Diagnostic First, Choice Second: How I Do Reproducible Social Data Science

When I cleaned the corpus for a study on how the press frames African elections, I had to decide what counted as a duplicate article. Wire stories get syndicated across dozens of outlets, so the same