"In an unrestricted science such as biology or geology, the effects under study are so complex that designed experiments with controls are often not possible. Quantitative models are more vulnerable and the chance of unknown factors dominating the observations is much greater."