The second Radial Velocity Experiment (RAVE) Data Release (DR2) derives log g values. However, we present a simpler and cleaner method of identifying dwarfs and giants by using only magnitudes, which does not require spectroscopic analysis. We confirm the method presented in an earlier paper, with which we estimate the number of dwarfs and giants using their positions in the J–V two-magnitude diagram, by applying it to RAVE DR2. It is effective in estimating the number of dwarfs and giants at J–H > 0.4 compared to RAVE’s log g values. For J–H≤ 0.4, where dwarfs and subgiants show a continuous transition in the J magnitude histogram, we used the Besançon Galaxy model predictions to statistically isolate giants. The percentages of giants for red stars and for the whole sample are 85 and 34 per cent, respectively. If we add the subgiants, the percentage of evolved stars for the whole sample increases to 59 per cent. For the first time in the literature, we have analysed the effect of CHISQ on RAVE’s log g values (CHISQ is the penalized χ2 from RAVE’s technique of finding an optimal match between the observed spectrum and synthetic spectra to derive stellar parameters). Neither the CHISQ values nor the signal-to-noise ratio bias the RAVE log g values. Therefore, the method of identifying dwarfs and giants using the two-magnitude diagram has been verified against an unbiased data set.