Quant #22: Sampling Bias

Sampling bias is a bias that arises when some members of the population have a higher likelihood to be selected than others. Common causes of sampling bias include the research design and the data collection method. Sampling bias can happen in random sampling or nonrandom sampling. In random sampling, for example, bias can be introduced when a sampling frame does not reflect the population. Nonrandom sampling often yields biased samples as some members of the population are more likely to be selected than others. Common types of sampling bias include:

1. Self-selection: This occurs when respondents with specific characteristics are more likely to agree to partake in a research study than others.
2. Non-response: This happens when respondents who refuse to participate or drop out from a study exhibit systematic differences from those who participated.
3. Undercoverage: This arises when some members of a population are underrepresented or inadequately represented in the selected sample.
4. Survivorship: This scenario happens when positive observations/outcomes or successful cases are more likely to be represented in the sample than unsuccessful ones.
5. Advertising or pre-screening: This bias arises from the way participants are pre-screened or where a study is advertised. In other words, this happens when selection criteria in a research study discourage some groups from participating in the research.
6. Healthy user: this occurs when participants in preventative interventions are more likely to adopt healthy behaviors than other members of the population.