Would like to see the partial residuals plotted...
Are they over fitting their model? I cannot understand how we can look at a set of data like that, see that there are, perhaps, *maybe* some associations, and then make such serious conclusions from the stats.