The location of the red edge of the ZZ Ceti instability strip is defined observationally as being the lowest temperature for which a white dwarf with a H-rich atmosphere (DA) is known to exhibit periodic brightness variations. Whether this cut-off in flux variations is actually due to a cessation of pulsation or merely due to the attenuation of any variations by the convection zone, rendering them invisible, is not clear. The latter is a theoretical possibility because with decreasing effective temperature, the emergent flux variations become an ever smaller fraction of the amplitude of the flux variations in the interior. In contrast to the flux variations, the visibility of the velocity variations associated with the pulsations is not thought to be similarly affected. Thus, models imply that were it still pulsating, a white dwarf just below the observed red edge should show velocity variations. In order to test this possibility, we used time-resolved spectra of three DA white dwarfs that do not show photometric variability, but which have derived temperatures only slightly lower than the coolest ZZ Ceti variables. We find that none of our three targets show significant periodic velocity variations, and set 95% confidence limits on amplitudes of 3.0, 5.2, and 8.8km/s. Thus, for two out of our three objects, we can rule out velocity variations as large as 5.4km/s observed for the strongest mode in the cool white dwarf pulsator ZZ Psc. In order to verify our procedures, we also examined similar data of a known ZZ Ceti, HL Tau 76. Applying external information from the light curve, we detect significant velocity variations for this object with amplitudes of up to 4km/s, making it only the third DAV for which velocity variations have been measured. (abridged)