Functional-form sensitivity checks
| Specification | Rank correlation | Interpretation |
|---|---|---|
| Multiplicative | Primary model | TA×(1-IR/100); treats institutional resistance as a partial gate on technical suitability |
| Additive mean | Spearman ρ = 0.44 | Equivalent in ranking to TA+(100-IR) after rescaling. It gives less force to high institutional resistance than the multiplicative model |
| Min-rule | Spearman ρ = 0.80 | Uses the lower of TA and realised permission. It compresses high-resistance occupations and supports broad-band rather than fine-rank interpretation |
| Geometric mean | Spearman ρ = 1.00 | Equal-weight geometric mean is rank-equivalent to the multiplicative model because both are monotonic functions of TA×(100-IR) |
| O*NET weighting check | Spearman ρ = 0.998 | Weighted and unweighted aggregation produce nearly identical rankings |
| Specification | Rank correlation | Interpretation |
|---|---|---|
| Multiplicative | Primary model | |
| Additive mean | Spearman | Equivalent in ranking to |
| Min-rule | Spearman | Uses the lower of TA and realised permission. It compresses high-resistance occupations and supports broad-band rather than fine-rank interpretation |
| Geometric mean | Spearman | Equal-weight geometric mean is rank-equivalent to the multiplicative model because both are monotonic functions of |
| O*NET weighting check | Spearman | Weighted and unweighted aggregation produce nearly identical rankings |
Note(s): Additive mean, min-rule, and geometric mean are computed from the same occupation-level TA and IR scores as the primary multiplicative model. Full occupation-level scores are provided in the replication repository
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.