Lecture 6: Regression Analyses Flashcards
Which parametric test do I do if both the dependent and the independent variables are continuous?
linear regression
Which parametric test do I do if the dependent variable is categorical and the independent variables is continuous?
Logistic regression
Which parametric test do I do if the dependent variable is continuous and the independent variables is categorical?
ANCOVA
Which non-parametric test do I do if both the dependent and the independent variables are continuous?
Non-parametric regression
Which non-parametric test do I do if the dependent variable is categorical and the independent variables is continuous?
Non-parametric regression
Which non-parametric test do I do if the dependent variable is continuous and the independent variables is categorical?
Non-parametric regression
Simple linear regression
one dependent variable and one independent variable
Multiple linear regression
one dependent variable and one independent variable
Bo (y-intercept)
point where regression line crosses the y-axis
B1 (slope)
tells you how much y changes as you move along the values of x
What is the null hypothesis in regression?
that B1=0
a straight line meaning that there is no relationship between your X and Y values
What does the R squared value tell us?
The R squared value is telling a proportion of variance that is explained in your model. So how much variability in your outcome is explained by your X
For example, if the R squared value is 0.82 (82%), then the model is 82% accurate and this is a strong relationship. For 1 unit increase in the independent variable, the dependent variable changes by 0.82 units.
What does the B value tell us?
B value is the slope so every 1 unit increase in independent variable there is an increase in the dependent variable of that B value.
What is the least squares criterion?
method of minimizing the sum of squared residuals in a model
What is the equation for total sum of squares?
Total sum of squares = explained sum of squares + residual sum of squares
What are the assumptions of linear regression?
- residuals are normally distributed
- linear relationship between dependent and independent variables
- no extreme outliers
- ample sample size for variables in the model (at least 10 cases per independent variable)
- no multicollinearity between independent variables (predictors) (occurs when there is too strong correlation between independent variables (p>0.7)
What are confounders?
a variable that is associated with an exposure, and independent of that association, is also a risk factor for an outcome. Distorts estimate of association between independent and dependent variable.
- can be controlled for in statistical analyses, but only if the variable has been measured during data collection!
How to identify confounders?
General approach: conduct linear regression between a predictor variable and an outcome variable.
- repeat the regression with the potential confounder (covariate) now included in the model
Variable considered a confounder if:
- a p-value that was initially significant is attenuated
- B coefficient for your main predictor variable changes by more than 10%
What is residual confounding?
Residual confounding is the distortion that remains after controlling for confounding in the design and/or analysis of a study.
Dummy variable
method of dichotomizing variables
What is odds ratio?
This is a ratio of ratio (odds) it is the odds that someone who has been exposed to the risk factor actually has the disease compared to the odds that someone who has not been exposed to the risk factor has the disease
What is multicolinearity? What are the consequences?
occurs when there is too strong correlation between independent variables (p>0.7). They are measuring the same thing
EX: BMI & waist circumference
we only adjust one of them in the model otherwise you are measuring the same thing and you will have inflated statistical results and maybe overestimating the significance of the relationship between your variables.
Which test do I do if my main predictor variable is categorical?
ANCOVA
Which test do I do if my outcome variable is categorical and dichotomous?
Binary logistic regression
Which test do I do if my predictor is either categorical or continuous but my outcome is always continuous?
Linear regression
What do I have to do to my independent categorical variable before I do a linear regression?
dichotomize the categorical variable
What is least square mean?
Mean that has been adjusted for covariates.