Common sampling biases?

Common sampling biases?

• A. Convenience bias
• B. Sampling bias is impossible in random samples
• C. Selection bias, nonresponse bias, measurement bias, and recall bias
• D. Overfitting bias

Answer: (C)

In sampling and data collection, biases undermine how well a sample represents the population, and several familiar sources can slip in at different stages. Selection bias happens when the method of choosing participants favors certain groups, so the sample isn’t representative. Nonresponse bias appears when those who participate differ in important ways from those who don’t respond, altering the results. Measurement bias arises from errors or inconsistencies in how measurements are taken, which can skew results regardless of who is in the sample. Recall bias occurs when participants’ memories distort the accuracy of their responses, especially for past events or exposures. Together, these biases cover the main ways data can become unrepresentative or distorted in practice. The other options are less accurate. One mentions a single bias and misses the broader set of common issues. The idea that sampling bias cannot occur in random samples is false, since biases can still arise from nonresponse or measurement errors even with random selection. Overfitting bias is a modeling problem, not a sampling bias.

In sampling and data collection, biases undermine how well a sample represents the population, and several familiar sources can slip in at different stages. Selection bias happens when the method of choosing participants favors certain groups, so the sample isn’t representative. Nonresponse bias appears when those who participate differ in important ways from those who don’t respond, altering the results. Measurement bias arises from errors or inconsistencies in how measurements are taken, which can skew results regardless of who is in the sample. Recall bias occurs when participants’ memories distort the accuracy of their responses, especially for past events or exposures. Together, these biases cover the main ways data can become unrepresentative or distorted in practice.

The other options are less accurate. One mentions a single bias and misses the broader set of common issues. The idea that sampling bias cannot occur in random samples is false, since biases can still arise from nonresponse or measurement errors even with random selection. Overfitting bias is a modeling problem, not a sampling bias.