1. The Simple Regression Model (SRM) requires that a histogram of the response looks like a normal distribution.
2. In the ideal situation, the SRM assumes that observations of the explanatory variable are independent of one another.
3. Errors in the SRM represent the net effect on the response of other variables that are not accounted for by the model.
4. The errors in the SRM are the deviations from the least squares regression line.
5. To estimate the effect of advertising on sales using regression, you should look for periods with steady levels of advertising rather than periods in which advertising varies.