The error (-related) negativity (Ne or ERN) has been related to detecting the mismatch between incorrectly executed and appropriate responses or, alternatively, to the degree of conflict between different response alternatives. In this study different levels of response conflict were generated by manipulating task difficulty in a Simon task. According to the product of incorrect and subsequent correct EMG activation, the amount of conflict in error trials was indeed larger for the easy than for the hard condition. In contrast, Ne/ERN amplitudes did not differ between difficulty conditions, nor was the amount of conflict mirrored by Ne/ERN amplitude. Therefore, the present data are at variance with the hypothesis that the Ne/ERN reflects the degree of response conflict.