The Panel on Decennial Census Methodology met during the year 1984 to investigate the following technical issues related to the 1990 United States census: (1) adjustment of census counts and characteristics; (2) sampling; and (3) administrative records. A number of recommendations were made. It was suggested that a low priority be given to research and testing directed toward taking a sample survey instead of a census for the population count and basic characteristics; that the costs of each stage of followup and the feasibility of sampling for followup be studied; and that machine-readable records be tried for a sample of the data. Further testing of telephone followups and further research in the area of verification of content were also supported. Suggestions for alternate ways of obtaining housing data were made. Further study of the feasibility of a two-stage data collection, using the short form and long form, was encouraged. Further studies evaluating coverage were strongly recommended, while it was felt that the pretest of the post-enumeration study methodology was too ambitious. Continuation of the Forward Trace Study was supported. Other issues of reliability, ancillary data, statistical estimation, and operational constraints were discussed. (GDC)