I should add that the data is applicable to the male half of the population. I suspect there is a geographical and occupational bias which over-represents certain watch types (e.g. divers) but I have no way of knowing from the way the survey was conducted.