# T8 — Political-science faculty perspective

**AI content review, not participation by a real professor.** Lens: comparative politics and research methods. I first read the student-facing text and How to Play, then consulted the permitted facilitator instructions, assessment guidance, source register and clarifications. Locations below use `Player_Screens.txt` line numbers. This was a cold content reading without a browser; I make no claims about interaction, adaptive behavior, accessibility or actual student performance.

## Overall judgment

This can support a serious introductory exercise in explaining political opportunities and limiting inferences. It is less capable, with its present evidence, of adjudicating why national electoral outcomes differed. That distinction should define its instructional promise. The best feature is its permission to conclude something substantive but bounded: an actor lost ballot access, opponents still organized under unequal conditions, or a fresh contest was implemented, while the effect on national voting remains unresolved.

A weaker student could assemble accurate facts, repeat “does not measure votes,” supply a different factor as a supposed rival, and complete a superficially plausible account. A stronger student can distinguish legal opportunity from electoral success, identify a missing baseline, separate observation from interpretation, and specify evidence with different expected patterns. The latter operations constitute analysis. The former mainly demonstrates careful summary. The materials acknowledge human judgment; the priorities below would make that distinction easier to teach and assess.

## Priorities

### 1. Fix the explanatory target before comparing cases — High

**Location/text:** Russia asks “How much could political opponents realistically challenge” (line 88); Malaysia asks what changed enough to defeat an incumbent (994); Malawi separates reopening a contest from its winner (1925); Serbia asks how alternatives remained active under advantage (2866). Compare asks students to “First name the result and period” (579).

These are different dependent variables: opportunity, turnover, institutional reopening and coexistence of activity with inequality. The comparison warning is sound, but naming an outcome alone does not make the outcomes comparable. A student can slide from evidence of exclusion to an explanation of Putin’s retention, then contrast that with Malawi’s winner, without testing one proposition.

**Minimal correction:** Add one short target statement before the connection: “I am explaining this observed option or event, for this actor and period.” Carry it into Compare. Ask whether the second case examines the same kind of option, a different stage, or a scope limit. Retain the prohibition on ranking surveys, votes and seats.

### 2. Give every offered opening a viable evidentiary and writing route — High

**Location/text:** Russia offers “Reported public backing” (93), supported by RU4/RU6. Its actor/action starters instead list commissions, authorities and organizers, followed by registration refusal, prosecution, protest and non-invitation (329–359). Serbia offers a seat-allocation account (2873), while RS6 explicitly says displayed totals do not reconstruct each rule’s contribution (3064).

The openings are reasonably plural, but the scaffolding makes restriction accounts easier to express than support accounts. Conversely, numerical conversion routes invite more precision than the cards supply. An offered hypothesis should not imply that the packet can establish its electoral effect. This is a question of evidentiary affordance, not equal political weight for every interpretation.

**Minimal correction:** Add optional Russia starters for respondents/voters, reported support, and the incumbent’s reported base of backing. Label openings as either directly documented pathways or hypotheses whose effects need further evidence. For Serbia, explicitly permit explaining the direction of conversion without claiming to reproduce 129 seats mathematically.

### 3. Make institutional differences visible without assigning one regime label — Moderate

**Location/text:** “One question. Four ways to test it” (12), the four opening briefs, and “Country files are historical windows, not current country assessments” in How to Play. The source register appropriately separates the 2020 political reading from later case windows.

Students receive useful office and electoral-rule context. They receive less help comparing which institutions actually constrained actors. Russia’s blocked entry, Malaysia’s district contests and later Senate veto, Malawi’s implemented judicial remedy, and Serbia’s pending/rejected challenges are analytically different configurations. “Advantaged incumbent” should not flatten them into equivalent contests. Nor should turnover serve as a retrospective democracy classification.

**Minimal correction:** Add a compact comparison prompt using existing evidence: “Which route was available here: entry, organizing, information, or remedy? Which institution controlled it, and what check actually operated?” Allow “not established.” This creates dated institutional context without inventing judicial independence, assuming selective registration, or classifying all four cases alike.

### 4. Operationalize source independence at the observation level — Moderate

**Location/text:** “Shared documents are not independent confirmation” (378); MY3/MY5 share MY-B; RS0 reproduces REC figures whereas RS2/RS3 report ODIHR observations; MW3 identifies MEC as an appellant and discloses indirect access to judgments.

