Anchor extrapolated ever-smoking targets on a youth survey series

Replaces the ever_smoke() targets for cohorts beyond the trend model's own data with values anchored on an external youth smoking series, linked onto the target scale by a factor estimated on the cohorts both sources observe. Cohorts within the trend model's data are returned unchanged.

Usage

anchor_recent_cohorts(
  ever_smoke_data,
  youth_anchor_data,
  ref_age,
  anchor_age_centre,
  taper_cohorts,
  quiet = FALSE
)

Arguments

ever_smoke_data

List returned by ever_smoke().

youth_anchor_data

Data table with columns survey_year and p_ever_smoked (a fraction), and optionally sex. If sex is present its values must exactly match the sexes in the trend model output; if absent, a single link and ratio path is estimated and applied to all sexes.

ref_age

Integer - the calibration age used by init_adj.

anchor_age_centre

Integer - the age the youth series represents: the single age surveyed, or a representative age for a band (13 for an 11-15 band).

taper_cohorts

Integer - number of cohorts over which the handover from trend to anchor is blended. 0 disables the blend.

quiet

Logical - suppress messages.

Details

A youth survey in year t reporting the proportion who have ever smoked is read as an observation on the cohort born in t - anchor_age_centre. The series is linearly interpolated onto integer cohorts, so gap years in a biennial survey are covered. The link factor is the mean of target / youth value over the overlap cohorts (those at or before the last cohort the trend model observed at ref_age), estimated separately by sex when the series has a sex column, and absorbs both later initiation between the survey age and ref_age and differential recall between the two instruments. Anchored targets are applied as a cohort-level ratio to the fitted trend, so gradients across strata within a cohort pass through. The ratio is 1 up to the last trend-supported cohort, blends linearly over taper_cohorts, follows the linked youth series, and is held at its final value for cohorts born after the last youth survey.

Assumed: the link factor is stable across cohorts, and (for a series covering an age band rather than a single age) that anchor_age_centre is a fixed representative age for the band. In the bootstrap the youth series is fixed external data; the link factor is re-estimated each iteration because the trend model is refit.

The link factor(s) and ratio path are attached to the returned list as $anchor.