How Many Participants Should a Research Study Have? What Is the Right Sample Size?

How many participants should a research study have? This is one of the most common questions researchers face when planning data collection. It is not unusual to hear that a study should have at least 100, 200 or approximately 400 participants. In reality, however, there is no single Sample Size that can be considered appropriate for every research study.

An appropriate Sample Size depends on several factors, including the population size, research objectives, sampling method, desired level of precision, number of groups to be compared and the statistical methods that will eventually be used to analyse the data.

Before deciding how many responses to collect, researchers should therefore consider what the study aims to answer, who the target population is and how the resulting data will be analysed. All of these factors can influence the appropriate Sample Size.

What Is Sample Size?

Sample Size refers to the number of units or participants selected for a study. These may include consumers, employees, students, patients, households or organisations.

For example, a study may aim to investigate customer satisfaction among bank customers in Bangkok. If the bank has hundreds of thousands of customers, collecting data from every customer would be impractical. A sample of customers is therefore selected for the study.

The important question is how many participants are enough.

Sample Size is not determined by population size alone. It should also be considered in relation to the research objectives, data collection method and planned statistical analysis.

Does a Larger Sample Size Always Mean Better Research?

In general, a larger Sample Size can improve the precision of statistical estimates and, in hypothesis-testing studies, may increase Statistical Power or the ability to detect an effect that genuinely exists.

However, this does not mean that collecting as many responses as possible will always produce better research.

If an appropriate Sample Size for a particular study is approximately 300 participants, increasing this to 3,000 may require considerably more time and budget without providing a proportional benefit to the Research Objective.

Conversely, a Sample Size that is too small may result in estimates with greater uncertainty or insufficient Statistical Power to detect the relationship or difference being investigated.

The objective is therefore not to obtain the largest possible sample, but to determine a Sample Size that is sufficient and methodologically justified.

So, How Many Participants Should a Research Study Have?

There is no universal rule stating that every study should have 100, 200, 300 or 400 participants. The appropriate Sample Size depends on the Research Design.

Consider a study examining employee satisfaction with company benefits. If the organisation has 800 employees and the Population Size is clearly known, Sample Size may be determined by considering the size of that population together with the acceptable level of sampling error.

Now consider a study investigating factors influencing consumers' intentions to purchase electric vehicles in Thailand. The target population is much larger, and if Multiple Regression is planned to examine several predictors simultaneously, the statistical model should also be considered when determining the Sample Size.

Another study might compare satisfaction among Generation X, Generation Y and Generation Z. Even if the total sample contains 300 respondents, the number of participants within each group becomes important when comparisons between those groups are part of the Research Objective.

Simply stating that a study has 300 or 400 respondents is therefore not enough to determine whether its Sample Size is appropriate.

Why Do So Many Research Studies Use Around 400 Participants?

The frequently seen figure of approximately 384–400 participants is not arbitrary. It is associated with Sample Size calculations used for certain types of survey research under commonly selected assumptions.

One approach that helps explain this number is Cochran's Formula for determining Sample Size when estimating a population proportion.

The basic formula is:

n = Z² × p(1−p) / e²

In this formula, Z corresponds to the selected Confidence Level, p represents the expected population proportion and e represents the acceptable Margin of Error.

Under commonly used survey assumptions, researchers might specify a 95% Confidence Level, giving Z ≈ 1.96, use p = 0.50 when the population proportion is unknown and set the Margin of Error at ±5%, or 0.05.

Substituting these values into the formula gives:

n = (1.96² × 0.50 × 0.50) / 0.05²

The resulting Sample Size is approximately 384.16, which would normally be rounded up to approximately 385 participants. In practice, some studies may target around 400 responses for convenience or to allow for unusable or incomplete questionnaires.

This is one important reason why Sample Sizes of approximately 384–400 appear so frequently in Survey Research.

Yamane's Formula can also produce figures approaching 400 when the population is large. A commonly cited form of the formula is:

n = N / [1 + N(e²)]

Here, N represents Population Size and e represents the specified level of precision.

