Interpretational errors in statistical causal inference