This study was conducted to assess the determinants of plasma concentrations of alpha-tocopherol, beta-carotene, retinol, and cholesterol fractions in a randomly selected subset of 203 workers participating in a worksite-based health intervention trial. Workers were from four companies in eastern and central Massachusetts, and all completed an 84-item semiquantitative food frequency questionnaire as part of baseline (preintervention) self-assessment instruments. At the time of fasting blood sampling, each participant also completed a short screening questionnaire for assessment of changes in dietary habits and tobacco exposure and for collection of data on use of vitamins and nutritional supplements. On the basis of the self-reported data, the authors found that they could explain 35% of the variability in plasma beta-carotene, 73% of the variability in alpha-tocopherol, 36% of the variability in retinol, and 19% of the variability in cholesterol. Plasma beta-carotene levels appeared to be affected by the use of supplements that did not contain carotene, indicating a beta-carotene sparing capability of other agents contained in these preparations. Plasma alpha-tocopherol levels were not similarly affected. These results compare favorably with those from studies that used more intensive dietary assessment techniques as the comparison criterion. Results are discussed in terms of implications for use of self-reported data in epidemiologic study analyses.