Regression Analysis

Regression analysis helps answer questions such as: how much do sales rise when advertising spending increases? How much does a share price tend to move when the overall market moves? How does house price relate to size, location and age?

The analysis fits a mathematical relationship to data. The simplest version, linear regression, finds the straight line that best fits the data points. The result shows the estimated effect of each independent variable on the dependent variable, along with measures of how well the model fits and how reliable the estimates are.

For example, a retailer might find that each additional 1,000 spent on advertising is associated with about 4,000 of extra sales, after accounting for seasonal effects.

Regression is widely used in finance for forecasting, estimating risk measures such as beta, valuing assets and testing economic theories.

Like correlation, regression shows association, not necessarily cause and effect. Results can be misleading if important variables are left out, if the relationship is not really linear or if the data contain unusual values. Careful interpretation is essential.