When N becomes very large and e is set at 0.05, the resulting Sample Size approaches 400.

Krejcie and Morgan's approach also produces a Sample Size of approximately 384 for very large populations under its specified assumptions. This is another reason why a number around 384 is frequently encountered in research methodology.

However, the important point is that 384 or 400 is the result of particular assumptions. It is not a universal requirement for every research study.

For example, reducing the Margin of Error from ±5% to ±3% would substantially increase the required Sample Size. Conversely, when the population itself is relatively small, the appropriate Sample Size may be considerably below 400.

The number 400 should therefore be understood as a figure that emerges under certain Sample Size assumptions rather than as a standard that should automatically be applied to every study.

What Methods Can Be Used to Determine Sample Size?

There are several approaches to determining Sample Size. Each is based on different principles and assumptions, so the method should be selected according to the Research Design rather than simply choosing whichever formula is easiest to calculate.

1. Cochran

Cochran's Formula is commonly used when determining Sample Size for estimating proportions, particularly when the target population is large.

Factors involved include the Confidence Level, Margin of Error and expected population proportion.

When the total population is finite and known, an adjustment such as a Finite Population Correction may also be considered.

2. Yamane

Yamane's Formula is frequently encountered in survey research where the Population Size is known. It provides a relatively straightforward way of relating Population Size to a specified level of precision.

Its simplicity makes it convenient, but this does not mean it is automatically suitable for every Research Design.

3. Krejcie and Morgan

Krejcie and Morgan provide another widely recognised approach to determining Sample Size based on Population Size.

Their method is frequently presented as a table showing the recommended Sample Size corresponding to different population sizes under specified statistical assumptions.

The table is convenient when the population is known, but it should still be understood within the assumptions under which it was developed rather than treated as a universal rule.

4. Power Analysis

Power Analysis takes a different approach. Rather than determining Sample Size primarily from Population Size, it considers factors such as Effect Size, Significance Level, desired Statistical Power and the Statistical Test that will be performed.

This approach is particularly relevant to studies designed to test hypotheses, including those involving t-tests, ANOVA, Correlation or Regression.

Instead of asking which Sample Size formula is best, it is therefore more useful to ask which method is most appropriate for the Research Question and Analysis Plan.

Does Sample Size Depend on the Statistical Analysis?

Yes. Sample Size and statistical analysis are related and should ideally be considered before data collection begins.

A study using only Descriptive Statistics to describe a sample may have different Sample Size considerations from a study designed to test Hypotheses or build a Statistical Model.

Analyses such as t-tests, ANOVA, Pearson Correlation, Multiple Regression, Logistic Regression and Chi-Square all involve considerations relating to the amount and structure of the available data.

This does not mean that each statistical test has one fixed minimum Sample Size. Other factors may also matter, including the number of groups, number of predictors, Effect Size and characteristics of the data.

For example, a study may have 400 respondents overall but divide them into eight groups. If some groups contain only 15–20 participants, the total Sample Size of 400 does not automatically mean that every planned comparison will have sufficient data.

For this reason, Sample Size should be considered together with questionnaire design and the statistical analysis plan before data collection begins, helping ensure that the resulting dataset can support the Research Objectives and intended analysis.

If a Study Has Several Groups, Does It Need a Larger Sample Size?

Possibly. It depends on what the study intends to analyse.

Suppose a study aims to compare satisfaction among Generation X, Generation Y and Generation Z. Having a total Sample Size of 300 does not automatically mean that all three groups contain enough participants.

If the sample contains 30 Gen X respondents, 220 Gen Y respondents and 50 Gen Z respondents, the structure of the data is very different from having approximately 100 respondents in each group.

When subgroup comparisons are part of the Research Objective, the required number of participants within each subgroup should therefore be considered during Sample Size planning rather than looking only at the Total Sample Size.

What If the Sample Size Is Large Enough but the Participants Are Wrong?

Having enough respondents does not automatically mean that the sample is appropriate for the Research Objective.

