There is increasing information on the importance of genetic polymorphisms in human genes. Polymorphisms occur on average once every 500-1000 base pairs in the human genome and are useful in the identification of genes involved in human disease. Some genetic polymorphisms have functionally significant effects on the gene product and are the most useful type of polymorphism in disease association studies while others are simply useful markers. There are two main approaches using polymorphisms in the identification of genes involved in polygenic diseases. The first involves examining inheritance patterns for genetic polymorphisms in family studies and the second case-control studies which compare genotype frequencies for candidate disease genes in unrelated individuals with the disease and healthy controls. Use of family studies is generally the preferred approach but this is only feasible if the genetic component of the disease is relatively strong, DNA samples are available from other family members and the disease is relatively easy to diagnose and is not stigmatized. Population case-control studies are useful both as an alternative and an adjunct to family studies. When performing case-control studies factors such as study design, methods for recruitment of cases and controls, functional significance of polymorphisms chosen for study and statistical analysis of data require close attention to ensure that only genuine associations are detected. To illustrate some potential problems in the design and interpretation of association studies, some specific examples of association studies on drug response and on disease susceptibility involving receptor genes, cytochrome P450 and other xenobiotic metabolizing enzyme genes and immune system genes including TNF-alpha, IL-10 and the IL-4 receptor are discussed.