The methods used to analyze and interpret clinical trials of chemotherapy may have a major impact on the conclusions that are drawn in papers reporting them. To illustrate this problem, we constructed a hypothetical clinical trial in which patients with metastatic cancer were treated with chemotherapy. The following two articles provide reports of this trial, analyzed by methods that we would interpret as being of low and high quality, respectively. In the present report, we describe briefly the methodological differences that led to the opposite conclusions based on this single set of data. The errors of reporting and omissions of the first article (A) are similar to those that have been extracted from recent issues of the Journal of Clinical Oncology and other leading cancer journals, although they have not all appeared within a single report. This demonstration model illustrates problems in the reporting of clinical trials and suggests guidelines for improved reporting.