*Dear R-Sig-Ecology,* *I did a GLS analysis for my experiement which investigates the abundance of bacteria influenced by the presence of contaminants and different poresizes (of the microcomsm to enter for bacteria) over a certain timeperiod (T1, T2, T3). The microcosms were filled with sand which was spiked with contaminats. As a control no contamiants were added to the sand: (contamination: contamin = contaminated sand, noncontamin = noncontaminated sand. Futhermore, there were different poresizes for the microorganisms to enter: (poresize: pore a = 2?m, pore b = 30?m). The samples were analysed after three timepoints: (timepoint: T1, T2, T3 = timepoint) * *I got an significant result for the three-way-interaction timepoint:poresize:contamination. Now I would like to show my results in a graph with a linear regression for the different interactions. But I don`t know how to calculate the intercepts as well as the slopes for the following interactions: **timepoint:pore a:contamin**, **timepoint:pore a:noncontamin, **timepoint:pore b:contamin, **timepoint:pore b:noncontamin.* *In my opinion the slope can`t be negative. * *The output was:* Coefficients: Value Std.Error t-value p-value (Intercept) -2477166 1103724 -2.244371 0.0329 timepoint 2477158 1103599 2.244617 0.0329 poreb -49849874 6934889 -7.188273 0.0000 contaminnon 2407662 1109466 2.170108 0.0386 timepoint:poreb 49854182 6934107 7.189705 0.0000 timepoint:contaminnon -2407655 1109341 -2.170346 0.0386 poreb:contaminnon 48366511 6970970 6.938275 0.0000 timepoint:poreb:contaminnon -48357964 6970184 -6.937832 0.0000 Correlation: (Intr) timpnt poreb cntmnntmpnt:p tmpnt:c prb:cn timepoint -1.000 poreb -0.159 0.159 contaminnon -0.995 0.995 0.158 timepoint:poreb 0.159 -0.159 -1.000 -0.158 timepoint:contaminnon 0.995 -0.995 -0.158 -1.000 0.158 poreb:contaminnon 0.158 -0.158 -0.995 -0.159 0.995 0.159 timepoint:poreb:contaminnon -0.158 0.158 0.995 0.159 -0.995 -0.159 -1.000 *I am in urgent need of help. Somebody please give me a valuable suggestion for my problem. Thanks a lot,* Caro
urgent help required regarding GLS - forgot to send the output in my previous email
1 message · Caroline Dreßler