MULTIVARIABLE DATA ANALYSIS 10/2/17 Flashcards
1
Q
What is a general linear modelling (GLM)?
A
- linear modelling is a very well established tool in statistics
- very flexible
- provides easily interpreted models = key element for epidemiology
2
Q
What is a linear model?
A
helps to indicate if there is an association between two variables
3
Q
What are nuisance variables?
A
confounders and competing exposures
- undermines the interpretation of these associations
- but can adjust for these
4
Q
On stata what does ‘regress’ mean? continuous variables
A
it is used to make linear models/linear regressions
5
Q
What do you do with regress on stata if the variables are categorical?
A
xi precedes regress
e.g. xi: regress weight i. sex
6
Q
what is a logistic regression and when is it used?
A
it is used if the outcome is binary and is categorical
e.g. xi:logistic comply not i. sex
