The present study investigated methodological differences between the clinical version of the Stroop Color and Word Test and the computerized single-trial version. Three experiments show that different presentations of the Stroop task can produce different levels of interference. The 1st experiment examined the effect of blocking; the 2nd experiment examined different control conditions. Greater interference in the blocked clinical version appears to result from lower response times (RTs) in the neutral condition, not from greater RTs in the incongruent condition. Experiment 3 examined the impact of shifting attention across locations while responding to Stroop stimuli. The present set of findings sheds light on the inconsistency in the clinical literature and demonstrates that the method and selection of neutral stimuli (that provide the baseline by which interference is measured) are critical because they clearly can change performance.