Worker-firm analysis datasets

These datasets organize HCT scores at worker–firm-period or worker–firm-pair grain for the main individual-level analyses.

Blue nodes are upstream inputs and purple nodes are the datasets read by analysis code. This subsection has no separately persisted support or reusable analysis-layer dataset.

flowchart TB
    scores[(Worker-firm-period scores)]
    periods[(Model relationship periods)]
    firms[(Firms)]

    disaggregated[(Disaggregated regression data)]
    pairs[(Worker-firm pair data)]

    scores --> disaggregated
    periods --> disaggregated
    firms --> disaggregated
    scores --> pairs

    classDef input fill:#e8eef7,stroke:#52739e,stroke-width:2px
    classDef analysis fill:#eee5f5,stroke:#76528c,stroke-width:2px
    class scores,periods,firms input
    class disaggregated,pairs analysis

Datasets and roles

Dataset	Grain	Role
Disaggregated regression data	Worker–firm-period	Direct analysis input
Worker-firm pair data	Worker–firm pair	Direct analysis input

Some analyses read processed data or model outputs directly and therefore do not appear as a constructed dataset in this diagram. The disaggregated data duplicate selected score and relationship-period fields intentionally to provide one aligned source for large regressions.