This project’s core idea is a simple two-stage argument about identifying and transporting social status variation.
- Conjoint identification of a status-assigning function.
- A randomized conjoint experiment assigns independent attribute bundles and elicits bounded status judgments (0–100).
- The experiment identifies the average status-assigning rule in the design via randomized assignment and standard identification logic.
- An external-validity assumption (ν-a.e. equality of score functions) is required later to transport the experimental score to target-population moments; it is not part of within-design identification.
- The model is well specified: all relevant attributes are included and there are no meaningful interaction effects, so a linear-in-terms model recovers the true rule.
- Transport to the population distribution.
- A separate evaluation sample is reweighted to match the target population distribution of attributes.
- Weighted sample moments converge to the true population moments.
- Applying the estimated status-assigning rule to this weighted population sample yields a predicted status distribution whose standard deviation converges to the population SD.
Under these assumptions, the estimated model converges to the true status assignment function, and the predicted population status dispersion converges to the true dispersion including for subcomponents that can be decomposed from the additive status scoring function.