Should Participant Recruitment Be Planned Separately for Each Study Site?

If a Study Needs 600 Participants From Four Universities, Should Each University Recruit 150?

You are conducting a multi-university study investigating students' experiences with online learning. Your sample-size calculation indicates that 600 participants are required.

You plan to recruit students from four universities and initially assign 150 participants to each institution.

However, the universities differ substantially. One has several thousand eligible students, another has only a few hundred, and the remaining institutions have different academic calendars and recruitment procedures.

Should every university recruit the same number? Would allocating participants proportionally to institutional enrollment be more appropriate? What happens if one institution fails to meet its target while another exceeds it?

These questions concern the relationship between overall sample-size requirements, site-specific recruitment feasibility, and the analytical purpose of including multiple study sites.

Yes, Site-Specific Planning Is Usually Advisable, but Equal Targets Are Not Always Appropriate

Designing and Managing Site-Specific Recruitment Targets

Why Should Recruitment Be Considered Separately for Each Site?

Study sites may differ in the number of eligible participants, recruitment procedures, institutional permissions, and expected response rates.

For example, a university with 3,000 eligible students may be able to contribute more participants than an institution with only 300 eligible students.

Other differences may concern academic schedules, student engagement, research participation policies, or the availability of personnel who can support recruitment.

A pooled recruitment target can conceal these differences. Site-specific planning makes it possible to assess whether each institution can realistically contribute the participants required by the study design.

Does Every Site Need the Same Number of Participants?

No. Equal allocation is one possible strategy, but its appropriateness depends on the research objective.

If the study seeks balanced comparisons between universities, approximately equal site samples may be useful.

If the objective is to estimate a characteristic of the combined student population, proportional allocation may better reflect the relative sizes of the institutions under an appropriate probability-sampling design.

If the study focuses on underrepresented institutional contexts, disproportionate allocation may be justified, although weighting or other adjustments may be necessary for population-level estimation.

The allocation should follow the research question rather than a preference for numerically symmetrical groups.

What Are the Main Approaches to Site Allocation?

These strategies are not mutually exclusive. A study might establish minimum site targets for comparison while allocating remaining recruitment according to eligible population size.

Any adjustment should preserve the analytical requirements and comply with the approved research protocol.

How Do You Calculate Proportional Site Targets?

For proportional allocation, the target for each site can be calculated from its share of the total eligible population.

Proportional allocation is not automatically optimal for every statistical objective. It does not by itself correct selection bias, and the resulting precision depends on the sampling procedure, site variability, and analysis.

How Is Site-Level Recruitment Capacity Calculated?

Each site's expected recruitment yield can be estimated using its screening, eligibility, consent, and usable-data rates.

These are expected-yield calculations. Actual recruitment may differ, and the contact requirement may exceed the site's accessible population.

The general method for calculating how many people must be approached should therefore be applied using assumptions appropriate to each site.

What If Sites Have Different Eligibility Rates?

Eligibility differences can substantially change recruitment feasibility.

One institution may have a large student population but relatively few students meeting the study's specific criteria. Another may have a smaller population with a high proportion of eligible participants.

For example, a study of remote laboratory experiences may find that only certain programs offer the relevant learning activities.

Recruitment targets should therefore be based on eligible populations rather than total institutional enrollment alone.

The influence of eligibility, refusal, and attrition on recruitment targets should be evaluated separately for each site when credible evidence indicates differences.

Should Site-Level Sample Size Depend on the Intended Analysis?

Yes. The analytical objective is central to site allocation.

If the study aims to estimate an overall association while accounting for site effects, it may not require equal sample sizes at every institution.

If it aims to estimate separate effects or descriptive parameters for each university, the sample within each site must support those estimates with appropriate precision.

If the objective is to compare institutional differences, both the number of sites and the number of participants within them may matter.

Researchers should establish these requirements during sample-size planning rather than assume that an adequate pooled sample automatically supports every site-specific conclusion.

How Does Clustering Affect Multi-Site Sample Size?

Participants recruited within the same institution may share characteristics or experiences because of common policies, teaching practices, or institutional environments.

Such dependence can affect statistical precision.

For approximately equal cluster sizes, a conventional design-effect expression illustrates how intracluster correlation can reduce information relative to independent sampling.

This approximation does not apply universally to every multi-site analysis. Unequal cluster sizes, covariate adjustment, site selection, and the number of clusters can alter precision.

Importantly, recruiting more individuals within a small number of institutions may not substitute for including additional independent institutions when the intended inference concerns between-site variation.

Is Every Multi-Site Study a Cluster-Randomized Study?

No. Recruiting participants from several institutions does not automatically mean that the study uses cluster randomization.

In an individually randomized trial, participants may be randomized within each institution. In a cluster-randomized trial, institutions or other groups may be randomized as units.

Observational multi-site studies may involve no random assignment at all.

