Dissertation Sample Size Help: How to Calculate and Justify Your Sample for a UK Dissertation Guide

Last updated: November 2026 · Reviewed by the HND Assignment Help tutoring team

Sample size is one of the questions UK dissertation students ask most often, and one of the places supervisors most often write "justify this" in the margin. This guide is for undergraduate (Level 6), master’s (Level 7) and MBA dissertation students at UK universities. It shows you how to work out a sensible number, how to explain it in your methodology chapter and how to talk about it honestly when you fall short. Your own university handbook and supervisor have the final word on what is expected.

Dissertation sample size at a glance

Item	What to know
Where it appears	Methodology chapter, usually under "Sampling" or "Participants", and again in Limitations.
Quantitative surveys	Justify with a margin-of-error calculation, adjusted for a small population and your expected response rate.
Experiments and group comparisons	Justify with a power analysis (effect size, alpha, power), often done in free G*Power software.
Interviews and focus groups	Justify with information power, the depth of each interview, homogeneity of the sample and what the time allows.
Secondary data	Explain the dataset, the inclusion rules and how many cases or documents remain after screening.
What earns marks	A clear rationale linked to your research question, honest limitations, and consistency with your analysis method.
Check your brief	Word limit, ethics rules and any minimum participant numbers set by your department.

How to approach sample size, step by step

1. Start from your research question and method

The right sample size depends on what you are trying to show. A survey describing attitudes needs enough responses for a stated margin of error. A study testing whether two groups differ needs enough people to detect a realistic difference. An interview study exploring experiences needs enough rich conversations to answer the question in depth. Write one sentence that links your question to the type of evidence you need before you pick any number.

2. Define the population and sampling frame

The population is everyone your findings are about (for example, "front-line staff at a UK retail chain"). The sampling frame is the list you can actually reach (for example, "staff on the company’s internal mailing list"). Markers look for this distinction because it explains possible bias.

3. Choose a sampling technique and name it

Technique	Best for	What to say about it
Simple random	Surveys with a full list of the population	Every member has an equal chance; strongest for generalising.
Stratified	Populations with clear sub-groups (e.g. departments)	Ensures each sub-group is represented in proportion.
Convenience	Student projects with limited access	Quick and cheap, but limits generalisation; say so in Limitations.
Purposive	Interviews with people who hold specific knowledge	Participants chosen for relevance to the question.
Snowball	Hard-to-reach groups	Participants refer others; risk of similar views.

4. Calculate the number (quantitative)

For a survey estimating a proportion, the common formula (often attributed to Cochran) is n = Z² × p(1 − p) / e², where Z is 1.96 for 95% confidence, p is the expected proportion (use 0.5 if unknown, because it gives the largest sample) and e is the margin of error. For smaller populations, apply the finite population correction: n_adj = n / (1 + (n − 1) / N).

Worked method example (not an answer, invented numbers): a student surveys staff at an organisation of 2,000 employees, with 95% confidence and a 5% margin of error.

• n = 1.96² × 0.5 × 0.5 / 0.05² = 3.8416 × 0.25 / 0.0025 = 384.16
• Finite correction: 384.16 / (1 + 383.16 / 2,000) = 384.16 / 1.1916 ≈ 322.4, so round up to 323 completed responses
• If the student expects a 40% response rate, they need to invite 323 / 0.40 ≈ 808 people

If the student only gets 150 responses, they should not hide it. They report the achieved margin of error (rework the formula for e) and discuss what that means for confidence in the findings.

For experiments or group comparisons, use a power analysis instead. With the conventional settings (alpha 0.05, power 0.80), detecting a medium effect (Cohen’s d = 0.5) with an independent-samples t-test needs about 64 participants per group; detecting a correlation of r = 0.3 needs about 84 participants. G*Power will give you these figures and a screenshot you can put in an appendix. For multiple regression, a widely quoted rule of thumb (Green, 1991) is at least 50 + 8m participants, where m is the number of predictors.

