I have a dataset with poll opinions. The respondents have 4 types of responses: very bad, bad, good, very good. The subject of the questions asked are of three types: concern-environment, concern-economy, concern-personal. So, the dataset includes a subject column that specifies the subject of the question and the 4 corresponding responses columns. I have to find the aggregated proportion of respondents who answers very bad or bad about the economy. Please help me with the Stata code required to perform this exercise.