The appropriate sample-size calculation depends on the actual design, the level of assignment where relevant, and the statistical analysis.

Researchers should not apply a cluster-randomized trial calculation merely because participants come from different universities.

Why Does the Number of Sites Matter?

The number of participating sites can be important when the analysis seeks to estimate institutional variation or generalize across institutions.

For example, collecting 1,000 students from two universities provides different information about between-university variation from collecting 1,000 students across twenty universities.

A large number of participants within very few sites may provide precise information about those particular institutions while offering limited evidence about institutional differences more broadly.

Statistical models involving random site effects may also require careful attention to the number of sites available for estimating between-site variability.

Increasing participant numbers within existing sites does not necessarily resolve insufficient site-level information.

Can One Site Compensate for Another Site's Recruitment Shortfall?

Sometimes, but not automatically.

If the study's objective concerns a pooled population and the sampling and analysis permit reallocation, additional participants from another site may help meet the overall sample target.

However, reallocation may alter population representation, weighting requirements, or the precision of site-specific estimates.

If the design requires a minimum number from each institution, excess recruitment at one site may not compensate for a shortfall elsewhere.

Researchers should examine the statistical and substantive implications before reallocating targets.

How Do Site Differences Affect Selection Bias?

Recruitment rates may vary across institutions in ways related to the study's outcomes.

For example, students with low engagement in an online learning platform may be less likely to respond to recruitment messages distributed through that same platform.

If recruitment channels and response patterns differ between universities, the final sample may reflect those processes rather than the intended student population.

Site-specific monitoring can identify such differences, but it does not automatically eliminate selection bias.

Researchers should consider appropriate recruitment channels, sampling procedures, and analytical adjustments when justified.

Should Recruitment Begin at Every Site Simultaneously?

Not necessarily. Simultaneous recruitment can improve calendar comparability but may increase operational demands.

Staggered recruitment can allow researchers to evaluate procedures before expanding to additional sites.

However, differences in semester timing, examinations, or institutional events may affect participant availability and responses.

The decision about staggering recruitment across study sites should consider both feasibility and the potential consequences of recruitment timing.

How Should Site Recruitment Be Monitored?

A multi-site recruitment record should track progress at each institution using consistent definitions.

Relevant measures include contacts, screening completions, eligible individuals, consent, enrollment, retention, and usable data.

Researchers should also monitor recruitment dates, approved channels, and relevant site-specific barriers.

For example, one site may be behind schedule because of delayed institutional approval, while another may have a lower consent rate despite adequate outreach.

These situations require different operational responses.

What If a Site Cannot Meet Its Target?

Begin by identifying whether the shortfall reflects a small eligible population, limited contact capacity, low participation, or attrition.

Possible responses include extending recruitment, improving approved communication procedures, or reallocating recruitment where the study design permits.

However, researchers should not automatically remove a poorly recruiting site if doing so changes the population or undermines planned comparisons.

Material changes to site participation, eligibility, or allocation may require ethics or institutional authorization.

The consequences for sample-size adequacy and generalizability should be evaluated before implementation.

How Does Site-Level Planning Differ in Qualitative Research?

Qualitative multi-site studies may use purposive recruitment to investigate contextual variation rather than achieve statistically proportional representation.

For example, a qualitative study of online course design might deliberately include institutions with different technological resources or assessment policies.

The number of participants within each site should reflect the intended analytical depth and comparisons.

Equal numbers are not automatically required, and a statistical proportional-allocation formula may be inappropriate.

Researchers should justify the selection of sites, the participants within them, and the adequacy of the resulting material according to the qualitative methodology.

How Should Multi-Site Recruitment Be Reported?

Reports should describe how sites were selected, how participants were recruited within them, and how the achieved sample was distributed.

Where relevant, researchers should provide site-level recruitment and participation counts.

The STROBE statement recommends transparent reporting of participant selection and study settings for observational research.

For multi-site randomized trials, the applicable CONSORT guidance should also be considered.

Researchers should explain important deviations from planned site allocations and their implications for the analysis and interpretation.

Allocating 600 Student Participants Across Four Universities

The example illustrates proportional allocation rather than establishing that it is the best strategy for every multi-site study.

If the study requires equally precise site comparisons, or if recruitment rates differ substantially, another allocation may be more appropriate.

Common Misconceptions About Multi-Site Recruitment

Plan Recruitment at Both the Overall and Site Levels

Begin with the overall research objective and determine what information each site must contribute.

Then establish site-specific targets using eligible population sizes, anticipated recruitment yield, analytical requirements, and feasible capacity.

Separate site-level planning improves operational visibility, but its methodological value depends on alignment with the intended analysis.

Before Finalizing Multi-Site Recruitment Targets

Frequently Asked Questions About Recruitment by Study Site

Site-Level Recruitment Planning Protects Both Feasibility and Analytical Adequacy

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.