# Eight missing card-face limits

These proposed limits use only `content/case_content.json` and `content/source_register.json`. They are short learner-facing boundaries; retain the full registered source limitations and source metadata in details. No application content was edited.

| Card | Proposed face limit | Source association and grounding |
|---|---|---|
| **RU3** | The absence of this mission does not measure fabricated votes or sincere support. | **RU-C**, OSCE/ODIHR and OSCE PA, 29 January 2024. The statement establishes no invitation and no deployment; its registered limitation explicitly excludes measuring fabricated votes or sincere support. |
| **RU4** | The survey records reported votes and attitudes from different respondent groups; it is not an official count or an estimate of preferences under open competition. | **RU-D**, Levada, fieldwork 21–27 March 2024, published 24 May 2024. The card distinguishes all respondents from self-reported voters and notes home interviews. The register distinguishes recalled votes/attitudes, denominators, official returns, and open-competition preferences. |
| **RU6** | Intended choices were recorded under existing restrictions; this is neither an election forecast nor evidence of preferences under unrestricted competition. | **RU-F**, Levada, fieldwork 21–28 February 2024, published 19 April 2024. The card calls the result sympathies at that moment, not a forecast. The register states that restrictions already shaped choices and the answers do not establish unrestricted competition. |
| **MY2** | Reported restrictions before the election do not measure how many votes they changed or the incumbent’s sincere support. | **MY-C**, HRW World Report 2018, events of 2017, expression/association sections. The card documents investigations, arrests, prosecution and registration powers. The registered limitation excludes vote-effect estimation and measuring sincere incumbent support. |
| **MW1** | The figures measure monitored 2019 airtime, not audience size, beliefs or votes; the preliminary report predates final tabulation and court petitions. | **MW-A**, EU preliminary statement, 23 May 2019; monitoring 17 April–18 May 2019, printed pp. 1–2 and 9. Both card and register distinguish monitored content from audiences/beliefs/votes. The register specifies the report's timing before tabulation and petitions. |
| **MW2** | These accounts do not reconstruct an irregularity-free count; court details here rely on MEC and scholarly analysis, not directly inspected judgments. | **MW-B**, MEC 2019 report, printed pp. 66 and 72; **MW-D**, Nkhata, Mwenifumbo and Majamanda (2021), printed pp. 59–60. MW-B's declared results were later set aside and cannot reconstruct an irregularity-free count. MW-D's register states original judgments were not retrieved and the court details rely on scholarly analysis and MEC. |
| **RS1** | A government’s stated rationale does not verify its democratic claims or establish equal campaign conditions. | **RS-B**, Government of Serbia, 30 October 2023, first three substantive paragraphs. The card and source record the government's public rationale. The register explicitly does not verify its democratic claims or establish equal conditions. |
| **RS4** | This documents one operation, not a criminal conviction or a national estimate of votes changed. | **RS-C**, CINS, 29 November 2023, recruitment, recorded exchanges and cash-payment sections. Both card and register distinguish firsthand investigation of a particular operation from a conviction and a national count of affected votes. |

RU4's final wording above refines the earlier message's “recalled answers” phrasing: attitudes and reported personal pressure are not themselves recalled votes. This wording keeps the separate respondent groups explicit without implying all survey questions measure the same thing.
