Inconsistency in timing of `dead` counts in time-variant model

[alburezg]

Inconsistency in timing of dead counts in time-variant model (kin_time_variant_2sex)

There appears to be a mismatch between the timing of mortality rates and the reported dead counts in the time-variant implementation of kin_time_variant_2sex.

Description

Within the main loop, results are stored as:

for (t in 1:n_years_data) {
  kin_all[[t + 1]] <- ...
}

This produces a list kin_all of length n_years_data + 1. However, names are assigned as:

names(kin_all) <- as.character(years_data)

Since years_data has length n_years_data, the final element (kin_all[[n_years_data + 1]]) is left unnamed (NA) and is not selected downstream (e.g. in output_period_cohort_combination).

Consequence

The practical effect is that:

• dead reported for year Y reflects mortality rates from year Y−1,
• while deaths computed using the rates of the final input year are effectively dropped.

Wanted behaviour

kin_full$dead at time Y should reflect mortality rates at time Y.