Seven Norwegian centres recruited 61 female and 54 male patients with non-ulcer dyspepsia (NUD). Their mean age was 40 years. After 6 weeks' double-blind alternating treatment with 150 mg ranitidine twice daily and placebo, 1 week of each alternative (part I), an effect score (Xs) and an efficacy index (Ei) were calculated. Ranitidine was significantly superior to placebo for symptomatic relief (p less than 0.01). Twenty-eight, 49, and 38 patients were Xs-classified as 'responders', 'unclassified', and 'nonresponders' to ranitidine, respectively. The symptomatic effect was impressive in the responder group and moderate but significant also in the unclassified group. The nonresponders had a significantly unfavourable effect of ranitidine compared with placebo. The Xs-classified responders and unclassified continued single-blind treatment with ranitidine for 4 weeks (part II) and were reclassified as 'new responders/nonresponders'. The new responders received single-blind treatment with placebo until relapse or maximum 8 weeks (part III). Parts II and III verified the good Xs classification of responders to ranitidine treatment. The overall effect of ranitidine in patients with NUD was due to good symptomatic effect in a subpopulation characterized by meal-related heartburn and/or regurgitation, large body mass index, first-degree relatives with gastrointestinal diseases, a relatively low frequency of gastrointestinal symptoms per week, and absence of soft stools.