Contradiction results
The contradiction test examined 3,859 candidate pairs of claim groups. Five pairs were retained as direct contradictions after repeated judgments and review, all concerning hybrid-journal payments. The table identifies the groups in the claim-group data. Its position descriptions are extracted paraphrases.
| Conditional payment position | Group | Documents | Opposed position | Group | Documents |
|---|---|---|---|---|---|
| Support hybrid journals only when their open-access share exceeds their subscription share. | osft-doc-1373/c185 |
3 | Do not pay APCs at hybrid journals. | open-science-source-02/c321 |
2 |
| Reduce hybrid-journal APCs to the open-access-only market average. | osft-doc-1373/c122 |
3 | Do not pay APCs at hybrid journals. | open-science-source-02/c321 |
2 |
| Mandate offsetting agreements for hybrid journals while subscriptions remain active. | osft-doc-1373/c125 |
3 | Do not pay APCs at hybrid journals. | open-science-source-02/c321 |
2 |
| Reduce hybrid-journal APCs to the open-access-only market average. | osft-doc-1373/c122 |
3 | Fund only publishing services offering full and immediate open access. | osft-doc-1398/c46 |
2 |
| Mandate offsetting agreements for hybrid journals while subscriptions remain active. | osft-doc-1373/c125 |
3 | Fund only publishing services offering full and immediate open access. | osft-doc-1398/c46 |
2 |
The five rows involve three distinct document records on the conditional-payment side and four on the refusal side. Each individual contrast spans three documents against two.
No pair was retained as a direct contradiction on the other four disagreement themes. The test treated context-dependent reproducibility as a qualification of general calls for reproducibility. A preference for nonprofit publishing and conditional private-sector participation were judged compatible. The candidate search and judgment rules can miss opposing claims; the retained pairs do not enumerate every possible disagreement.