On the analysis of clonogenic survival data: Statistical alternatives to the linear-quadratic model