Re: Linearizing Graphs

[John Cooper <jcooper@BUCKNELL.EDU>]

On Wed, 20 Oct 1999, Michael Edmiston wrote:

> (2) If we are trying to determine if the equation (the theory) is
> upheld by the data, it is much easier to judge (by eye) if the data
> points form a straight line, than to determine if they follow a
> specific curve.
   Have you thought of plotting residuals of a function from the data, or
vice versa?  Such a residual plot against the independent data gives an
immediate visualization of the deviation from the proposed fit, whether
that proposed fit is linear or not, and strongly suggests possible
modifications of the fit.
   The Durbin-Watson statistic provides a numerical indicator that semi-
quantifies the distribution of the residuals, whether ~random,
systematically deviant.