A cikin main rubutu, zan tattauna yin causal da'awar daga wadanda ba gwaji data amfani halitta gwaje-gwajen da matching. A cikin wannan appendix, zan gabatar da m sakamakon model, da kuma ayyana more daidai da yanayin da ake bukata domin causal hasashe daga observational data. Wannan babi zai zana a kan Morgan and Winship (2014) da kuma Imbens and Rubin (2015) .