Census Survival Ratio Consistency Check

Home > Research Notes > 1997-census-survival-ratio-check

Notes

The Demeny-Shorter procedure takes as input the age distributions for two censuses five years apart and a set of life table survival ratios for the intercensal period and estimates the true age distributions—i.e., corrects for age misreporting errors—on the assumption that the schedule of age-specific correction factors is the same for the two censuses.

The Demeny-Shorter formulation applies only to censues five years apart. While it may be adapted to censuses ten years apart by “interleaving” corhorts, this is clumsy (and going to ten year age groups looses potentially important detail. Sometime in the early 1990s, I believe it was, I invented the notion of “synthetic” survival ratios and showed how to use them to generalize the Demeny-Shorter calculations to censuses any number of years apart.

Files

• Census Survival Ratio Consistency Check [Word | PDF]
• Table 1 Spreadsheet: Application to Japanese Females, 1970-1975 [Excel]
• Table 2 Spreadsheet: Application to Thailand, Both Sexes, 1980-1990 [Excel]
• Directory Listing For the above and other files.

Related Work

Synthetic survival ratios are discussed in the context of census survival estimation in the Methods for Estimating Adult Mortality that I prepared for the United Nations Popuation Division (ESA/P/WP.175, 1 July 2002) (Publications). See Section D of Chapter I. Using synthetic survival ratios for census survival estimation turns out to be equivalent to the method described in Preston and Bennett‘s “A census-based method of estimating adult mortality,”, Popualtion Studies 37(1):91-104 (1983).