Lecture 13: Bias and Misclassification

Question 1

Bias

Answer 1

systematic (non-random) error in study design or conduct leading to erroneous results - distorts relationship between exposure and outcome

Question 2

Recall/Reporting Bias

Answer 2

differential level of accuracy/detail between subjects; diseased individuals may be more sensitive to recalling history
MEASUREMENT BIAS

Question 3

Hawthorne effect

Answer 3

subject unable to report accurately; overly encouraging, everything is positive, giving more info than needed
MEASUREMENT BIAS

Question 4

Contamination bias

Answer 4

subjects in control group accidentally/against protocol receive treatment/exposure that is being studied
MEASUREMENT BIAS

Question 5

Compliance bias

Answer 5

differential level of compliance/adherence to protocol between subjects
MEASUREMENT BIAS

Question 6

Lost to follow-up bias

Answer 6

different withdrawal or lost to follow up rates in studies leads to unequal distribution of groups
MEASUREMENT BIAS

Question 7

Diagnosis/Surveillance/Expectation bias

Answer 7

observers/interviewers may have preconceived expectations of what should be found in examination, evaluation, and follow up
MEASUREMENT BIAS

Question 8

Interviewer Proficiency bias

Answer 8

difference in the way questions are asked/recorded/interpreted by interviewer
MEASUREMENT BIAS

Question 9

Neyman bias

Answer 9

Mild, clinically resolved, or fatal cases of disease are not included in a study
SELECTION BIAS

Question 10

Lead-Time/Length Bias

Answer 10

The outcome of one test identifies disease earlier, giving the impression that survival was prolonged
SELECTION BIAS

Question 11

Selection bias

Answer 11

Selecting study subjects that are not representative of your primary population of interest or generate differences between members of the group
SELECTION BIAS

Question 12

Self-Selection/Responder/Participant bias

Answer 12

Individuals who readily volunteer for studies may be different than those would not not normally volunteer
SELECTION BIAS

Question 13

Healthy Worker Effect

Answer 13

Individuals no longer employed/dead/sick - not at employer facility where study is being conducted - are often ignored
SELECTION BIAS

Question 14

Misclassification

Answer 14

a measurement bias; error in classifying disease, exposure status, or both

Question 15

Non-differential misclassification

Answer 15

misclassification in which error occurs in both groups equally; balanced/conservative error; moves OR closer to 1.0

Question 16

Differential misclassification

Answer 16

misclassification in which error occurs in one group differently than the other; unbalanced/worse; may move OR toward or away from 1.0