plot_simulate_y_structure_optim: Plot the minimum Y-structure required to explain some...

In explodecomputer/collidR: Examining bias in observational associations due to sampling ascertainment

Description

For a given Y-structure, can an association of b_xy can be obtained through ascertainment on a collider? Use optimisation to determine the minimum OR_sz for a range of rsq_xs and rsq_ys values

Usage

plot_simulate_y_structure_optim(
  prop,
  b_xy_thresh,
  b_xy = 0,
  rsq_xs_range = c(0, 1),
  rsq_ys_range = c(0, 1),
  gran = 101,
  max_or_sz = 20
)

Arguments

prop	Proportion of population included in sample
b_xy_thresh	Target b_xy association - e.g. what value of b_xy are you suspicious could be due to ascertainment on a collider
b_xy	Suspected true X-Y effect. Default = 0
rsq_xs_range	Range of rsq_xs values to enumerate over. Default = c(0,1)
rsq_ys_range	Range of rsq_ys values to enumerate over. Default = c(0,1)
gran	Granularity of ranges. Default = 101
max_or_sz	Maximum OR_sz to allow in optimisation.

Value

ggplot of simulations

explodecomputer/collidR documentation built on May 8, 2020, 3:33 a.m.