The statistical methods used in occupational health studies were evaluated by analyzing the papers published from 1986 to 1990 in the proceedings of the annual meetings of the Italian Society of Occupational Medicine and Industrial Hygiene, in order to calculate the degree of understanding of the readers provided with only one-sided knowledge of statistical subjects and to improve the educational objectives in postgraduate schools of occupational health. Almost 70% of the 1151 articles reviewed contained some kind of statistical analysis: methods more complicated than descriptive statistics were used in about 35%. Student's t test (15%) and chi square (12%) were the most common methods. Other methods were less frequently used, so that it was possible to estimate that the learning of any new method would improve the understanding of about 1-2% of the articles. A wider use of statistical methods in data analysis is recommended; the attainment of a higher level of statistical knowledge should be a priority target in occupational health training.