feat(psychometric): recover 2017-era commented asymTRAITVAR as trait/a² on main

[seonghobae]

Independent scalar increment from protected main 6444a81. Does not stack on CWC #327, Kish-weighted CWC #312, Eq. 5 T0TOTALVAR #328, or Driver std restorations #299–#326.

Map

2017-era ctsem 2.3.1 summary.ctsemFit.R comments

asymTRAITVAR <- solve(DRIFT) %*% TRAITVAR %*% t(solve(DRIFT))

and does not put it in the active outlist. Driver, Oud, and Voelkle (2017, Eq. 1, p. 4) write dη = (Aη + ξ + Bz + Mx) dt + G dW. At a stable equilibrium η_∞ = −ξ / a, so process-mean variance from that random intercept is trait / a².

Form 1 / a first, then square, then multiply by trait.

• Zero trait is exactly zero even if a ≥ 0.
• a ≥ 0 with a nonzero trait fails closed (AsymptoticTraitVarianceRequiresStableDrift).
• Non-event clock fails closed (EventTimeRequired).
• Overflowing 1 / a or product fails closed (InvalidNumericInput).
• Numeric: a = −0.5, trait = 1 → 4.

Refusals

• trait / a² is not Table 2 TRAITVAR φ_ξ.
• trait / a² is not §7.2 addedTIPREDVAR (B / a)² v.
• trait / a² is not asymDIFFUSION −q / (2 a).

The commented if (asymptotes == TRUE) TRAITVAR <- DRIFT %&% TRAITVAR is a² · trait and is the opposite transform, not this map. Section 4.3's extra a = 0 trait process is already in process units and is not this random-CINT map.

Local verification

• cargo fmt --all
• cargo clippy -p psychometric_core --all-targets --all-features -- -D warnings
• cargo test -p psychometric_core: lib 190, multilevel 109, scientific-claim 53, rubin/mean-gate 6, plausible-value 4

JSS PDF re-opened 2026-08-30T21:10Z from https://www.jstatsoft.org/index.php/jss/article/download/v077i05/1104.

Meredith (1993) remains unread (Unpaywall 2026-08-30T21:10Z: is_oa: false, 0 locations). Mislevy (1991) remains unread on the same terms. Still not DSEM, not a Kalman filter, not a matrix expm, not ctsem estimation. Do not self-approve. Do not merge without independent non-author APPROVE and exact-head required-check success.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[seonghobae]

Requesting independent Copilot review of exact head 0cbb221. Do not self-approve.

[seonghobae]

Exact head 0cbb221a9c13d3068dfea90ec85bb6b065e40396 from protected main 6444a81.

Local verification:

• cargo fmt --all
• cargo clippy -p psychometric_core --all-targets --all-features -- -D warnings
• cargo test -p psychometric_core: lib 190, multilevel 109, scientific-claim 53, rubin/mean-gate 6, plausible-value 4

github___request_copilot_review MCP returned unauthenticated this cycle. REST requested_reviewers for copilot-pull-request-reviewer is 422 (not a collaborator). Copilot /copilot/reviews is 404. Independent review still needed. Author COMMENT is not APPROVE.

Unpaywall 2026-08-30T21:10Z: Meredith (1993) DOI 10.1007/BF02294825 and Mislevy (1991) DOI 10.1007/bf02294457 remain is_oa: false, 0 locations. Still unread. Not DSEM. Do not self-approve. Do not merge without independent non-author APPROVE and exact-head required-check success.