Which feature best distinguishes selection bias from information bias in epidemiological studies?

## Definition and Mechanism **Key Point:** Selection bias and information bias are fundamentally different types of bias that operate at different stages of a study. ### Selection Bias - Occurs during the **recruitment and enrollment phase** - Affects **who is included or excluded** from the study - Arises from systematic differences between those who participate and those who do not - Example: Berkson's bias in hospital-based case-control studies ### Information Bias (Measurement Bias) - Occurs during the **data collection and measurement phase** - Affects **how accurately data about subjects are collected, recorded, or classified** - Arises from errors in measuring exposure, outcome, or confounders - Example: Recall bias, observer bias, misclassification ## Comparison Table | Feature | Selection Bias | Information Bias | | --- | --- | --- | | **Stage of occurrence** | Subject recruitment/enrollment | Data collection/measurement | | **What is affected** | Who enters the study | How data are recorded | | **Nature** | Can be random or systematic | Usually systematic (but can be random) | | **Applicability** | All study designs | All study designs | | **Example** | Loss to follow-up, Berkson's bias | Recall bias, observer error | **High-Yield:** The key discriminator is the **timing and target**: selection bias happens at the **entry gate** (who gets in), while information bias happens at the **measurement station** (what we learn about them). **Clinical Pearl:** In a case-control study of myocardial infarction and oral contraceptive use, selection bias would occur if cases and controls were recruited from different populations (e.g., cases from tertiary care, controls from community); information bias would occur if cases over-reported OCP use due to recall bias. [cite:Park 26e Ch 8]