This paper presents a novel methodology for measuring the drag latency of touchscreens. Drag latency is a characteristic of touch hardware that can affect key metrics of human performance as typically collected in human-in-the-loop (simulator) experiments. The proposed methodology uses a laser and laser sensor to obtain an external measurement of stylus position that is compared, in time, with the digitally measured touch event data. The drag latency is estimated as the time shift required to align the locations of touch sensor activation with the recorded laser beam crossings. Application of this methodology on two touchscreens from different vendors showed that touchscreen drag latency is in general strongly dependent on the input speed and varies between 28 (fast inputs) and 200 ms (slow inputs) for the tested hardware. Furthermore, while drag-latency characteristics seem to generally follow an exponential decay curve as a function of input speed, the latency was found to differ by a factor 2 at low input speeds between both touchscreens. Using the measurements from our proposed methodology to obtain predictive models of touchscreens' drag-latency dependency on input speed was found to facilitate accurate compensation for the drag latency for three different input tasks (chirp, multi-sine, and step) relevant to human-in-the-loop testing. The results suggest that measuring, and compensating for, touchscreen drag latency is essential for human-in-the-loop experiments, to ensure consistent measures of human performance and to enable replication of measured effects.