... used.¹

Statistics Canada is rarely explicit about which rounding technique they use, but Boudreau implies that unbiased random rounding is used [7].

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... conversion.²

The only data on these dwellings are province-wide, in [41] and [48].

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... children.³

Of course, according to the census definition of family, a “mother” could in fact be a stepmother, and there is a small but non-zero probability that she could be younger than her “children.” This is not evident anywhere in the Canada-wide PUMS, but there are two other baffling families: one with a 27 year old father, a 24 year old mother, and a child of 25 years or older; the other has a 17 year-old single mother and a child of 18 years or older.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... simultaneously.⁴

It is conceivable that an IPF procedure could be devised where the two populations are fitted in parallel and could be constrained against each other; however, the convergence and discrimination information-minimizing properties of such a process are unknown.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... households⁵

From the HH86A01 table, there are 849,950 one-family households and 27,720 multifamily households. Assuming 1,000 of these are three-family households, this gives 906,390 census families in total, quite close to the 906,385 total family count found in various family tables.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Hhnumcf.⁶

The details are a little complicated. After synthesizing a dwelling, a special conditional probability table is used to add a CFSIZE attribute using a Monte Carlo draw. The conditional probability is $P(\textsc{Cfsize}\,\vert\, \textsc{Room}, \textsc{Hhsize}, \textsc{Hhnumcf})$, and is calculated by reweighting the Person PUMS for family persons to the family universe. Finally, the dwelling with this additional attribute is used to synthesize the family, conditioning on the shared attributes ROOM, CFSIZE, TENURH, HHNUMCF and CTCODE.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... I8).⁷

Experiments I9 and I10 are excluded from this discussion, since they use the PUMS as a margin instead of using it as the starting table.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.