Studies already do that, it's called pack-years. If your friends do 1/12 a pack per day for 12 years, that'd mean 1 pack year of smoking. I'm analyzing data fo

Studies already do that, it's called pack-years. If your friends do 1/12 a pack per day for 12 years, that'd mean 1 pack year of smoking. I'm analyzing data for an epidemiological study now, we have such a binary smoking status indicator that you critique. Our issue is the data source has immense complexity, 3500 plain text entries on smoking status questions, it would take weeks of painstaking effort to develop the measure you speak of...whereas the current measure took us a morning to implement as someone had already done the legwork for 0/1 smoking or not.