Biostats mod 6 Flashcards
what does t-test test
difference between 2 means
how many groups can a t-test be used for
1 - is group different than expected mean
2 (more common) - are t groups different
paired samples t-test
same group before and after
4 things given in t-test output
T value
df - number of values that can vary
sample estimates - means
probability - probabliity that difference is due to chance
what kind of data must you have for t-test
continuous and normal dist.
what is ANOVA used for
test difference between 2 or more groups
what kind of data must you have for ANOVA
continuous and normal dist.
what does ANOVA tell you
that there is a diff. between groups, but not which
what do you get on ANOVA output
F-test - anova value
SS - sum of the deviation calculated from each value of the outcome variable
Mean square - SS/df
residuals - leftover from explanatory variables
df - same as t-test
probability - compares F-test relative to df
when is CHI square used
analysis of contingency tables
what does chi square show
predicts expected value in each cell of contingency table and the determines if found values are differnet
data type of chi square
2 or more pops and 2 or more outcome variables
what types of variables in chi squre
outcome must be categorical with 2 or more groups - no assumption of normality
what are trying to show with linear regression
association of one continuous variable with another cont. variable
what kind of data needed in linear regression
at least one population and 2 variables
what kind of variables needed for linear regression
continuous and normal
what is intercept and slope of linear regression lin
intercept - where is crosses Y axis
slope - rise/run
what is regression formula
outcome = intercept + (regresison coefficient*explanatory variable) + error
what is multiple regression
use multiple IVs to predict DV
3 thing on linear regression output
estimate - refers to the regression coeffcient
pr - p-value for each coefficient
F-stat - test of overall difference
when is logistic regression used
show how the change in one explanatory variable (cont. or cat.) affects the status of the outcome variable (always cat.)
what is survival analysis
determine if your intervention or explanatory variable is associated with the outcome of interest from a defined start date
what kind of data for survival
data from 2 populations (exposed vs. not exposed)
what kind of variables for survival
usually need categorical explanatory variable. time is the outcome variable
