Unexplained postcensus treatment of data from the 1980 census to some extent has hurt the prestige of national population estimates in Mexico. Deficient vital statistics coverage has led to underestimation of deaths of approximately 5% and overestimation of births of approximately 10% over the past decade. The 1980 census suffered from a higher degree of underenumeration than the census of 1990. Several attempts have been made to correct 1980 census estimates at different levels of disaggregation using data from the 1990 and 1970 censuses. The National Institute of Geographic Studies and Informatics estimated the 1980 population at 64.9 million and the 1990 population at 81.4 million, similar to the 81.2 million actually enumerated in the 1990 census. The method of demographic reconciliation is used to correct data on age and sex structure gathered at different time periods, so that the results are consistent among themselves and with estimates of fertility, mortality, and migration. At times, when the basic data are deficient, demographic criteria must be used to select the most reasonable results. A demographic reconciliation was carried out for 1960-90 in which the census of 1980 was not considered, and data from UN model life tables, recent fertility surveys, and a study of international migration were used. The results indicated that the 1990 census had an omission rate of 3.6% for men, 2.1% for women, and 2.9% overall, making it the most complete census conducted in Mexico and one of the most complete in Latin America. The 1994 population was estimated at 91,222,077 based on the corrected 1990 results. 36.3% were under 15 years old and 4.4% were 65 or over.