However, even with a latent variable approach using raw RTs, the correlation was only slightly improved. While correlations of such magnitudes had previously been found in some studies, measures reflecting the same competency should be expected to correlate strongly and positively. Explicit tests of a unitary-construct model favor the interpretation that the Stroop interference. Similarly, even when the Stroop interference measures used were the more comparable adjusted RT measures. Even though the largest observed correlation surpassed that found in some previous studies, the ratio of common to unique variance between the two measures indicates way more divergence than convergence. We have found no compelling evidence of a robust relationship between Stroop and stop-signal RT measures of inhibition. But what does this mean? Stroop and stop-signal RT measures may indeed be capturing different competencies. It may also still be possible that they involve the same inhibitory mechanism, but with some major differences in how the mechanism is executed or is reflected in dependent measures. However, that their behavioral measures are not strongly related poses empirical and practical problems when the tasks are often used synonymously in the literature to assess the same cognitive functions. The present data suggest that Stroop inhibitory ability, is at least moderately related to stop-signal inhibitory efficiency. However, inhibitory efficiency measured by RT scores of the two tasks should not be taken to index similar constructs or LY2835219 processes–a participant may be classified as a poor inhibitor when measured on one task yet do well on the other. In addition, all versions of our Stroop interference RT measures were not significantly correlated with Stroop errors. Hence, though accuracy and RT both reflect inhibitory performance on the Stroop task, they likely reflect different competencies. The low reliability of conventional measures of Stroop interference is problematic and warrants further research in identifying a more reliable measure of Stroop inhibition. The inverse efficiency score shows some promise with its substantially higher reliability and its attempts at reconciling speed with accuracy performance. However, its reliability is still less than satisfactory and its suitability as a measure of inhibitory performance needs to be further examined. Some doubts that have been raised about the IE score include the increased variability associated with multi-component measures; the limits to its application when there is not a high, positive correlation between RT & accuracy; and most importantly, whether dividing RT by percentage accuracy is the most appropriate way of reconciling the two aspects of performance. Cognitive studies are often interested in comparing performances on conditions comprising as little as 20 to 30 trials each. A small difference in the number of errors between conditions can translate to a large difference in terms of percentage accuracy. The effect of errors on IE scores is hence not linear but accelerates with lower percentage accuracies.