Past research has demonstrated an age-related increase in the Stroop effect. Some theorists have suggested that this increase results from a decline in the ability to inhibit word information on incongruent trials, whereas others have suggested that the decline reflects general slowing. These two hypotheses were evaluated using event-related brain potentials (ERPs) measured while younger and older adults performed the Stroop task. As expected, the Stroop effect was greater for older than younger adults. The ERP data revealed a selective age-related attenuation of two modulations reflecting the inhibition of word information on incongruent trials. Latency of the P3 wave did not increase to a greater extend for older than younger adults from the congruent to incongruent trials as expected based on the general slowing hypothesis. Taken together, these findings support the inhibitory deficit hypothesis by demonstrating an age-related decline in a conceptual level inhibitory process that supports the suppression of word information in the Stroop task.