Bias and Confounding Lecture Notes

Vocabulary flashcards focused on key terms and definitions related to bias and confounding in study designs.

Bias

A systematic error in study design that leads to false associations between exposure and disease.

Confounding bias

Distortion of the association between exposure and disease by a third variable, known as a confounder.

Internal validity

The extent to which the study is conducted properly to produce accurate results.

External validity

The ability to generalize study results to a larger population.

Randomization

A method to assign study participants randomly to ensure a balanced distribution of confounders.

Selection bias

Bias resulting from systematic differences in selecting study participants.

Information bias

Bias arising from systematic differences in the way information on exposures and diseases is gathered.

Recall bias

A type of information bias where case participants may remember past exposures differently than control participants.

Attrition bias

Bias that occurs when participants who drop out of a study differ from those who remain.

Performance bias

Bias that occurs when participants change their behavior because they are aware of group allocations.

Misclassification bias

Error in classifying subjects regarding exposure or disease status, leading to biased results.

Confounding

Occurs when another variable influences both the exposure and outcome being studied, distorting the association.

Recency bias

A form of recall bias where memories of events that occurred recently are more accurate than those from the past.

Observer bias

When the researcher's knowledge of participants affects their observation or reporting.

Stratification

A technique that involves dividing a study population into sub-groups based on a specific criterion to control confounding.

Matching

A method of pairing participants in different study groups to balance known confounders.

Variance

Random error in a study that causes deviations in results due to chance rather than systematic differences.

Measurement error

Any distortion in the data collection process leading to incorrect results.

Blinding

A process to prevent participants or researchers from knowing assignments to control bias.

Regression to the mean

The phenomenon where extreme measurements tend to be closer to the average on subsequent assessments.

Null hypothesis

A statement that there is no effect or no difference, serving as a baseline for statistical testing.