In addition to providing the mathematical foundation for statistical inference in nested case-control designs and variants thereof, Dr. Langholz has been exploring the improvement in statistical efficiency that can be obtained by exploiting knowledge of exposures, confounders, or surrogates for them on the full cohort. He has shown that up to four-fold gains in efficiency for estimation of exposure or modifier effects can be obtained by "counter-matching" potentially exposed cases (as determined by a surrogate) with potentially unexposed controls and vice versa. For example, job title could be used as a surrogate for occupational exposure in a case-cohort study, or wiring codes as a surrogate for magnetic fields in a two-stage population-based case-control study. Another use of these cohort sampling methods is to optimally target subjects for inclusion in validation or exposure measurement substudies. The state-of-the-art molecular biology techniques that form an important theme of our Center are often quite expensive, yet the need for large sample sizes will remain. The stratified sampling techniques we are developing (such as counter-matching) will allow analysis of all the data, using detailed information available on a sample, in a statistically optimal way.