Regression (pt. 1)

Conceptual resources

• Why we use “least squares” regression instead of “least absolute deviations” regression

Regression calculation demonstration

## regression demonstration
x <- mtcars$wt
y <- mtcars$mpg
n <- nrow(mtcars)

## find a and b

Multiple linear regression

mtcars

## conduct bivariate linear regression

## conduct multiple linear regression

## evaluate model performance

Assessing assumptions

## load packages: dplyr, MASS, car, haffutils

## assumption 1: interval or ratio data

## complete kitchen sink regression with mtcars

## assumption 2: Linear relationship

## assumption 3: Normality of residuals

## assumption 4: homoscedasticity of residuals

## assumption 5: limited multicollinearity

## assumption 6: independent events

Using stepwise regression and assessing assumptions

## load packages: dplyr, MASS, car, haffutils

## assumption 1: interval or ratio data

## complete stepwise regression on mtcars

## assumption 2: Linear relationship

## assumption 3: Normality of residuals

## assumption 4: homoscedasticity of residuals

## assumption 5: limited multicollinearity

## assumption 6: independent events