Sampling Techniques & Participant Recruitment Flashcards

Intraclass correlation coefficient (ICC) in cluster sampling measures:

Answer: The degree to which cluster members resemble each other more than the overall population

ICC captures how similar members within a cluster are to each other; a higher ICC means more homogeneity within clusters and a larger design effect.

A researcher conducting research with a rare disease population uses existing patient advocacy group networks to recruit participants. This is best described as:

Answer: Convenience sampling via community partnerships

Recruiting through established community or patient networks is a form of convenience sampling where an accessible existing group is leveraged rather than random selection.

Which statement about panel surveys is accurate?

Answer: Panel attrition can introduce bias as dropouts may differ from continuers

Panel attrition—where some participants drop out between waves—can bias results if dropouts systematically differ from those who continue participating.

When oversampling a small subgroup in a survey, researchers must apply post-survey weights to:

Answer: Restore the correct proportional representation in aggregate estimates

Oversampling inflates the representation of a subgroup, so survey weights are needed to scale results back to actual population proportions for overall estimates.

The central limit theorem supports the use of normal-distribution-based calculations in survey research primarily because:

Answer: Sample means from sufficiently large samples approximate a normal distribution regardless of population distribution

The central limit theorem states that sampling distributions of means approach normality as sample size increases, justifying normal-based confidence intervals and hypothesis tests.

A telephone survey that calls only listed landline numbers will most likely underrepresent:

Answer: Younger adults who rely exclusively on mobile phones

Landline-only frames miss households with only cell phones, which disproportionately include younger adults, lower-income individuals, and renters.

Respondent-driven sampling (RDS) differs from basic snowball sampling primarily because RDS:

Answer: Uses probabilistic referral chains and adjusts for network size to estimate population parameters

RDS extends snowball sampling by using structured peer recruitment chains and mathematical adjustments for network size, allowing population-level estimates with defined statistical properties.