PRC Sampling Techniques & Participant Recruitment 3 — Questions and Answers
Question 1: A market research firm needs 100 completed interviews and assigns interviewers to recruit specific numbers of men and women in each age bracket. This describes:
- Disproportionate stratified sampling
- Quota sampling (Correct answer)
- Multistage sampling
- Voluntary response sampling
Correct answer: Quota sampling
Quota sampling sets predetermined targets for subgroup representation and allows interviewers to recruit non-randomly until quotas are filled.
Question 2: When calculating the minimum sample size, increasing the desired confidence level from 90% to 95% will:
- Decrease the required sample size
- Have no effect on sample size
- Increase the required sample size (Correct answer)
- Eliminate the need for a sampling frame
Correct answer: Increase the required sample size
A higher confidence level requires a larger critical value (z-score), which increases the calculated minimum sample size needed to achieve that confidence.
Question 3: In a multistage sampling design, which sequence correctly describes the process?
- Select individuals first, then clusters
- Select clusters, then subgroups, then individuals (Correct answer)
- Stratify by quotas, then select clusters
- Select all individuals within randomly chosen strata
Correct answer: Select clusters, then subgroups, then individuals
Multistage sampling progressively narrows selection from larger units (e.g., counties) to intermediate units (e.g., schools) to final units (e.g., students).
Question 4: Which participant recruitment method poses the greatest risk of self-selection bias?
- Random digit dialing
- Probability panel recruitment
- Opt-in online surveys (Correct answer)
- Address-based sampling
Correct answer: Opt-in online surveys
Opt-in online surveys attract only those who choose to participate, systematically overrepresenting people with strong opinions or high interest in the topic.
Question 5: A design effect (DEFF) greater than 1.0 in cluster sampling indicates:
- The sample is more efficient than simple random sampling
- The sample requires fewer respondents than SRS
- The variance is larger than under simple random sampling (Correct answer)
- The clustering has no impact on standard errors
Correct answer: The variance is larger than under simple random sampling
A DEFF > 1.0 means that clustering increases variance compared to SRS, requiring a larger sample to achieve the same precision.
Question 6: Which IRB requirement is most directly related to participant recruitment practices?
- Data storage encryption standards
- Obtaining informed consent before participation (Correct answer)
- Selecting the appropriate statistical test
- Establishing a data retention schedule
Correct answer: Obtaining informed consent before participation
IRBs require researchers to obtain voluntary informed consent from participants before collecting data, ensuring ethical recruitment practices.
Question 7: Address-based sampling (ABS) is considered an improvement over random digit dialing (RDD) primarily because:
- It eliminates the need for survey weights
- It achieves higher cooperation rates automatically
- It provides broader coverage including cell-only households (Correct answer)
- It requires no sampling frame
Correct answer: It provides broader coverage including cell-only households
ABS uses postal address files which cover nearly all households including those with only cell phones, reducing the undercoverage that affects RDD frames limited to landlines.
PRC Sampling Techniques & Participant Recruitment 3
A market research firm needs 100 completed interviews and assigns interviewers to recruit specific numbers of men and women in each age bracket.
This describes: