ABSTRACT
‘Demographic array’ is our term for cross-tabulated counts, rates, or other values. The dimensions of a demographic array include attributes such as age and sex, plus, in most applications, time. A demographic array contains one cell for every possible combination of the dimensions. In this chapter, we use the example of 12 fictitious individuals to illustrate how individual-level data are turned into demographic arrays, and we also look at how these arrays are combined or manipulated.
We construct arrays of population counts and death counts. We look at ways of representing changing statuses, including the origin-destination format, pool format and net format for movements. We construct an array of totals for non-demographic events, and arrays of exposures.
We discuss how placing events into ‘Lexis triangles’ allows us to switch between age-based arrays and cohort-based arrays. We look at four common types of measures that can be calculated from arrays of counts and totals: rates, proportions, means, and ratios. We distinguish between super population quantities for underlying propensities, and finite-population quantities for observed values. We show how to collapse one or more dimensions of an array.
