Cliometrics is the explicit use of economic theory and measurement in the study of economic history. As the essays in this reader demonstrate, cliometrics has become a dominant method among economic historians. One of the central tools of cliometrics, and of all econometrics, is regression analysis. Multiple regression analysis is a means of fitting economic relationships to data. It lets us quantify economic relationships and test hypotheses about them.

Examine the data plotted in Figure A.1 for the Michigan furniture industry in 1889. Each dot on the scatter plot represents one worker. Clearly, there is a relationship between age and earnings. If we could draw a line through the scatter of points summarizing what is going on, the line would have an upward slope, since earnings generally rise with age. This is exactly what regression analysis does – it selects the line that provides a best fit to the data.

Estimating this line must begin with some economic theory. Most models of the labor market assume that wages are based on the value of the worker's contribution to output. Furthermore, in many jobs, productivity of young adults rises with maturity and experience. The theory, therefore, says that earnings depend on (rise with) age. Earnings are the dependent variable; age is the independent variable. It would be foolish to argue that the causation runs the other way, that age depends on or is caused by level of earnings.