Sampling is where the gap between research intention and research validity is most often created. A study with a perfect research question and a rigorous data collection instrument can produce entirely unreliable findings if the sample is wrong. Nigerian researchers make specific, identifiable sampling mistakes that this guide addresses directly.

Sampling is not just a technical decision about numbers. It is a conceptual decision about whose experience, whose data, and whose perspective your research will represent. Getting it wrong means your findings generalise to people you did not study rather than to the people you intended to understand. In a country as diverse as Nigeria, where experience varies dramatically by state, ethnicity, gender, income level, and urban-rural location, sampling decisions are particularly consequential.

Probability vs Non-Probability Sampling: The Core Distinction

Probability sampling gives every member of the target population a known, non-zero chance of being selected for the study. Non-probability sampling does not. This distinction matters for generalisation: findings from a probability sample can be generalised to the broader population from which the sample was drawn, within the bounds of sampling error. Findings from a non-probability sample cannot be statistically generalised, though they may be analytically transferable in qualitative research where generalisation means something different.

The choice between probability and non-probability sampling is determined primarily by your research purpose and your access to a sampling frame. A sampling frame is a list or enumeration of all members of the target population from which a sample can be drawn. If a reliable sampling frame exists and you want to generalise statistically, probability sampling is appropriate. If no reliable sampling frame exists, or if your research purpose is understanding rather than measurement, non-probability sampling is appropriate.

Probability Sampling Methods

Simple random sampling assigns every population member a number and selects the required sample size randomly from those numbers. It is conceptually simple and statistically rigorous but requires a complete sampling frame and is logistically challenging for large, geographically dispersed Nigerian populations. Systematic sampling selects every nth member from a ordered list, which is simpler to implement but requires a truly random starting point to maintain probability properties.

Stratified random sampling divides the population into subgroups based on relevant characteristics, such as gender, geographic zone, or income level, and samples randomly within each subgroup. This approach is particularly valuable for Nigerian research because it ensures that important subgroups within the diverse Nigerian population are represented in proportions that either reflect their real population proportions (proportionate stratification) or ensure adequate representation of smaller but analytically important groups (disproportionate stratification).

Cluster sampling divides the population into natural clusters, such as schools, health facilities, or local government areas, randomly selects clusters, and then surveys all or a random sample of members within selected clusters. This is frequently the most practical probability sampling approach for nationally representative Nigerian research, as it reduces the logistical burden of studying a geographically dispersed population. The Nigerian National Demographic and Health Surveys use a multi-stage cluster sampling design for exactly this reason.

Non-Probability Sampling Methods

Purposive sampling selects participants based on specific characteristics that make them particularly informative for the research question. A study of the experience of first-generation university students in Nigeria would purposively select students who are the first in their family to attend university, not a random sample of all students. Purposive sampling is the standard approach for qualitative research and is legitimate and rigorous when the selection criteria are explicitly justified.

Snowball sampling recruits initial participants through purposive selection and then asks those participants to refer other suitable participants from their networks. It is particularly useful for studying hard-to-reach populations in Nigeria: communities with specific practices, individuals with rare experiences, or groups that are not accessible through standard institutional channels. The limitation is that snowball samples tend to be homogeneous within social networks, which can bias the sample toward participants who know each other.

Convenience sampling recruits whoever is most accessible: fellow students, colleagues, or people in a specific location. It is the most common sampling approach in Nigerian student research and the one that creates the most validity problems. Convenience samples are not inherently invalid, but the researcher must be explicit about their limitations and must not make generalisation claims that the sample cannot support.

Sample Size: How Much Is Enough?

Sample size is one of the most anxiety-producing questions in Nigerian student research and one of the most frequently answered incorrectly. The correct answer depends entirely on your research design, your statistical analysis approach, your expected effect sizes, and your acceptable level of precision. There is no universal minimum sample size that applies across all research designs.

For quantitative research with statistical hypothesis testing, power analysis is the appropriate method for sample size calculation. Power analysis uses the expected effect size, desired statistical power (typically 0.80), and significance level (typically 0.05) to calculate the minimum sample needed to detect a real effect if one exists. Software like G*Power (free) performs these calculations. Research that reports "a sample size of 384 was used following Yamane's formula" without specifying the effect size and power assumptions has not justified its sample size; it has selected a number from a table without engaging with the underlying logic.

For qualitative research, sample size is determined by theoretical saturation rather than statistical calculation. You continue sampling until new data no longer produces new insights or themes. In practice, most qualitative studies in Nigerian student projects reach saturation between 15 and 30 purposively selected participants, though this varies significantly by research topic and design.

Nigerian-Specific Sampling Challenges

The absence of current, reliable population registers for most Nigerian target populations makes true probability sampling extremely difficult for most student researchers. The 2006 National Population Census remains the most recent complete enumeration, and significant population movement since then means its sampling frames are outdated for many applications. The practical implication is that most Nigerian student research necessarily uses non-probability sampling, which is methodologically defensible as long as it is transparently acknowledged and appropriately justified.

Researcher positionality in Nigerian research contexts matters for sampling. A researcher studying a community they belong to, or one that has complex relationships with academic institutions, needs to consider how their identity and relationships affect who agrees to participate, what participants say, and how the data is collected. These positionality effects should be acknowledged and managed as part of sampling design, not ignored.

"Your sample is not just a number on a methodology form. It is the boundary of your findings. Define it clearly, justify it honestly, and acknowledge its limitations explicitly. Your examiners are not looking for a perfect sample. They are looking for a researcher who knows exactly what their sample can and cannot support."

Project and Research at projectandresearch.com provides sampling calculators, methodology frameworks, and a research community for Nigerian students navigating sampling decisions. The peer review function is particularly useful for getting feedback on your sampling justification before your supervisor sees it.

Platform that has it all

Sampling done right is sampling defended confidently. Get the framework and the community support at projectandresearch.com.

Project and Research