Sub provincial estimation of undercoverage in the 1991 Canadian census / Peter Dick and Don Royce. : CS95-0169/1991E-PDF

"Following each Canadian Census, an evaluation study known as the Reverse Record Check is carried out to estimate the level of undercoverage. The sample size of this study is such that reliable estimates of undercoverage can be produced for each province, and for certain age-sex groups at the national level, but not for age-sex groups at the province level. The use of synthetic estimation techniques has been investigated for this latter purpose, but these have the disadvantage that they are based on inherently unverifiable assumptions. Cressie (1989) proposed a compromise model that combines the direct survey estimate and a synthetic estimate. We generalize Cressie's model to admit the possibility of bias in the survey estimates, and include some results on possible gains to be made by employing this model. These results are illustrated with data from the 1986 Reverse Record Check."--Abstract.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.839690&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs