Variation in the characteristics of irritable bowel syndrome patients recruited for clinical trials from different sources could affect their response and the generalizability of trial results. To describe and compare the characteristics of three different groups of irritable bowel syndrome patients recruited into a 'mock clinical trial.' We enrolled 245 irritable bowel syndrome patients from three sources: (i) 121 from British primary practitioners; (ii) 72 from California newspaper advertisements; and (iii) 52 from a California gastroenterologist's practice. We obtained demographic, clinical, and Hospital Anxiety and Depression (HAD) Scale data. Most patients were young to middle-aged women; the majority reported symptoms for > 5 years in all three groups. Subject characteristics varied among the groups. Typically, primary care patients were anxious, smokers and daily alcohol drinkers who had sought care recently for irritable bowel syndrome and tried antispasmodic drugs. Their symptoms were intermediate in severity between those of the other two groups. Advertisement subjects were the oldest, most highly educated, most often depressed, and were least likely to have sought care recently for symptoms, which were almost uniformly only moderate in severity. Gastroenterologist patients tended to be anxious and had nearly all sought care recently for symptoms, which were the most severe and most likely to include all three pain-related Rome I criteria. Recruitment methodology affects important characteristics of an irritable bowel syndrome study group.