Public health nutrition programs are intended to serve low-income families who are at greater nutritional risk than the general population. Not all persons who are program-eligible are at equal risk, however. It would be desirable to evaluate a program's ability to enroll persons from higher risk backgrounds in the population (coverage) and, conversely, the extent to which those enrolled in this program are at higher risk (targeting). A method for the evaluation of coverage and targeting was developed using data from the Tennessee Women, Infants, and Children Special Supplemental Food Program (WIC) linked with birth certificates. The linked computer file was created by matching the name and date of birth in both record files. The birth records were the common source of information used to characterize the risk background for both the WIC and non-WIC participants. Maternal sociodemographic information on the birth records was used to define the health risk background of each child. The coverage and targeting of "at-risk" children were computed and compared for 50 counties or county-aggregates in Tennessee. Considerable variation in the coverage and targeting rates of at-risk children was observed among Tennessee counties, although the counties within each WIC administrative region tended to have similar coverage and targeting patterns. Using the existing data in linked program and vital records provides a direct evaluation of a program. Coverage and targeting evaluation can be used to detect underserved populations within small geographic areas.