How common are conditional biases that affect presentation and response timing?
Unfortunately they have existed in every paradigm we have looked at regardless of platform, operating system or whether a commercial experiment generator has been used or not. In all cases the researcher has been unaware of the presentation of response biases. More worryingly on many occasions there has been a conditional bias caused by the equipment itself. For example in cross-modal studies it is common with trails where audio is present to have markedly different response timing characteristics. This can account for apparent statistically significant experimental results leading to spurious conclusions. In the majority of cases timing errors can be improved by use of the Black Box Toolkit. Either the experiment parameters can be altered or post-hoc statistical manipulation can be carried out where response data has a low variability and constant absolute error.
