This EW distribution is interesting. Comparing the high L galaxies and the low L galaxies, you can find the same trend as in my paper.
She think there is no evolution of the Luminosity function from redshift 2 to 6, but is this really true? Just based on this plot, we can see that the faint end of luminosity function may be quite the same, but there could be much different in the high L galaxies. The idea that the most massive galaxies have already formed in redshift 2 but not at redshift 3 is widely accepted. This could be the reason of this plot.
What is more, it is quite interesting that in my paper Xingxing Huang et al. 2013, there are more galaxies with high EW in the faint end. Maybe this is because that the massive galaxies are formed by less massive galaxies and they are higher metallicity and higher dust extinctions. Then if we check the galaxies higher than redshift 3, massive galaxies have not formed, and there should be more high EW galaxies as discovered in Shim et al. 2013 or other paper.