Weeks 1 And 2 Flashcards
Pearson correlation
Linear regression for Gaussian data
-null is that the population correlation is 0 (no linear relationship)
Spearman correlation
Linear ranking method used with extreme values and applied to regression for a more Gaussian look
Regression equation
Y hat= bx+a
When there are Equal SD for regression, blank = blank
B=r
Multivariate
Techniques using multiple factors to remove confounding (except analysis of variance)
Multiple linear regression
For predicting a numeric dependent variable with multiple variables
Logistic regression
Used for 2 levels of dependent variables (yes/no, success/failure, etc)
Proportional hazards
Dependent variable is the time until a certain event
Relative risk
How many times more likely it is for 1 outcome vs another
Odds ratio
An approx of relative risk
Used in logistic regression
Hazard ration
Approx of relative risk, used in proportional hazards
Mean duration of survival
Best if all subjects die; mean amount of time they live
Median duration of survival
How long pts live, works better with censored data
Case fatality rate
% of deaths from condition
5 yr mortality rate
Proportion of deaths in a 5 yr period
Mortality rate per person yrs of observation
of deaths/total pt years (alive and dead)
Survival curves
Kaplan Meier, life table
Censored events
If event of interest doesn’t occur by the end of the study or there is a competing cause of death, etc
True experiment
Randomized design
Crossover trial
Each subjects gets 2 or more tx (each subject is his own control)
Equivalence trial
Shows that 2 tx or equivalent or close enough
Non inferiority trial
1 tx is not worse than an existing tx
Quasi experimental design
Strong element of control BUT no random assignment of individuals to groups
Single subject multiple baseline
Quasi exp
-many observations, intervention, many observations