World rice production
Use this diagram to show how to model a linear trend.
Click Show residuals, drag the line to make the residuals small, then click Least squares.
Show that the intercept is very large when Year is used as an explanatory variable. Use the pop-up menu at the top to change the explanatory variable to (Year - 1960) and show that the least squares slope and the fitted values remain the same but the intercept is smaller.
(This trick is more important when fitting polynomial models such as the quadratic model on the next page where use of the raw Year makes the parameter estimates numerically unstable.)
The data set gives the total world rice production (million tonnes) between 1961 and 2001.