Veil-of-darkness test — vodtst • crimlab

Compares each group's share of stops made in daylight, when a driver's race can be seen, with its share in darkness, when it cannot. The ratio of the two odds against a reference group is the factor by which visibility changes that group's chance of being stopped; 1 means no change.

Usage

vodtst(day, night = NULL, ref = 1, shift = NULL)

Arguments

day: Stops by group in daylight: non-negative numbers, as a numeric vector, a 1-d table, a list of single numbers or a one-row data frame. With night left out, one table of groups by stops: a matrix, 2-d table or data frame whose two columns are named day and night, in either order (columns are never taken by position).
night: Stops by group in darkness: positive numbers, in the forms of day; named inputs are matched to day by name, and an error lists any group named in one and not the other. Unnamed inputs are matched by position: only names protect against a mismatch.
ref: The reference group, by name (matched exactly, case included; a factor by its label) or by position in day as given.
shift: NULL, or one positive number per group: daylight exposure over darkness exposure, in the forms of day, matched by name.

Value

A data frame with one row per group, in the order of day: group, ref (TRUE for the reference group), shday, shnight, ratio, sign (the sign of shday - shnight) and, when shift is given, adj.

Details

The ratio identifies the visibility effect only when the mix of drivers on the road is the same in daylight and darkness. shift states how much each group's exposure differs between the two, and adj is the ratio with that difference taken out. See proof("vodtst").

References

Grogger, J. and Ridgeway, G. (2006). Testing for racial profiling in traffic stops from behind a veil of darkness. Journal of the American Statistical Association, 101(475), 878-887. doi:10.1198/016214506000000168

Examples

vodtst(day = c(white = 600, black = 400),
       night = c(white = 700, black = 300))
#>   group   ref shday shnight    ratio sign
#> 1 white  TRUE   0.6     0.7 1.000000   -1
#> 2 black FALSE   0.4     0.3 1.555556    1
# one table of groups by day and night
stops <- data.frame(day = c(600, 400), night = c(700, 300),
                    row.names = c("white", "black"))
vodtst(stops)
#>   group   ref shday shnight    ratio sign
#> 1 white  TRUE   0.6     0.7 1.000000   -1
#> 2 black FALSE   0.4     0.3 1.555556    1