Causation/Random Error/Bias

Question 1

5 Possible Reasons E Related to O

Answer 1

Random Association
Systematic Error
Confounding
Reverse Causation
Causation

Question 2

Standard Error

Answer 2

Standard deviation of a sample. Often the variance

Question 3

Type 1 vs. 2 Error

Answer 3

Rejects null when null actually true vs. accepts null when null false

Question 4

CI Interval and Null Value

Answer 4

If null contained w/in CI, not statistically significant

Question 5

2 Main Types of Bias

Answer 5

Selection bias

Information bias - interviewer or recall

Question 6

Nondifferential vs. Differential Misclassification Bias

Answer 6

Non: misclassification random, equal problem for both groups, but this dilutes association so BIASES TOWARDS NULL. (in all studies, so most studies actually underestimate effect size)
Differential: nonrandom, non-equal misclassification b/w groups so bigger problem and effect on association is unpredictable

Question 7

3 Forms of Selection Bias

Answer 7

Participant selection procedure incorrect
Non response
Attrition

Question 8

Healthy Worker Effect

Answer 8

Major issue in occupational health studies (esp cohort) - employed people have overall better health than unemployed. Attenuates the risk of an occupational hazard

Question 9

Case Control Study Selection Bias (& key principle to avoid)

Answer 9

Difficult to pick good control group - exposure rates should approximate exposure rates seen in non-diseased members of source population