# Preregistration template — unregistered draft

## Administrative

- Title, investigators, institution, ethics protocol, registration timestamp, frozen commit.
- Recruitment source, eligibility, compensation, consent, privacy, and withdrawal procedure.

## Design

- Questions and hypotheses H1–H8.
- Selected 20-cell matrix and justification against 16/24/180-cell alternatives.
- Unit of assignment, blocks, sessions, concealment, sample size, and fixed stopping rule.

## Outcomes and analysis

- Operational definitions, ITT estimands, models, clustering, repeated rounds.
- Missingness, exclusions, manipulation checks, multiplicity, heterogeneity, robustness.
- Exact-model comparisons labeled as synthetic ground truth, never human effects.

## Reporting

- Deviations, attrition, calibration failures, all primary outcomes, null results,
  code/data availability, and the no-human-data status of the development package.

Completing this file does not constitute preregistration.

## Selective-attention v2 amendment draft

- Questions and hypotheses H9–H14.
- Nine appended treatment cells T20–T28 and the frozen 29-cell assignment matrix.
- Attention, implementation, and conditional-policy multiplicity families.
- Eleven declared synthetic response scenarios, including heterogeneous
  salience, partial compliance, memory leakage, role aversion, and social
  preference heterogeneity.
- Prominent statement: **No participants were recruited. No human data were
  collected. No experiment was conducted.**