As the literature on heart rate variability (HRV) continues to burgeon, so too do the challenges faced with comparing results across studies conducted under different recording conditions and analysis options. Two important methodological considerations are (1) what sampling frequency (SF) to use when digitizing the electrocardiogram (ECG), and (2) whether to interpolate an ECG to enhance the accuracy of R-peak detection. Although specific recommendations have been offered on both points, the evidence used to support them can be seen to possess a number of methodological limitations. The present study takes a new and careful look at how SF influences 24 widely used time- and frequency-domain measures of HRV through the use of a Monte Carlo-based analysis of false positive rates (FPRs) associated with two-sample tests on independent sets of healthy subjects. HRV values from the first sample were calculated at 1000 Hz, and HRV values from the second sample were calculated at progressively lower SFs (and either with or without R-peak interpolation). When R-peak interpolation was applied prior to HRV calculation, FPRs for all HRV measures remained very close to 0.05 (i.e. the theoretically expected value), even when the second sample had an SF well below 100 Hz. Without R-peak interpolation, all HRV measures held their expected FPR down to 125 Hz (and far lower, in the case of some measures). These results provide concrete insights into the statistical validity of comparing datasets obtained at (potentially) very different SFs; comparisons which are particularly relevant for the domains of meta-analysis and mobile health.