only marginally significantP = 0.055
Although the main effect of atomoxetine on decision noise was only marginally significant, F (1,252) = 3.71, P = 0.055, there was a significant interaction of treatment with horizon, F (2, 252) = 4.73, P < .01, indicating that the increase in decision noise from baseline to longer horizons was smaller in the drug condition than in the placebo condition ( Fig 3b and 3c ; Table 1 ).