The Cu, C2, and MI thresholds and geometric means are reported for each participant in Table 1. Repeated-measures ANOVA of urge-to-cough ratings showed significant within-subject effects for the three main factors of conditioning (F1,10 = 5.2, P < .05), treatment (F1,10 = 102.0, P < t10 = 3.2, P < .001), and challenge (F3,30 = 15.6, P < .001). Two-way interactions between the main factors were also significant (conditioning by treatment [F1,10 = 8.1, P < .05], conditioning by challenge [F3,30 = 4.8, P < .001], and treatment by challenge [F3,30 = 21.6, P < .001]), as was the three-way interaction (F3,30 = 7.4, P < .001). A tendency for urge-to-cough ratings to generally decrease in magnitude across the course of the experimental protocol explained in part the significant effects of conditioning and its interaction with the inhaler. When expressed as proportional effects, placebo was associated with a 37.9% decrease in urge-to-cough ratings during the conditioning phase and a 28.3% decrease in the testing phase.