Peng et al (2010, http://de.arxiv.org/abs/1003.4747) take a data-driven approach to study how star-formation is turned off in different environments/different mass scales using SDSS and zCOSMOS. They combine a number of results but one of the fundaments for their work is the plot above which shows the relative quenching efficiency as a function of local density in several mass bins (top) and as a function of mass in several local density bins (bottom) - this plot is for the SDSS.

They define quenching here to be the number of objects on the red sequence that would have been on the blue sequence in the lowest density environment - so it is all done relative to the lowest density environment.

Their conclusion is that the effect of environment and mass seem to be decoupled and can be treated separately - ie. that there is one source of quenching that depends on stellar mass and one that depends on local density but that they do not interact significantly. They also find the same for zCosmos at z~0.8 and postulate that it extends somewhat higher in redshift.

They are able to fit a range of observables with a very simple model and quantify the relative importance of what they call 'mass quenching' and 'environmental quenching' for galaxies of different mass at different redshifts.