2000 年 65 巻 534 号 p. 211-217
The author deduced the estimate households by CENSUS in 1993 from CENSUS in 1990, 1995.And then the author compared them with the statistics households by HOUSING SURVEY in 1993. The results of the comparison of these values are the following: the estimate households are large over the range of sampling errors at owned tenure, at detached owned and semidetached owned tenure, at public rental tenure and at private semidetached rental tenure in 10 large cities in HOUSING SURVEY 1993. The differences between the estimate values in CENSUS and the statistics values in HOUSING SURVEY were scattered between0.9% and 8.3% of each estimate at tenure values in large cities on Oct. 1993. The ratio of households at unknowned tenure over total households is scattered between 0.3% and 6.7% of the statistics values in large cities on Oct. 1993 HOUSING SURVEY. The coefficient of correlation between the differences of CENSUS, HOUSING SURVEY and the unknowned tenure is 0.68. So both values are deeply related.