Multiple Linear Regression

In order to use multiple regression, you must ensure that your data can actually be used to conduct multiple regression. When using multiple linear regression, the data must pass the following eight assumptions:

1. The dependent variable must be measured on a continuous scale (it must be on the interval or ratio scale). 
2. There must be two or more variables, which can be continuous (on the interval or ratio scale) or categorical (on the ordinal or nominal scale). 
3. There must be independence of observations. 
4. The relationship between the dependent variable and every one of the independent variables must collectively be linear. 
5. The data must reflect homoscadasticity, which means that the variables must have the same finite variance along the line of best fit. 
6. The data must not reflect muticollinearity, which means the data must not depict high intercorrelations among the independent variables.
7. There shouldn’t be any significant outliers, high leverage points, or highly influential points within the data. 
8. Residual errors must be (approximately) normally distributed.