Bias

Question 1

Selection bias

(Sampling bias)

Answer 1

• Nonrandom assignment into groups leading to unrepresentative sample
• Non-generalizable
• Ex: study population selected from hospital is generally less healthy

Question 2

Recall Bias

Answer 2

• Knowledge of presence of disease changes subject’s response
• Common in retrospective studies

Question 3

Measurement Bias

Answer 3

• Information is gathered in a way that distorts it
• Ex: miscalibrated scale

Question 4

Procedure Bias

Answer 4

• Subjects in different groups not treated the same
• Ex: patients in treatment group spend more time in specialized units

Question 5

Observer Bias

Answer 5

• Researcher’s decision affected by prior knowledge of subject’s exposure status

Question 6

Hawthorne Effect

Answer 6

• Subjects change behavior when they know they’re being studied
• Subject’s behavior

Question 7

Confounding Bias

Answer 7

• One factor distorts/confuses the effect of another closely related one
• Factor is related to both exposure and outcome, but is not in causal path

Question 8

Lead-time Bias

Answer 8

• Early detection confused with increased survival
• Early detection makes it seem like survival has increased, but natural course of disease has not changed
• Seen in screening trials

Question 9

Selection bias can lead to what type of error?

Answer 9

both type 1 and type 2

Question 10

Small samples decrease power, which can lead to what type of error?

Answer 10

type 2

• increases the effect of random error, which bias toward the null

Question 11

Non-compliance can lead to what type of error?

Answer 11

Type 2

bias toward the null

Question 12

Co-interventions in a control group can lead to what kind of error?

Answer 12

Type 2

bias toward the null

Question 13

Co-interventions in a treatment group can lead to what kind of error?

Answer 13

Type 1

Question 14

If you decrease prevalence, what happens to the +PV?

Answer 14

decreases

Question 15

General advantages of cohort studies:

Answer 15

1. assess exposure and potential confounders at the same time
2. can study multiple outcomes
3. less recall bias and selection bias
4. directly measure RR and AR

Question 16

General disadvantages of cohort studies:

Answer 16

1. long wait for chronic diseases with long induction periods
2. costly long term F/U
3. loss to F/U
4. inefficient for rare disease