We have developed an empirical technique for measuring perceived flicker on refresh displays. This technique is based in vision science, is technology independent and has proved to be a useful tool in display development and product assurance. It has been used to evaluate a large number of displays varying in size, luminance, refresh rate, phosphor decay and colour. In this paper we describe the flicker-matching technique (FMT) and provide detailed information on its operation. In the IBM flicker-matching technique, observers of known flicker sensitivity compare the amount of flicker on the display to the amount of flicker on a standard light. The temporal frequency of the standard light is then varied until the display and light appear to flicker equally. We measure this ‘matching frequency’ at three levels of display contrast. In the ‘standard test’ version of this technique, we distil a figure of merit for each display based on the performance of observers in the top 5 per cent of the population. This figure of merit corresponds well with field reports and success in receiving a ‘safety label’ from the German TCA (Berufsgenossenschaft). © 1986 Taylor & Francis Group, LLC.