Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
20 Cards in this Set
- Front
- Back
hill's criteria have to do with what?
|
causation
|
|
if i reject the null hypothesis but the results are not statistically significant, what?
|
type I error; level of alpha is 5%
|
|
if p value is < alpha, what?
if p value is > alpha, what? |
reject null hypothesis; accept null hypothesis
|
|
if i accept the null hypothesis but the results are statistically significant, what?
|
type II error; level of beta is 20%
|
|
when do we consider type I error? type II?
|
after study; before study
|
|
what does statistical power refer to?
|
how big a study needs to be in order not to miss anything
|
|
what is the probability of rejecting null hypothesis when it is false?
|
1-beta=80%
|
|
in follow up, wide range between best and worst case scenario indicates what?
|
study is inadequate
|
|
what is berkson's bias?
|
patients getting care are not like general population
|
|
what are neyman bias?
|
errors WHEN we collect data
|
|
when you can draw a vertical line across cases in a horizontal line diagram and almost always hit one kind of disease, what does that mean?
|
prevalence-incidence bias; the disease is mil, or else it would be short because patient would die
|
|
what is it called when an effect disappears after controlling for a variable?
|
simpson's paradox
|
|
ways to control confounding factors?
|
restriction, randomization, stratification, multivariate analysis, and matching
|
|
ways to control bias?
|
rely on study design, standardize, blinding
|
|
what is stratification?
|
recall use of indirect and direct standardization
|
|
what is hawthorne effect?
|
production of workers improved solely because they knew they were being watched
|
|
what is most robust clinical trial format?
|
cross-over
|
|
advantage of factorial study?
|
can see if drugs are synergistic or antagonistic
|
|
what is nnt?
|
number needed to treat to actually treat 1 case of disease
|
|
nnt formula?
|
1/(adv. outcomes in placebo group - adverse outcomes in treated group)=nnt
|