RTs for correct trial types, CRN amplitudes, and N2 amplitudes and latencies were subjected to ANOVAs with stimulus conflict (Hs, Ls) and response conflict (Hr, Lr) as repeated measure factors and diagnosis (controls, alcoholics) as a between-groups factor. Because the number of years of education was significantly different between groups, it was used as a covariate in all the statistical analyses meeting the homogeneity of variance assumption. Univariate ANOVAs were used to investigate error trial RT and ERN amplitudes with diagnosis as a between-group factor and years of education as a covariate; however, because years of education did not correlate with all the dependent variables, we also report the results for each variable without using years of education as a covariate. Correlation analyses were conducted separately for the control and alcoholic groups. Pearson correlations were conducted between the amplitudes of the fronto-central negativities (ERN and CRN) and (1) demographic variables [age, SES, NART IQ, total lifetime alcohol consumed, and number of days abstinent], (2) accuracy, and (3) RT measures (RTs for error, correct, post-error, and post-correct trials). RTs for correct
