OBJECTIVE: Social class mortality differences in New Zealand men aged 15-64 years have previously been examined for the periods 1975-77 and 1985-87 using the Elley-Irving social class scale. The objective was to repeat these analyses for 1995-97 in order to examine time trends, and to assess current social class patterns of mortality. METHODS: Age-standardised mortality rates were calculated for each social class and a weighted estimate of the social class mortality gradient was obtained. RESULTS: Male mortality declined 21% between 1985-87 and 1995-97, but the social class mortality differences have not diminished and may have even increased. The Relative Index of Inequality has increased from 1.8 in 1975-77 to 2.1 in 1985-87 and 2.3 in 1995-97. Unlike previous analyses, the relative social class mortality gradient was just as strong in the older age groups as in the younger age groups, indicating that the possible increase in social class gradient has largely occurred in the older age groups. CONCLUSIONS: These findings indicate that the potential to address the excess preventable mortality caused by socio-economic factors has not been fully realised in New Zealand. IMPLICATIONS: Social class analyses identify groups in the community that have an excess mortality that is potentially preventable. There are still major social class differences in mortality in New Zealand, and these differences may even have increased. It is important that these patterns are taken into account in public health planning and that further research is conducted to identify the mechanisms by which these differences occur.