.
• Response capacity — the Pro plan supports up to 5,000 responses per month, which is more than sufficient for most dissertation and thesis research projects.
• AI-powered insights — once data is collected, the platform generates a narrative summary of key findings, patterns, and concerns with confidence context. A useful starting point before you move to deeper statistical analysis.
• Data export — export raw data in formats compatible with SPSS and other statistical tools.
• Secure storage — data is encrypted in transit and at rest. Response data is not used to train AI models — an important consideration when collecting data under an ethics approval.

The Pro plan at $49/month is priced accessibly for postgraduate researchers. For institutional access, the Enterprise plan supports university-wide deployment with SSO and role-based access control.