How would you approach a situation where your initial hypothesis is contradicted by early experimental data?