Probability and Statistics Flashcards
2 ways a study can screw the pooch
caused by chance (random error)
not caused by chance (bias or systematic error)
what deals with random errors
statistical inference
T/F random errors bias a study
false
may be wrong but not biased
what is a systematic error
error that is inherent to the study method being used and results in a predictable and repeatable error for each observation
T/F systemic errors bias a study
true
T/F there is a formal method to deal with systematic errors
false
tests of statistical inference
estimate the likelihood that a study result was caused by chance
T/F statistically significant means clinically important or meaningful
false
what is a chance occurrence
something that happen unpredictably without discernible human intention or with no observable cause
random variation
there is error in every measurement
if measure something over an over again will get slight variations in measurements
what does statistical inference tell us
if we measure something only once, how sure are we that our measurement has been caused by chance
what 2 methods are used to estimate random variation in a study
confidence intervals
p-values
T/F width of the CI is related to sample
true
- small samples have large CI
- large samples have small CI
T/F if the 95% CI for the OR does not include one the OR is statistically significant
true
T/F the same rules apply for OR, PR, RR when it comes to CI
true
if CI spans one they are statistically insignificant and there is no association
what do P-values estimate
whether a measured association was likely to have been caused by chance
Does P-value give you information on size of sample and range of true value
it sure as shit doesn’t
to be statistically significant p-value must be less than
0.05
how do you calculate p-value
chi-squared test
student’s T-test
correlation
null hypothesis
hypothesis of no association
alternative hypothesis
the research question
there is an association between exposure and disease
T/F we use p-values or CIs accept or reject the null hypothesis
true
wha is a type 1 error
false positive
- reject the null when it is not false
- saying there is an association when there isn’t
what is a type II error
false negative
- not rejecting the null when it is false
- saying there is no association when there is
types of data
categorical
continuous
categorical data
broken into discrete categories
- nominal
- ordinal
continuous data
variable is numeric and can have any one of many possible values
nominal
named, not ordered
horse vs. donkey; stallion vs. gelding vs mare vs colt vs filly
ordinal
named ad ordered by nor constant value between ranks
neonate vs juvenile vs adult vs geriatric
describe categorical data
frequency distribution
may be represented as a table or bar chart
statistical test
describe continuous data
frequency distribution and histogram
describe the center of the distribution
describes the amount of dispersion
describe the shape of distribution
statistical tests
what is central tendency
describes the center of the distribution
mean, median, mode
what is dispersion (spread)
describes how closely the values are gathered around the center of distribution
range, standard deviation
Chi-squared test
test of independence between 2 categorical variables
p-value
used for categorical data
student’s T-test
difference in means
compare averages of 2 groups
used for continuous data
H0=means of 2 groups are the same
Ha= can be one or two tailed
correlation
measures the strength and direction of a linear relationship between 2 continuous variables
correlation coefficient (r)
often used for dose response relationships
both variables are numerical, usually continuous
- strong: r>0.80
- weak: r
