Post date: Oct 09, 2016 5:30:54 PM
Here are three figures that further explore/summarize the effect of temperature mentioned here. First, one quick note on a2. This is the overall effect of temperature, but things are a bit more complicated than that. Specifically, we assume a base effect of temperature, that is modified by host (coded as A = 0, C = 1) and mountain range (coded as not refugio = 0, refugio = 1). So, this is really the effect of temp. for bugs on A and not on refugio. You need to add c2 or d2 or both to get the C, refugio or both. So, with that said, I included two scatterplots, one that shows the best fit line based solely on a2 (and a1 as the intercept), and a second one that gives lines for A and C (orange vs. blue) on refugio (solid line) or not (dashed lines). Points are similarly colored with filled circles for refugio. I also included the posterior for a2 with a vertical line at 0. One other quick note, temperature (as well as elevation, which is not shown) has been centered and standardized (i.e., mean = 0, s.d. = 1), hence the scale on the x-axis.