How Participants Respond to Computer Delays
Désirée Scholz, Linda Graefe, Thomas M. Prinz
TL;DR
This study investigates whether unintended computer delays distort reaction-time measurements in a simple two-number comparison task. Participants were randomly assigned to four delay conditions (constant, low variability, high variability, and control) in an online setting, with measured metrics including system and user response times, error rates, repulsion, and demographic covariates. Across analyses, the primary hypotheses—that greater delay variability would increase reaction times, errors, and repulsion—were not robustly supported, though age and gender showed some covariate-related effects; power may have been insufficient to detect small effects. The findings suggest reaction-time measurements are relatively robust to common computer delays but highlight the importance of reporting unintended times and considering covariates in analyses for future RT studies.
Abstract
Reaction time studies with computers investigate how and how quickly participants respond to changing sensory input. They promise simple and precise measurement of time and inputs and offer interesting insights into human behavior. However, several previous studies have discovered imprecisions in timing appearing as delays, depending on the browser, software and programming used for conducting such studies. Since the accuaracy of the collected data is widely discussed, we aim to provide new results on the effect of unintended delays on participants' behavior. For this purpose, a new reaction time study was conducted. Computer delays were added to the experiment to investigate their effects on participants' performance and repulsion. Minimal changes in participants' behavior did occur and should be furtherly investigated, as the power of this study was rather low and might not have uncovered all underlying effects. The following report details our study design and results and offers several suggestions for improvements in further studies.