These distinctions are unusually good. However, publisher counting still offers an easy shortcut. Two passages can add separate premises without corroborating each other; different publications can repeat one underlying claim. DOJ allegations in MY4 do not independently corroborate voters’ responses to 1MDB. ODIHR’s reproduction of results is different from an independently observed campaign finding.

**Minimal correction:** Add one board question: “Do these cards supply different premises, or independently check the same claim?” Require only a phrase, not an extra essay. Where relevant, identify the underlying origin—official count, monitored broadcasts, survey wave, reported detention account. Do not require two different publishers: that would discard legitimate within-document reasoning.

### 5. Attach the rival to an uncertain link, not a settled procedural consequence — High

**Location/text:** The rival field asks “What else could explain the same event or link?” (492); the evidence request supplies contrasting expectations (494). Connection step three correctly allows an inferred baseline (348).

This is stronger than merely asking for “another factor.” Yet a student choosing a tightly bounded pathway may encounter a false obligation to dispute an established relationship. If registration was refused, ballot entry was closed; an alternative explanation of why the commission acted is a different question from whether refusal closed that route. Likewise, court ordering and election implementation are successive steps, not necessarily rival explanations. The text already warns against treating later steps as rivals, but needs a workable next move.

**Minimal correction:** Precede the rival with “Which uncertain link are these accounts explaining?” Permit accounts of the action’s motivation or broader political effect, clearly identified as such. Reward a researchable contrast and acknowledge that the supplied file may not distinguish it. Do not make students manufacture doubt about directly recorded facts.

### 6. Promote Malawi’s threshold qualification into the opening choice — Moderate

**Location/text:** The third opening says “Changed turnout, candidate choices and the majority rule may together help explain the new result” (1932). MW4’s source note more precisely says the 2020 winner had both a majority and the largest share, so the threshold alone would not select a different winner (2117).

The qualification is essential. Otherwise a student may infer that changing the winning threshold mechanically explains the different 2020 winner. A possible strategic effect on alliance formation is a different causal proposition and is not established by the count. This problem is localized; the broader legal sequence is carefully separated.

**Minimal correction:** Revise the opening to distinguish changed candidate choices/participation from possible strategic effects of the interpreted majority requirement. Keep the explicit statement that the threshold alone does not change the winner on the reported 2020 count. Preserve the distinction between interpreting a constitution and amending its text in the facilitator guidance.

### 7. Teach what a source warrants alongside what it cannot establish — Moderate

**Location/text:** Repeated “What it cannot establish alone”; RU3 records the missing observation mission; MW5 records implementation; the final paragraph requires a claim-specific limit (833).

The cautions prevent overclaiming, but repeated negative formulations can encourage indiscriminate skepticism or a ritual final disclaimer. Evidence can strongly establish a local institutional fact while remaining unable to estimate its nationwide electoral effect. Reported support also remains relevant evidence; restrictions do not automatically make every survey answer false.

**Minimal correction:** Pair selected limits with one positive warrant: “This establishes ___; the further inference requiring evidence is ___.” Ask students to explain how their limit narrows their own claim. The existing requirement already says this in principle; one concrete model would make its substance clearer.

### 8. Separate successful rehearsal from independent analytical transfer — Moderate

**Location/text:** Hearing alternatives repeatedly contrast bounded readings with conspicuous overclaims; the optional Vale test (904–907) repeats the monitored-coverage versus voter-effect distinction. The guide’s assessment appropriately reserves quality for human review.

These items can teach measurement discipline. They cannot show that students independently select an explanatory target, build an institutional mechanism, or design a discriminating comparison. A polished caveat can be reproduced without those skills. This is a content limitation, not a claim that a particular student will exploit the checks.

**Minimal correction:** Add a brief unprompted debrief: identify the strongest supported inference, its uncertain link, and one observation that would alter the explanation. Use a different evidence type for later transfer. Keep participation judgments separate from analytical mastery.

## Retain and verify

Retain all-source access, a genuinely meaningful complication, revisable initial ideas, claim-specific limitations, open political conclusions, and the unusually careful distinction between reported results and free preferences. Retain the Malawi discrepancy disclosure without treating it as evidence of fraud or explaining it away. Preserve the separation of turnover, equal opportunity and later reform.

I did not independently verify original documents, translations, numerical calculations, publication histories or the legal summaries. The unavailable original Malawi judgments and unresolved official-total discrepancy remain disclosed limitations, not defects this review has resolved. I have not verified learning gains, completion times or how any controls behave.
