Biostats

Question 1

Type I Error

Answer 1

incorrectly positive

FP / (FP+TN)

Question 2

Type II Error

Answer 2

incorrectly negative

FN / (FN+TP)

Question 3

Power

Answer 3

1-beta

Question 4

Meaning of power

Answer 4

probability of rejecting null when it is truly false (usu 80%)

Question 5

alpha

Answer 5

maximum probability of making type I error someone willing to accept

typically p=0.05

Question 6

beta

Answer 6

probability of committing type II error

Question 7

cross-sectional study

Answer 7

measures exposure AND outcome simultaneously
“snapshot” design frequently used in surveys

limited by unclear temporal relationship b/w exposure and outcome

Question 8

cohort study

Answer 8

select a cohort, DETERMINE their exposure status, then follow them in time

Question 9

case-control study

Answer 9

selecting individuals with particular disease (cases), individuals without disease (controls), and then evaluating PREVIOUS exposure status

Question 10

Berkson’s bias

Answer 10

selection bias created by selecting hospitalized pts as control group

Question 11

Pygmalion effect aka Rosenthal effect

Answer 11

the greater the expectation on placed on people, the better they perform

Question 12

Lead-time bias

Answer 12

prolongation of survival after applying a screening test w/out any real effect on prognosis

Question 13

Hawthorne effect

Answer 13

tendency of a study population to affect an outcome due to the knowledge of being studied