This experiment tested the hypothesis that the setting in which research is conducted may dominate symptom magnitude. Cough was induced with inhaled citric acid and its magnitude was influenced by changing the setting in 30 adolescents with asthma. Cough thresholds for citric acid were determined. The participants were assigned to a condition emphasising asthma, or a control condition, described as estimation of lemon flavors. All participants inhaled (in different order) thresholds for citric acid, 50% of these thresholds, or placebo. Results showed that both cough frequency and subjective 'cough tendency' were significantly higher in the asthma condition. State anxiety was higher in the asthma condition, but correlated only moderately with cough. Lung function, severity of asthma, trait anxiety, age or sex did not correlate with cough. It was concluded that patients with asthma cough more often in a situation which they have learned to associate with asthma.