Reproduction code and synthetic results for Reward History and Voluntary Re-engagement

This archive accompanies “Reward History and Voluntary Re-engagement: A Dynamical Model of Contact, Pressure, and Persistence.” It contains executable simulation and plotting code, fixed parameter and analysis specifications, the disclosed supplemental initial-state diagnostic, synthetic observations and numerical result files, table-generation code, exact-arithmetic checks, original figure assets, and a SHA-256 manifest.

All observations are theoretical model outputs or synthetic samples. No human participant data were collected or fitted. The archive supports a conditional account of post-withdrawal persistence and voluntary re-engagement, comparisons over a fixed 1,116-point pressure grid, mechanism ablations, equal-budget schedule comparisons, observation-window sensitivity, and conditional recovery of two pressure parameters. It does not establish that voluntary return measures attention, that tapering is universally preferable, or that the full model is identified from behavior.

The numerical payload and scientific simulation code are those of the completed fixed analysis; this release packages them for the CSR manuscript preparation version 2.2. The README supplies execution instructions and required libraries. This is a substantial update from the earlier-model reproduction code associated with the prior manuscript “The Dynamics of Entry into Attention.” It is a research-software release, not an accepted journal publication.

AI assistance is disclosed in the README and manuscript: ChatGPT (OpenAI) supported conceptual, mathematical and code development and verification; Claude (Anthropic) supported production of the earlier manuscript and associated computations. The author takes responsibility for the research and its interpretation.