In longitudinal designs, which problem commonly leads to loss of participants over time?

In longitudinal designs, which problem commonly leads to loss of participants over time?

• A. Blinding
• B. Attrition
• C. Random sampling error
• D. Measurement error

Answer: (B)

Attrition is the loss of participants over time in longitudinal studies. This happens when people drop out or become unavailable for follow-up assessments. It’s a central concern because if those who leave differ from those who stay—say, they have different severity levels, demographics, or treatment responses—the results can be biased and the effective sample size shrinks, reducing statistical power. For example, if more participants with worse outcomes drop out, the study may underestimate how severe the problem really is. To cope, researchers try to keep people involved through reminders and flexible scheduling, and they use missing-data techniques like multiple imputation or maximum likelihood, and consider patterns of missingness in their analyses. Blinding is about preventing bias from expectations by keeping participants or researchers unaware of group assignment, not about people dropping out. Random sampling error refers to random fluctuations due to selecting a sample, not the systematic loss of participants over time. Measurement error involves inaccuracies in the data collected, rather than participants leaving the study.

Attrition is the loss of participants over time in longitudinal studies. This happens when people drop out or become unavailable for follow-up assessments. It’s a central concern because if those who leave differ from those who stay—say, they have different severity levels, demographics, or treatment responses—the results can be biased and the effective sample size shrinks, reducing statistical power. For example, if more participants with worse outcomes drop out, the study may underestimate how severe the problem really is. To cope, researchers try to keep people involved through reminders and flexible scheduling, and they use missing-data techniques like multiple imputation or maximum likelihood, and consider patterns of missingness in their analyses.

Blinding is about preventing bias from expectations by keeping participants or researchers unaware of group assignment, not about people dropping out. Random sampling error refers to random fluctuations due to selecting a sample, not the systematic loss of participants over time. Measurement error involves inaccuracies in the data collected, rather than participants leaving the study.