The application of correction factors for maternal serum alpha-fetoprotein (MSAFP) is expected to increase the accuracy of this screening tool. Correction for maternal weight compensates for the dilution effect of a larger plasma volume in women of greater weight. We analyzed the effect of a previously published linear correction formula for weight on the frequency of abnormal MSAFP results. Serum samples from 8276 patients were studied for AFP and were grouped according to maternal weight in 50-lb increments. Abnormal results were defined as 0.4 or less or 2.5 or more multiples of the median for gestational age. Without correction, the highest rate (15%) of low MSAFP results was obtained in the obese population. Conversely, the rate of elevated MSAFP was highest (3.8%) in the lowest maternal weight group. Correction for weight up to 250 lb significantly increased the rate of abnormally high results in the obese group, indicating an overcorrection effect. The number of amniocenteses indicated for abnormal MSAFP was reduced by about 9% with weight correction. We suggest that linear correction of serum AFP results is adequate for maternal weights up to 200 lb. Results in women weighing more than this upper limit should be corrected as if the weight were 200 lb only. This balances the frequency of abnormal results in all weight groups and maintains the reduction in number of procedures performed.