The last five years, important progress has been made in understanding risk set sampling designs - like nested case-control studies - and in developing new and useful designs and statistical methods for sampled cohort data. The main purpose of this report is to present a fairly non-technical review of this development. We also illustrate the use of the general methodology in two particular situations: (i) a study on the effect of radon and smoking on the risk of lung cancer deaths among a cohort of uranium miners, and (it) introduction of a "neighborhood-matched counter-matched" design particularly developed to investigated factors which may explain the observed association between "very high current configuration" power lines and childhood leukemia. The report will appear as a chapter in the book Recent Advances in the Statistical Analysis of Medical Data to be published by Edward Arnold Publishers Ltd. with Brian Everitt and Graham Dunn as editors.