The measurement of specific heat by heat flux differential scanning calorimetry using the “ratio method” requires three consecutive runs (baseline, reference and sample) to be performed using identical temperature programs. Traditionally measurements are taken as the sample of interest is subjected to a controlled heating rate, but it is also possible to perform measurements on down-ramps, and this is often worthwhile as it provides a consistency check on the up-ramp data. In recent work a Netzsch DSC-404 has been used to investigate discrepancies observed between up- and down-ramp specific heat results. Based upon this experience several different “baseline correction” techniques have been developed, and have been applied to specific heat measurements on reference sapphire samples. The performance of these “baseline correction” techniques is shown to be extremely good, even up to 1400°C. An important conclusion is that, even in cases where agreement between up-ramp and down-ramp data is reasonable, a “baseline correction” method can still usefully be applied to improve the precision of the final results.