The SFRS is based on a coherent sample of 369 nearby star-forming galaxies.
The SFRS selection criteria were defined objectively, to guarantee that the sample spans the full ranges of three critical properties: far-infrared luminosity, the flux ratio of the 60 and 100 micron bands, and the flux ratio of the near-infrared and 60 micron bands. These three observables act as proxies for total star formation rate, for dust temperature (an indicator of the star-formation mode), and for the specific star formation rate (the star formation per unit stellar mass, a measure of the intensity of star formation), respectively.
The parent sample for our project is the PSCz catalog (Saunders etal 2000), a full-sky database of 15,000 nearby star-forming galaxies brighter than 0.6 Jy at 60 microns. By virtue of its coverage, uniformity, and far-infrared sensitivity, the PSCz is a fair representation of star-forming galaxies in the local Universe. We constructed a parameter space using the three critical measureables as the dimensions, and then binned it to capture all the variation of the PSCz within that space. By extracting a representative number of PSCz galaxies from each occupied bin of that 3-D space, we created the SFRS study sample of 369 galaxies -- a sample guaranteed to capture all the manifestations in star-forming galaxies seen in the much larger PSCz sample.
The figure at right illustrates how the selection works. The thick line in each of the three panels shows the PSCz galaxy distribution in one of the critical observables, referenced to the right-hand axis. The vertical dashed lines show the bin boundaries we imposed on the PSCz. Finally, the thin lines show the resulting distribution for the SFRS, referenced to the left-hand axis. The distributions are very similar in all three panels -- so the SFRS is representative of the much much larger dataset and by extension, of star formation in the local Universe generally.
For more details of the sample selection, see Ashby et al, 2011.