Sample and model evidence

These analyses describe the estimation sample, interpret MQ in the separation hazard, evaluate out-of-sample prediction, and measure the part of pair-level MQ not explained by worker and firm components.

flowchart TB
    employment[(Processed employment datasets)]
    scores[(Worker-firm-period scores)]
    regression[(Disaggregated regression data)]
    pairs[(Worker-firm pair data)]

    description_code[[Sample-description code]]
    hazard_code[[Hazard-estimation code]]
    performance_code[[Performance-evaluation code]]
    variation_code[[MQ-variation code]]

    hazard_cache[(Hazard fits and clustered covariance)]
    performance_cache[(OOS predictions and metrics)]
    variation_metrics[(Variation metrics)]

    description[/Sample-description LaTeX table/]
    hazard[/Hazard-interpretation LaTeX table/]
    performance[/Predictive-performance LaTeX table/]
    variation[/Match-quality-variation LaTeX table/]

    employment --> description_code --> description
    scores --> hazard_code
    regression --> hazard_code --> hazard_cache --> hazard
    scores --> performance_code
    regression --> performance_code --> performance_cache --> performance
    pairs --> variation_code --> variation_metrics --> variation

    classDef upstream fill:#e8eef7,stroke:#52739e,stroke-width:2px
    classDef input fill:#eee5f5,stroke:#76528c,stroke-width:2px
    classDef code fill:#f3f3f3,stroke:#666,stroke-width:2px
    classDef support fill:#fff0cc,stroke:#a66b00,stroke-width:2px
    classDef tableOutput fill:#e2f2e7,stroke:#3b7a4b,stroke-width:2px
    class employment,scores upstream
    class regression,pairs input
    class description_code,hazard_code,performance_code,variation_code code
    class hazard_cache,performance_cache,variation_metrics support
    class description,hazard,performance,variation tableOutput

Analysis	Main sample	Final fragment
Sample description	Processed employment data	sample_description.tex
Hazard estimates and MQ interpretation	Full-training in-sample	hazard_estimates.tex
Predictive performance	Training model, evaluated out of sample	model_performance.tex
Match-quality variation	Full-training in-sample pairs	pairs_variations.tex