The purpose of this study was to investigate the differences between the administrative record data for establishments which were non-respondents to the 1977 Censuses of Wholesale Trade, Retail Trade or Selected Services and the "reported" data for these establishments acquired by intensive follow-up procedures. The table below summarizes the findings. The ratio estimator of "reported" data to administrative data exceeded one for all trade areas and for each variable studied. This pattern was also observed when the ratio estimators we calculated for single unit establishments and multi-unit establishments separately.