To identify observed plasma and magnetic field discontinuities in interplanetary space as magnetohydrodynamic shocks requires fitting the plasma and magnetic field data to Rankine-Hugoniot conditions. By statistics and specific samples we show that unless the fluctuations in the observed parameters in the preshock and postshock regions are seriously taken into consideration, careful fitting may still lead to erroneous identification and description of a shock. An approach to the unbiased choice of the preshock and postshock values is to systematically consider all possible combinations of preshock and postshock parameters averaged over different time intervals measured away from the shock.