The three latent components behind each pair of counts in bicount_sim: the count recorded by both sources and the counts recorded by each source alone. Kept apart from bicount_sim because they would not be observed in practice; they give the true value of every count, including the source-1 counts that were not recorded, against which estimates and predictions can be checked.
Format
A data frame with 540 rows, aligned row for row with bicount_sim, and 5 columns:
- site, unit
As in bicount_sim.
- Ns
Integer, the count recorded by both sources.
- N1
Integer, the count recorded by source 1 only.
- N2
Integer, the count recorded by source 2 only.
Details
y1 = Ns + N1 on every unit, recorded or not, and y2 = Ns + N2.
Generating values and seed are given in bicount_sim.
Examples
# The true source-1 total over the units on which source 1 was not recorded
un <- bicount_sim$y1_obs == 0
tapply(bicount_sim_latent$Ns[un] + bicount_sim_latent$N1[un],
bicount_sim$site[un], sum)
#> A B C
#> 661 805 718