Surveillance, Epidemiology, and End Results (SEER) program data suggest that blacks with laryngeal carcinoma have a significantly lower 5-year survival rate than whites. Most of this difference persists despite adjustment for "crude stage." To evaluate possible factors contributing to this residual survival deficit, 190 white and 23 black patients treated at the Martinez Veterans Administration Hospital between 1968 and 1988 were studied. The independent impact of race on survival rate was analyzed with respect to various prognostic factors including treatment delay, elapsed time (diagnosis to treatment), age, stage, cancer subsites, and type of therapy. No independent prognostic significance could be attributed to race. The differences noted in SEER data probably reflect a tendency for the use of crude stage to underestimate the impact of prognostic groups within the categories of "local" and "regional" disease, the independent prognostic significance of subsites (glottic versus supraglottic), and the variable distribution of these subsites in different populations. This study suggests that when stage, subsite, and quality of care are adequately considered, survival from laryngeal cancer in blacks is comparable to that of whites.