Biostats module 1-2

Question 1

def clinical vs. stats significance

Answer 1

stats - probability that occured by chance alone is less than 5%
clinical - size and relevance of effect

Question 2

what is relative risk calculation for smokers

Answer 2

incidence of cancer in smokers/incidence of cancer in non-smokers

Question 3

hill criteria for causation (8)

Answer 3

strenght of assoc, - high RR
consistency of association - multiple studies and times
temporal relat. - did cause happen before
biological gradient - dose response
specificity - do other things cause the outcome as well?
biological plausibility
coherence to existing data - do other epi evidence support
experiment - gold standard
analogy - do other similar things do the same thing

Question 4

def. populaiton

Answer 4

group from which samples taken for statistical purposes

Question 5

target population

Answer 5

supset of overall pop that you study hopes to target

Question 6

study pop

Answer 6

subset of target population that you can actually sample

Question 7

sample

Answer 7

subset of study population that is actually used

Question 8

def. non-responder

Answer 8

those that don’t fill in or don’t show up

Question 9

def. probability

Answer 9

means of quantifying the chance that something will occur

Question 10

def. validity

Answer 10

how well the test measures what is says it does

Question 11

def. external validity

Answer 11

how well that results can be generalized outside - to population

Question 12

what to take into account to determine ext. validity (2)

Answer 12

• setting of study

- selection bias

Question 13

def. internal validity

Answer 13

are you measuring what you think you are

Question 14

4 types of internal validity

Answer 14

1. face - intuitive sense
2. content - does it encompass all aspects
3. construct - does it conform to theory about the construct
4. criterion related - does it correlate with other known measures

Question 15

def. precision

Answer 15

(reliability) - are the result reproducible

Question 16

accuracy

Answer 16

does it measure the truth?

Question 17

def. bias

Answer 17

systematic error that causes deviation from the truth

Question 18

3 types of bias related to internal validity

Answer 18

1. observer - misinterpret events or uses a different scale
2. subject bias - personal beleifs on part of subject
3. instrument bias - bad ruler

Question 19

5 other types of bias

Answer 19

1. selection bias
2. non-response bias - non random loss of indivs.
3. differential measurment - different groups being tested differently
4. recall bias
5. misclassification bias - people being put in wrong catagory

Question 20

7 ways to reduce bias

Answer 20

1. standardize measurement - manuals
2. train and certify observers
3. refine instruments
4. automate intruments
5. make unobstrusve measures
6. blind participants
7. calibrate instruments

Question 21

def. random error

Answer 21

non-sytemtic error that can generally be controlled by stats