I have been asked to comment on the analysis of a data set which has been modelled as two straight lines, intersecting at a "critical" value. The scientist involved has calculated a single R-squared for the fit of all the data.
Can list members give me some pointers?
- Is this common practice?
- references to similar usage of R-squared.
- references to any discussion of this usage of R-sqaured
- Any other (accepted) means of quantifying goodness of fit for models consisting of two (or more) parts joined at a discontinuity.
Thanks Alan Collins
|