bw_decomp: between-within decomposition of lifespan inequality measures

In alysonvanraalte/LifeIneq: Calculate indices of lifespan variation

between-within decomposition of lifespan inequality measures

Description

Partition a lifespan inequality index into additive components of between-group inequality and within-group inequality. Presently implemented for Theil's index, e-dagger, variance, mean log deviation, the gini coeficient, mean absolute deviatation, the absolute inter-individual difference, and life table entropy (H).

Usage

bw_decomp(
  age,
  ax,
  dx,
  lx,
  ex,
  distribution_type = "aad",
  prop = rep(1/ncol(dx), ncol(dx)),
  method = c("theil", "edag", "var", "mld", "gini", "mad", "aid", "H")
)

Arguments

age	numeric vector of lower age bounds.
ax	numeric matrix of the average time spent in the age interval of those dying within the interval with age in rows and subgroups in columns.
dx	numeric matrix of the lifetable death distribution with age in rows and subgroups in columns.
lx	numeric matrix of the lifetable survivorship with age in rows and subgroups in columns.
ex	numeric matrix of remaining life expectancy with age in rows and subgroups in columns.
distribution_type	"aad" for age-at death, "rl" for remaining life
prop	numeric vector of starting fractions for each of the subgroups.
method	character one of "theil", "edag","var","mld","gini","mad","aid","H"

Examples

 data(LT)
 data(LTm)
 
 # rescale dx to exactly sum to radix,
 # this is a nerd step
 LT$dx <- LT$dx / sum(LT$dx)
 LTm$dx <- LTm$dx / sum(LTm$dx)
 
 # make argument matrices (in this case just two columns)
 method <- "mld"
 ind <- 26:11
 age <- 0:110
 ax  <- cbind(LT$ax,LTm$ax)
 dx  <- cbind(LT$dx,LTm$dx)
 lx  <- cbind(LT$lx,LTm$lx)
 ex  <- cbind(LT$ex,LTm$ex)
 
 


 # Choice of remaining life vs age-at-death
 # makes no difference for age 0:
 
 method <- "mld"
 # same story with other method choice
 # (try "mld", theil", "edag","var","mld","gini","mad","aid", or "H")
 bw_decomp(age = age,
           ax = ax,
           dx = dx,
           lx = lx,
           ex = ex,
           prop = c(.4886,1-.4886),
           distribution_type = "rl",
           method = method)$fB
 bw_decomp(age = age,
           ax = ax,
           dx = dx,
           lx = lx,
           ex = ex,
           prop = c(.4886,1-.4886),
           distribution_type = "aad",
           method = method)$fB
 
# but if we age-condition, results should differ:
 method <- "theil"
 ind <- 26:111
 bw_decomp(age = age[ind],
           ax = ax[ind, ],
           dx = dx[ind, ],
           lx = lx[ind, ],
           ex = ex[ind, ],
           prop = c(.47,.53),
           distribution_type = "rl",
           method = method)$fB
 bw_decomp(age = age[ind],
           ax = ax[ind, ],
           dx = dx[ind, ],
           lx = lx[ind, ],
           ex = ex[ind, ],
           prop = c(.47,.53),
           distribution_type = "aad",
           method = method)$fB

alysonvanraalte/LifeIneq documentation built on March 12, 2024, 1:42 p.m.