Samuel Clark Department of Sociology, University of Washington Institute of Behavioral Science, University of Colorado at Boulder Agincourt Health and Population Unit, University of the Witwatersrand Age-Standardization & Decomposition
1 Period Age-Specific Death Rate Death Rate for ages x to x+n during the period spanning 0 to T: M is the death rate D is the number of deaths N is the population
2 Lexis Diagram
3 Components of the Crude Death Rate Dropping the period notation: n C x is the proportion of the population between ages x and x+n
4 Standardization CDR is a function of the mortality schedule AND the age distribution Changes in either or both affect the level of the CDR When comparing CDRs, it is important to isolate the source of the differences: – Differences in age-specific mortality rates? – Differences in age distributions? Age standardization holds the age structure constant so that the only source of difference is the mortality schedule Same applies to any division of the population that produces differing rates (or proportions)
5 Age-Standardized CDR = ASCDR Replacing the n,x notation with i: The Age-Standardized Crude Death Rate is: Where C s is a standard age distribution
6 Selection of a Standard There is no “correct” way to choose a standard As the covariance between the standard and the schedule increases, so will the value of the standardized rate The average of the proportionate distributions being compared is a good choice in general: Where there are N distributions indexed over x
7 Age Standardization: CDR We want to compare the crude death rate from two populations P1 has lower child and higher old-age mortality P2 has higher child and lower old-age mortality P1’s age distribution is almost constant, comparatively unloaded on young ages and loaded on old ages P2’s age distribution is loaded on younger ages and unloaded on older ages
8 Example Mortality Schedules
9 Example Age Distributions
10 Calculation of Standardized CDRs
11 Comparison of CDRs Crude Death Rate
12 Standardization: Income We want to compare male and female average income distributions for the working population The proportionate measure is the job category-specific average income, AI j, for the period 0 to T:
13 Job Category Standardized Average Income As with the CDR, AI j can be written as the product of two components: the job category-specific average income and the proportion of the population holding jobs of each category:
14 An Employment Distribution Effect - Chart
15 An Employment Distribution Effect
16 An Average Income Distribution Effect - Chart
17 An Average Income Distribution Effect
18 A Joint Effect - Chart
19 A Joint Effect
20 Decomposition Decomposition refers to a technique that identifies the proportion of the difference between two crude death rates that results from the differences in the mortality schedules and the differences in the age distributions As with the standardization technique described earlier, this is a general technique that can be used with any crude proportion formed as the sum of proportionate distribution and a proportional measure
21 Components of Difference in Crude Rates
22 Derivation of Decomposition
23 Composition & Rate Contributions to Difference
24 Decomposition Example: CDR
25 Check Decomposition