Suppose a study aims to investigate luxury purchasing behaviour among consumers aged 30–55 in Bangkok. Even if 500 responses are collected, the sample may still be unsuitable if most respondents have never purchased the type of luxury product being studied.

Researchers therefore need to consider both Sample Size and Sample Characteristics.

For online research, our Online Panel questionnaire data collection service can help identify potential participants according to predefined criteria. Existing participant profile information, such as age, gender, location and other relevant demographic characteristics, can be used for initial screening before inviting suitable respondents to participate. Additional Screening Questions can also be included according to the requirements of the study.

The objective is therefore not simply to reach the required number of responses, but also to obtain respondents who correspond appropriately with the study's Target Population.

Should You Collect More Responses Than the Minimum Required Sample Size?

This should often be considered, particularly when some collected responses may ultimately be unsuitable for analysis.

If a study requires 300 valid cases, collecting exactly 300 responses can create a risk that the final usable Sample Size will fall below the requirement after data checking.

Some questionnaires may be incomplete, respondents may fail the Screening Criteria, important variables may contain Missing Data, or cases may not satisfy predefined Inclusion Criteria.

The number of participants recruited may therefore need to be greater than the Minimum Required Sample Size.

For example, if 300 valid cases are required, researchers may plan to collect more than 300 responses by a reasonable amount based on the characteristics of the study. The allowance should take into account the Target Group, data collection method, questionnaire length and data-quality criteria rather than applying an arbitrary percentage.

Longitudinal or clinical studies may additionally need to account for expected Dropout or Loss to Follow-up.

Should Sample Size Be Determined Before or After Designing the Questionnaire?

Sample Size planning should begin before data collection and should be considered alongside the Research Objectives, Hypotheses, Sampling Method, Questionnaire Design and Analysis Plan.

For example, if a study plans to use Multiple Regression with several Independent Variables, this should be taken into account when planning the Sample Size.

Similarly, if ANOVA will be used to compare four consumer groups, the researcher should consider the number of participants required within each group rather than determining only the overall Sample Size and dividing the sample afterwards.

If Sample Size is determined without considering the Analysis Plan, problems may only become apparent after data collection. The overall sample may appear large enough while particular subgroups contain too few cases, or the structure of the collected data may not adequately support the intended statistical analysis.

For studies that are still at the planning stage, our questionnaire design and review service can help align the questionnaire, variables, target sample and analysis plan with the research objectives before data collection begins.

For studies where data have already been collected, Data Investigator's SPSS statistical data analysis service can help review the dataset and identify an analytical approach appropriate to the Research Objectives, Hypotheses, Sample Size and characteristics of the available data.

Conclusion: What Is the Right Sample Size for Research?

There is no universal rule that 100, 200 or 400 participants represent the correct Sample Size for every study.

The frequently encountered figure of approximately 384–400 participants can be explained by Sample Size calculations used in certain survey settings. For example, Cochran's Formula produces approximately 384 participants when using a 95% Confidence Level, ±5% Margin of Error and p = 0.50. Other approaches may produce similar numbers for large populations under comparable assumptions.

However, an appropriate Sample Size should take into account more than this figure. Population Size, Research Objectives, Sampling Design, required precision, subgroup comparisons and the Statistical Analysis being planned may all influence how many participants are needed.

Cochran, Yamane and Krejcie and Morgan provide approaches that can be used under their respective assumptions, while Power Analysis connects Sample Size more directly with Effect Size, Statistical Power and the Statistical Test being conducted.

The number of respondents is also only one part of the equation. A large sample cannot compensate for participants who do not correspond to the Target Population.

A well-planned Sample Size should therefore answer two questions: how many participants are needed, and who should those participants be? Considering both from the beginning can help ensure that the collected data are appropriate for answering the Research Objectives and supporting the intended statistical analysis. For more information, please kindly contact:

E-mail: info@datainvestigatorth.com

Line Official Account: @datainvestigator     Tel: 063-969-7944