The aim of this study is to study a possible social gradient for testicular cancer in Norway, and assess whether the gradient has changed since 1960. Testicular cancer has traditionally been considered a cancer of the affluent. The previous register-based studies are conflicting on whether there is a gradient, if the gradient is shifting and if there is a difference between seminomas and non-seminomas. We have used a linked file with data regarding social status from the censuses and data from the Cancer Registry of Norway regarding the occurrence of testicular cancer. We conducted a Cox-regression analysis of testicular cancer among all men >= 25 years when participating in the censuses of 1960, 1970 and 1980 and data extracted from the Personal register in 1990. For 1990, we did not have occupational data. The analysis was done for the following ten year intervals after each census. We found an increased risk associated with higher education during all periods in our study with an odds ratio over two when comparing the highest and lowest educated. Using occupational status we did only find an increased risk during the period 1981-1990, the period with both a large number of cases and occupational data. Even during that period, the effect of education was far more pronounced than occupation. The results were similar for seminomas and non-seminomas.