The threats to validity that affect clinical research 35 fall into three broad areas: chance, bias and 'generalizability'. The threat from chance, referred to as overfitting, is discussed in Ref ...