Dissertation Sampling Strategy: Methods, Sample Size and Justification

A dissertation sampling strategy explains who you will study, how you will select them and why that number is enough. Committees scrutinize it because a weak sample limits every conclusion that follows.

This guide covers the key terms, the main probability and non-probability methods, how to calculate sample size with a power analysis, how to justify qualitative sample size, and how to write the section clearly. Conventions differ between disciplines, so check your program's expectations too.

Key Terms First

Be explicit about the gap between target and accessible populations. It is the main limit on how far your findings generalize.

Probability Sampling

In probability sampling, every member of the sampling frame has a known, non-zero chance of selection. It supports statistical generalization to the population, which is why it is the ideal for many survey studies.

• Simple random: every member has an equal chance, for example through a random number generator.
• Systematic: every kth member from a list, after a random start.
• Stratified: the population is divided into groups (strata) such as job grade, then sampled within each.
• Cluster: whole groups, such as schools or wards, are randomly selected, then everyone or a sample within them is studied.

Cluster samples need analysis that accounts for clustering, since people in the same ward or school tend to be similar. Mention this if you use one.

Non-Probability Sampling

Most dissertations, especially qualitative and practice-based ones, use non-probability methods. They are legitimate when justified by the aims, but they limit statistical generalization.

Purposive sampling has its own variants, such as maximum variation, homogeneous, typical case and critical case sampling. Naming the specific variant shows the committee you chose deliberately.

Sample Size for Quantitative Studies: Power Analysis

A power analysis estimates the sample needed to detect an effect of a given size, if it exists. It needs four inputs: the statistical test, the significance level (alpha, usually .05), the desired power (commonly .80) and the expected effect size.

Free software such as G*Power is widely used. Base the effect size on previous studies or a meta-analysis where possible; Cohen's conventional values (for example d = 0.2, 0.5 and 0.8 for small, medium and large) are a fallback, and some reviewers dislike them without justification.

Different tests need different calculations. A multiple regression with three predictors, a medium effect (f² = 0.15), alpha .05 and power .80 needs about 77 participants, for example. Always run the calculation for your own planned analysis.

Sample Size for Qualitative Studies

Qualitative studies justify size by depth and purpose, not statistical power. There are no universal minimums, and published guidance varies.

• Saturation: continue until new data add no new codes or themes. Report how you judged it.
• Information power: fewer participants are needed when the aim is narrow, the sample is highly specific, theory is well established, dialogue is rich and the analysis is focused on a few cases.
• Tradition norms: interpretative phenomenological analysis often uses small, homogeneous samples; case studies may involve one or a few cases with several data sources.

Plan a provisional range, for example 10 to 16 interviews, and explain how you will decide when to stop. Committees prefer a reasoned range over a single number pulled from another dissertation.

Mixed Methods Sampling

Mixed methods studies need a sampling plan for each strand and a statement of how the samples relate. Common options include identical samples (the same people in both strands), nested samples (a subset of survey respondents interviewed) and parallel samples (different people from the same population).

In an explanatory sequential design, a nested sample is common: you select interviewees based on their survey results, such as high and low scorers.

Recruitment, Access and Bias

Describe exactly how people will be approached: who sends the invitation, through what channel and what they receive. Gatekeepers, such as managers or headteachers, often control access and may need to give written permission.

Recruiting from Your Own Workplace

It is common in practice doctorates, but it raises power and confidentiality issues. Ethics committees usually expect safeguards, such as an independent person sending invitations and clear assurance that declining has no consequences.

Sources of Sampling Bias

• Self-selection: volunteers may differ from non-volunteers.
• Non-response: those who decline may share characteristics.
• Coverage: the sampling frame misses part of the population.
• Survivorship: only those still present, such as staff who stayed, are studied.

Writing the Sampling Section

A clear sampling section usually follows this order: population and setting, sampling method and rationale, inclusion and exclusion criteria, sample size and justification, recruitment procedure, and ethical safeguards.

In the results chapter, report the achieved sample, response rate and any differences from the plan. If you fell short of your target, say so and discuss how it affects power or depth.

Common Committee Objections

• "Why this number?" Give a power analysis or qualitative rationale, not a round number.
• "Convenience sampling limits generalisability." Acknowledge it and explain why it fits the aims.
• "Your criteria are unclear." List inclusion and exclusion criteria explicitly.
• "Is access confirmed?" Show gatekeeper support where your program allows.
• "Your power analysis does not match your analysis." Run it for the test you will actually use, including the number of predictors or groups.
• "The sample cannot answer the question." If you ask about managers, sample managers, not only their staff.

Most of these objections are easy to prevent. Read your sampling section as a skeptical reviewer would, and make sure every number and choice has a stated reason.

How DissertationHorse Helps with Your Sampling Strategy

Send your research questions, design, target population, access details and planned analysis. A writer experienced in research methods prepares a custom sampling section, including a power analysis or qualitative size rationale, recruitment plan and bias discussion, for you to study and adapt.

We never invent participants, response rates or data. Free revisions within the scope of your original request are included.