Sampling bias occurs when the sample is not representative; how can it be minimized?

• A. Sampling bias occurs when the sample is not representative; minimized by random sampling, stratified sampling, and broad recruitment.
• B. Sampling bias is about measurement error.
• C. Sampling bias is eliminated by doing nothing.
• D. Sampling bias improves generalizability.

Answer: (A)

The main idea here is that sampling bias happens when the chosen participants don’t reflect the larger population, which can skew results. To minimize it, use methods that give everyone in the population a fair chance to be included and ensure subgroups are represented. Random sampling helps because each member has an equal chance of selection, reducing systematic preference that could distort findings. Stratified sampling ensures key subgroups are represented in proportions that match the population, preventing over- or under-representation of important groups. Broad recruitment expands who is invited or who volunteers, which reduces self-selection bias and improves diversity so the sample better mirrors the population.

Seeing sampling bias as measurement error isn’t accurate—the issue is about who is included, not how outcomes are measured. Doing nothing won’t fix bias, and saying bias improves generalizability is the opposite of what it does. Using these sampling techniques together is the best way to minimize bias and support more accurate generalizations.