5. Justify the number (qualitative)

There is no formula for interviews. Instead, explain why your number is adequate. Useful ideas to cite and apply include information power (Malterud and colleagues, 2016), which says fewer participants are needed when the aim is narrow, participants are highly relevant and the dialogue is strong, and data saturation, though note that Braun and Clarke argue saturation does not fit reflexive thematic analysis. Many UK master’s projects use roughly 6 to 15 interviews; treat that as a range seen in practice, not a rule, and link your number to your own design.

6. Write it up in your methodology chapter

Paragraph	Template sentence starter
Population and frame	"The target population was …; participants were accessed through …"
Technique	"A [purposive / stratified] sampling approach was used because …"
Size and calculation	"Using a 95% confidence level and 5% margin of error, the required sample was … (see Appendix X)."
Achieved sample	"[N] usable responses were received after removing [N] incomplete responses …"
Limitations	"Because the sample was …, findings may not generalise to …"

How assessors judge Dissertation sample size

What the marker looks for	Weak work	Strong work
Link to research question	Number appears from nowhere	Number follows from the question and method
Sampling technique	Not named, or named but not explained	Named, justified and its bias acknowledged
Calculation or rationale	Missing or copied without workings	Formula or power analysis shown, or qualitative rationale with sources
Achieved vs planned sample	Shortfall ignored	Response rate reported; achieved margin of error or impact discussed
Consistency with analysis	Complex tests on tiny groups	Analysis suits the sample (e.g. descriptive stats or thematic analysis)
Ethics	No mention of consent or access	Consent, anonymity and gatekeeper approval described

Pre-submission checklist

• Population, sampling frame and technique are each named and defined.
• The number is justified with a calculation or a referenced qualitative rationale.
• Workings (or a G*Power screenshot) are in an appendix and referred to in the text.
• Planned and achieved sample sizes and the response rate are both reported.
• Your statistical or qualitative analysis is realistic for the sample you actually have.
• Limitations explain how the sample affects generalisation.
• Ethics approval, consent and data protection are mentioned in line with your university’s policy.
• Every source (Cochran, Cohen, Malterud, Braun and Clarke) is in your reference list in your required style.

Dissertation sample size answer hints and sample structure

There is no answer to copy here, because your sample depends on your own question, access and method. What follows is a typical structure for the sampling part of a methodology chapter in a 10,000 to 15,000-word dissertation. Adjust it to your handbook.

Section	What goes in it	Suggested share
Population and sampling frame	Who the findings are about and who you could actually reach	About 100 words
Sampling technique	Named technique, why it fits, its main bias	100–150 words
Sample size justification	Formula or power analysis, or qualitative rationale with sources	150–250 words
Recruitment and response	How people were contacted, response rate, exclusions	100–150 words
Ethics	Consent, anonymity, approval, data storage	100 words
Limitations (in methodology or conclusion)	Effect of the sample on generalisation and analysis	100–150 words

Frequently asked questions

What is a good sample size for a UK undergraduate dissertation survey?

There is no single number. Many undergraduate surveys collect between 50 and 200 responses. What matters is that you explain how you reached your number and what margin of error or limitation follows from it.

How many interviews do I need for a master’s dissertation?

It depends on your aim and design. Many projects use roughly 6 to 15 interviews, but you should justify the number using ideas such as information power and the depth of each interview, not just quote a range.

What if I do not get enough survey responses?

Report the number you achieved, recalculate the margin of error, use analysis suited to that sample and discuss the limitation honestly. Supervisors expect this to happen in student research.

Do I need G*Power for my dissertation?

Only if you are testing differences or relationships and your department expects a power analysis. G*Power is free and widely used, and a screenshot of the settings makes a clear appendix.

Can I use convenience sampling?

Yes, many student projects do. Name it, explain why it was practical and state clearly that findings may not generalise to the wider population.

Where should the sample size calculation go?

Summarise it in the methodology chapter and put the full workings or software output in an appendix that you refer to in the text.

Does a bigger sample always get higher marks?

No. Markers reward a sample that fits the question, is clearly justified and is analysed appropriately. A large sample with no rationale or poor analysis will not score well.

Related help

• Dissertation methodology chapter help: UK step-by-step guide
• How to design a questionnaire for a UK dissertation
• Statistics and SPSS tutoring
• UK dissertation tutoring and feedback