Severity-weighted crime index — crsev • crimlab

Police-reported crime weighted by seriousness, as the Crime Severity Index of Statistics Canada is built (Wallace et al. 2009): each offence's count is multiplied by its weight (there, the incarceration rate of the offence times its mean prison sentence in days), the products are summed for each unit (a city, a neighbourhood) and period, divided by the population and expressed against a base cell as 100.

Usage

crsev(
  x,
  off,
  w,
  unit = NULL,
  time = NULL,
  pop = NULL,
  per = 1e+05,
  base = NULL,
  sent = NULL
)

Arguments

x: The counts: numbers from 0, one per row.
off: The offence of each row: labels.
w: The weight of each offence: numbers from 0, named by offence (names of offences that do not occur are ignored) or in the order of the offences (a factor's levels, or the sorted labels).
unit: NULL or the unit of each row: labels.
time: NULL or the period of each row: labels.
pop: NULL or the population of each row's unit and period, the same on all its rows: positive numbers.
per: The rates are per per people: one positive number.
base: NULL, or the row of the result whose index is 100.
sent: NULL, or the mean sentence of each offence in the same form as w; the weights are then w * sent (an incarceration rate times a mean sentence length).

Value

An object of class crsev, a list: tab, a data frame with one row per unit and period (in the order of the units, then the periods): unit and time when given, count (the counts summed), sev (the weighted sum), with pop also pop, rate and srate (the count and the weighted sum per per people), and with base also idx (srate, or sev without pop, as 100 times its ratio to the base row's); w, the weights used (off, w); per and base.

Details

Give one row per unit, period and offence (rows that repeat a unit, period and offence add up); an offence with no row in a cell counts 0 there. Beside the weighted rate the plain count rate shows whether the mix of offences moved, not only their number.

References

Wallace, M., Turner, J., Matarazzo, A. and Babyak, C. (2009). Measuring Crime in Canada: Introducing the Crime Severity Index and Improvements to the Uniform Crime Reporting Survey. Statistics Canada, catalogue 85-004-X. https://www150.statcan.gc.ca/n1/en/catalogue/85-004-X2009001

Examples

d <- data.frame(city = rep(c("A", "B"), each = 2),
                off = rep(c("theft", "robbery"), 2),
                n = c(10, 2, 8, 3), pop = 1000)
crsev(d$n, d$off, c(theft = 1, robbery = 50), unit = d$city,
      pop = d$pop, base = 1)
#> crsev: severity-weighted counts and rates per 1e+05, index 100 at row 1
#>  unit count sev  pop rate srate      idx
#>     A    12 110 1000 1200 11000 100.0000
#>     B    11 158 1000 1100 15800 143.6364