1. If you use a Linear Least Square fitting program or try to apply the standard formulas,
they ASSUME that the uncertainty in ONE of the variables is negligible and the uncertainty in
the other variable is constant. Usually it is the "x" variable that has negligible uncertainty.
2. The validity of a straight line fit is based on how well the data actually fit the straight line.
If the data are too sparse then the straight line fit may inspire little confidence.
3. As a calibration, one should only use it within a margin of the range of the data.
4. For a meaningful extrapolation from y to x, one must determine the uncertainty of the
extrapolated value (this will involve the uncertainty in y, AND the uncertainties in the fitted
slope and intercept.)
5. In the particular example, the standard value of r @ p=0 was omitted. Why?