Biostat Flashcards
what is sensitivity?
probability that person with disease has a positive test. if negative rule out disease
what is specificity?
probability that person without disease has a negative test. if positive, rule in disease
RR?
a/a+b/c/c+d
OR?
ad/bc
what is precision testing and what reduces precision?
reliability. random variation reduces precision.
what is accuracy testing and what reduces accuracy?
validity. systematic error (bias) decreases accuracy
increase in precision affects what 2 values?
increase power, decrease std deviation
odd ratio calculated in what observational study?
case control (looks at endpoint (disease) and than tries to determine risk factors)
relative risk calculated in what observation study?
cohort (has a risk factor and looks back to see who developed disease with risk factor)
when does OR=RR?
when outcome is uncommon in a population
what helps control confounding variables?
matching such that both groups have similar distribution (eg age, race) in accordance with those variables
biases pose a threat to what of the study?
validity
methods to decrease confounding bias?
matching, restriction, randomization
attributable risk?
a/a+b-c/c+d
relative risk and incidence can only be calculated from what?
prospective or experimental trials (ex prospective cohort)
what values significant for OR and RR?
any value besides 1
odds ration can be calculated for what?
retrospective (ex case control)
another name for prevelance surgery?
cross-sectional
what is an epidemic?
observed incidence greatly exceeds expected incidence
define and give an example for nominal, ordinal, and continuous types of data and type of test it is associated with?
nominal-no numeric value (eg day of week)
ordinal-ranking but no quantification (eg class rank does not specify how far number 1 is ahead of you)
continuous-numerical measurement
chi squared (nominal or ordinal), t-test or anova (continuous)
define p-value?
that the data were obtained by random error or chance
what correlates to chance of making type 1 error?
p-value because it claims that an effect or difference is presence when none really exists. so if p is 4%, there is less than 4% chance making type I error
what is type II error?
saying there is no difference (null hypothesis accepted) when one actually exists
what is power and how do you increase?
probability of rejecting null hypothesis when it is actually false. increase sample size