Re: [R] mgcv pack- gam() - question about the by argument

2008-10-28 Thread Gustaf Granath
Thanks!! The R-community is just amazing. Two follow-up questions: 1) IF, I had sufficient amount of data. How would that interaction term (Site x Time) be specified? Maybe like this? (adding stuff on the previous example) And I should look at the Parametric coefficients: for significant

[R] mgcv pack- gam() - question about the by argument

2008-10-22 Thread Gustaf Granath
R fellows, I hope my questions are not too much about statistical methodology. I do lab experiments where y is measured at different x, using (plant)material originating from different sites (2-4). The response is non-linear in a way that I have turned to gam() (mgcv package) to investigate