How Do Eligibility Rates, Refusal Rates, and Attrition Affect Recruitment Targets?

Why Might a Study Requiring 300 Participants Need 700 Invitations?

You have calculated that your study requires 300 participants with usable data. You identify 500 students who might qualify and plan to invite them to participate.

However, not every student meets the eligibility criteria. Some eligible students decline consent, while others enroll but fail to complete the required follow-up assessment.

Would inviting 500 students be sufficient? How should you account for participants who are ineligible, refuse participation, or withdraw after enrollment? Should you add 10% or 20% to the calculated sample size, as researchers sometimes do?

The answer depends on the successive stages through which potential participants must pass. Eligibility, refusal, and attrition affect different denominators and should be incorporated into recruitment planning accordingly.

Each Recruitment Loss Reduces the Number Expected to Contribute Usable Data

Understanding How Recruitment Losses Accumulate

What Is an Eligibility Rate?

The eligibility rate is the proportion of screened individuals who satisfy the study's inclusion and exclusion criteria.

For example, suppose 400 students complete screening and 320 meet the eligibility criteria. The observed eligibility rate is 320 divided by 400, or 80%.

Eligibility depends on the population approached and the criteria applied. Recruiting from a verified list of students enrolled in a particular course may produce a different eligibility rate from advertising the study across an entire university.

Eligibility should not be confused with willingness to participate. An eligible individual may still decline consent.

What Is a Refusal Rate?

The refusal rate describes the proportion of individuals who decline participation within a clearly specified denominator.

When calculated among eligible individuals who have received an invitation and made a participation decision, it is the complement of the corresponding consent proportion.

For example, if 200 eligible individuals make a decision and 50 refuse, the refusal rate is 25% and the consent rate is 75%.

However, individuals who never respond should not automatically be classified as explicit refusals. Noncontact, nonresponse, and refusal may represent different recruitment outcomes.

Researchers should define these categories consistently and report the denominators used.

What Is Attrition?

Attrition refers to the loss of participants after enrollment, commonly through withdrawal or failure to complete required follow-up activities.

For example, if 300 participants enroll and 255 provide the required final follow-up outcome, the observed loss proportion is 45 divided by 300, or 15%.

Attrition is distinct from refusal because it occurs after enrollment.

Its implications depend on the research design. A participant who misses one assessment may still contribute useful baseline or intermediate data, particularly under an appropriate longitudinal analysis.

Consequently, attrition should not automatically be treated as equivalent to complete analytical exclusion.

How Do These Rates Differ?

Definitions vary across designs and reporting standards. The essential requirement is that the numerator, denominator, and stage of recruitment are clearly identified.

How Do You Combine Eligibility, Consent, and Retention?

When everyone approached completes eligibility screening, the expected overall recruitment yield can be represented by the product of the eligibility, enrollment, and usable-data proportions.

If some approached individuals do not complete screening, their screening-completion proportion must also be included.

These conditional probabilities can be multiplied because they describe successive stages. They do not need to be statistically independent, but their definitions must be consistent.

How Do You Calculate the Required Contact Target?

Once the expected overall yield is established, divide the required analytical sample by that yield.

This calculation estimates the number needed to obtain 300 usable observations on average. It does not guarantee that 631 approaches will produce the required sample.

The calculation of how many people to approach can be extended to additional recruitment stages and assurance-based planning.

Why Is Adding an Arbitrary Percentage Often Incorrect?

Researchers sometimes increase the calculated sample by 10% or 20% without examining what the percentage represents.

Suppose 200 completed follow-up observations are needed and 20% of enrolled participants are expected to be lost.

Adding 20% gives 240 participants. However, 80% of 240 is only 192, below the required 200.

Dividing 200 by 0.80 gives 250 enrolled participants, whose expected retained count is 200.

The distinction arises because attrition is calculated relative to those enrolled, not the final required sample.

Should Refusal and Attrition Be Combined Into One Percentage?

They may be represented through a combined overall yield, but only when the underlying stages and denominators are understood.

For example, 25% refusal among eligible individuals and 20% attrition among enrolled participants produce an expected progression proportion of 0.75 × 0.80 = 0.60 from eligibility to retained participation.

It would be incorrect to add the two percentages and conclude that 45% of eligible individuals will necessarily be lost.

Sequential losses act on different groups. Multiplication reflects that structure.

What If Some Individuals Never Respond?

Nonresponse should be distinguished from explicit refusal where the recruitment process allows that distinction.

For example, an invitation may be delivered to 1,000 students, but only 600 open the screening form. The remaining 400 cannot automatically be classified as ineligible or as having explicitly refused.

The overall recruitment calculation should account for the proportion who progress from invitation to screening.

Researchers should also examine whether nonrespondents differ systematically from respondents in ways relevant to the study.

A larger number of invitations may improve the final count without eliminating nonresponse bias.

How Does Attrition Affect Statistical Power?

Attrition may reduce the information available for estimating an effect or testing a hypothesis, potentially reducing statistical power.

