Create new references under key `gsed2510`

[stefvanbuuren]

Background

dscore 2.0.0 sets key gsed2510 as the new default. This key is based on data from seven countries and replaces earlier versions used for calculating D-scores and DAZ scores.

As a temporary solution, the script save_builtin_references.R retains copies of earlier reference tables linked to previous keys.

To ensure consistency and backward compatibility with earlier versions, we need to create three updated reference sets under the new key gsed2510.

Proposed reference sets

• preliminary_standards
  Normative references calculated from a subgroup of healthy children across the seven GSED cohorts. Intended as default.
  Temporary placeholder to be replaced by global who_standards for healthy child development.

• who_descriptive
  Descriptive references calculated from all children with adequate data in the seven GSED cohorts. Replaces phase1 references.

• dutch
  Descriptive references calculated from Dutch GSED children only (for in-country use). Updates dutch references for keys dutch and gsed2406.

Next steps

• Create and verify preliminary_standards under key gsed2510

  • Describe selection criteria
  • Generate NO/LMS/BCT solution
  • Add smoothed centile curves (P3-P97)
  • Update get_mu()
  • Compare with existing keys (gsed2406) to confirm expected shifts
  • Validate DAZ scores
  • Add dataset metadata (e.g., source, cohort composition, date, and method
  • Update DESCRIPTION or internal metadata registry.
  • Include cohort list, variable definitions, and processing date.
  • Add summary of changes in NEWS.md.
  • Store results as builtin_references objects

• Create and verify who_descriptive under key gsed2510

  • Describe selection criteria
  • Generate NO/LMS/BCT solution
  • Add smoothed centile curves (P3-P97)
  • Update get_mu()
  • Compare with existing keys (gsed2406) to confirm expected shifts
  • Validate DAZ scores
  • Add dataset metadata (e.g., source, cohort composition, date, and method
  • Update DESCRIPTION or internal metadata registry.
  • Include cohort list, variable definitions, and processing date.
  • Add summary of changes in NEWS.md.
  • Store results as builtin_references objects

• Create and verify dutch under key gsed2510

  • Describe selection criteria
  • Generate NO/LMS/BCT solution
  • Add smoothed centile curves (P3-P97)
  • Update get_mu()
  • Compare with existing keys (gsed2406) to confirm expected shifts
  • Validate DAZ scores
  • Add dataset metadata (e.g., source, cohort composition, date, and method
  • Update DESCRIPTION or internal metadata registry.
  • Include cohort list, variable definitions, and processing date.
  • Add summary of changes in NEWS.md.
  • Store results as builtin_references objects

• Document the transition and rationale in the package NEWS and vignette

  • Update user vignette to explain new key and references.
  • Provide migration advice for users upgrading from dscore < 2.0.0.

• Release plan

  • Increment patch versions (2.0.1–2.0.x) for each intermediate update.
  • Use 2.1.0 once all references are integrated and validated.

[stefvanbuuren]

dscore 2.0.3 implements the who_descriptive references based on key gsed2510 and seven countries.

Use dscore(..., population = "who_descriptive") to use these instead of the default preliminary_standards (which are slightly outdated since these are based on key gsed2406 and a subset of children raised in favorable circumstances from three countries).

[stefvanbuuren]

dscore 2.1.0 adds references for populations GSED-BGD, GSED-BRA, GSED-CHN, GSED-CIV, GSED-NLD and GSED-TZA.