1.
Treatment-effect identification without parallel paths: an illustration in the case of Objective 1–Hainaut/Belgium, 1994–2006
2.