However, the effect depends on the timing and pattern of missing data, the analysis method, and the outcome distribution.

For example, a longitudinal mixed-effects model may use some incomplete repeated-measures records under specified assumptions, whereas an analysis requiring a complete final outcome may lose more information.

Researchers should align recruitment inflation with the actual sample-size calculation and missing-data plan.

Inflating enrollment does not itself correct bias caused by informative missingness.

Can Attrition Introduce Bias Even When the Final Sample Is Large?

Yes. A study may retain the required number of participants but still produce biased estimates if loss to follow-up is related to outcomes or other relevant characteristics.

For example, students experiencing severe academic difficulties may be less likely to complete a follow-up survey about learning experiences.

If those students are systematically underrepresented, the achieved sample may not adequately represent the intended outcome distribution.

Researchers should consider appropriate retention strategies and missing-data methods rather than relying solely on overrecruitment.

Do Eligibility Rates Differ Across Recruitment Channels?

They often can.

A recruitment invitation distributed through a course-specific mailing list may reach a high proportion of eligible students. A general social media advertisement may reach a larger audience but include many ineligible respondents.

Similarly, consent rates may differ according to study burden, recruitment mode, and participants' relationship with the institution.

Researchers should use channel-specific assumptions where credible evidence supports them.

A single pooled rate may conceal important differences and produce unrealistic operational targets.

Should Different Study Sites Use Different Recruitment Rates?

Yes, when site characteristics or previous recruitment evidence indicate meaningful differences.

For example, universities may differ in their eligible student populations, recruitment permissions, academic calendars, and capacity to contact prospective participants.

Site-specific calculations can identify whether each institution is likely to meet its assigned target.

The planning of recruitment separately by study site is particularly important when the analysis requires adequate representation of each site.

How Can Recruitment Assumptions Be Estimated?

Relevant evidence may come from previous institutional studies, administrative recruitment records, comparable published research, or a preliminary feasibility assessment.

Researchers should prioritize evidence reflecting the proposed population, recruitment channel, eligibility criteria, and study burden.

When reliable estimates are unavailable, assumptions should be presented as provisional.

For example, using a 70% consent rate without any supporting evidence may be reasonable as one sensitivity scenario but not as a confidently established prediction.

How Should Uncertainty in Recruitment Rates Be Handled?

Recruitment rates are uncertain even when based on previous evidence.

Researchers can evaluate alternative combinations of eligibility, enrollment, and usable-data proportions.

The figures assume all approached individuals are screened. The rates are hypothetical and the contacts are rounded upward.

The scenarios illustrate how modest differences at successive stages can substantially change the number of invitations needed.

What If the Recruitment Funnel Changes During the Study?

Researchers should monitor actual progression rates and compare them with the original assumptions.

If eligibility is lower than expected, the team may need to improve how potential participants are identified.

If refusal is high, the team should examine invitation clarity, participant burden, confidentiality concerns, and scheduling.

If attrition is high, retention procedures and the causes of missing follow-up should be investigated.

Material changes to recruitment or eligibility procedures should be scientifically justified and approved where required.

Should Recruitment Continue Until the Required Sample Is Achieved?

For many quantitative studies, recruitment aims to obtain the information required by the prespecified design, subject to ethical and operational limits.

However, the target should not be pursued through inappropriate pressure, unauthorized changes, or indefinite recruitment without reassessing feasibility.

Researchers should distinguish invitations from enrollment and avoid unnecessary overrecruitment once the required sample is reasonably secured.

In qualitative research, recruitment may instead be guided by methodology-specific adequacy criteria rather than a fixed numerical sample.

The development of a realistic recruitment plan should therefore follow the design's actual informational requirements.

How Should Recruitment Losses Be Reported?

Reports should distinguish individuals approached, screened, eligible, enrolled, retained, and analyzed where applicable.

The STROBE statement recommends reporting participant numbers at relevant stages of observational studies, reasons for nonparticipation, and consideration of a flow diagram.

For randomized trials, the applicable CONSORT reporting guidance should be followed.

Transparent reporting allows readers to assess whether recruitment losses may affect the study's precision, power, or susceptibility to selection bias.

Calculating Recruitment for a Longitudinal Student Study

The expected contact requirement of 631 does not guarantee that 300 participants will complete follow-up.

Researchers requiring a specified probability of achieving the final target may need additional assurance-based planning, particularly when recruitment rates are uncertain.

Common Misconceptions About Recruitment Losses

Calculate Recruitment Targets From Stage-Specific Rates

Start with the required usable sample, then identify the relevant recruitment stages and their conditional probabilities.

Use the combined yield to calculate the number of people who must be approached, and evaluate uncertainty before treating the target as operationally feasible.

Recruitment inflation should reflect the expected losses and informational requirements of the study, not an arbitrary allowance applied to every project.

Before Finalizing Recruitment Adjustments

Frequently Asked Questions About Eligibility, Refusal, and Attrition

Recruitment Losses Must Be Calculated at the Stages Where They Occur

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.