# S1 independent first-use pressure test

I simulated a willing, low-confidence 18-year-old introductory politics student with limited research vocabulary. I used only the visible game at `http://terminal.local:4173/?reviewer=S1`, its glossary, help and source drawers. Clicks and displayed text are observations; feelings and interpretations are simulated reactions. Tool latency does not measure human completion time.

## First understanding and first actions

Before opening help, I understood the assignment as explaining why one political side had an advantage, using election evidence, then checking another country. I thought “public hearing” meant I would argue my view and answer objections. I did not expect to alter history because the opening explicitly says the outcomes are known. “Two anchors and one complication” sounded like two pieces of support and one problem, although “anchor country” and “anchor” evidence initially seemed like the same thing.

Russia, For class and Solo were already selected. I entered without help, chose “Start with my own account or question,” and recorded: “Putin could keep winning because the people who opposed him were stopped from running or organizing.” I left confidence provisional. I opened Choose evidence, consulted Glossary and How to play, and read RU1, RU2 and RU4 with source details. I selected RU1 and RU2 as anchors and RU4 as the complication. I read the other card headlines, not their full records.

## Confusion and recovery log

- **“Your public exhibit has two anchors and one complication.”** I initially wondered whether the two anchors were countries. The evidence page and glossary definition, “An anchor supports a step in your account,” resolved this. Naming them “supporting evidence 1/2” alongside the metaphor would reduce the initial translation work.
- **“Each row has its own denominator.”** I knew denominator as the bottom of a fraction, but did not immediately know what to do with it here. I opened Glossary specifically for this; denominator was absent. The table's “Who answered” column let me infer that it meant which group the percentage was of. Add that definition and point to one table row.
- **“Why foreground it?”** I read this as “why did I choose it?” and entered short reasons. The glossary's anchor definition supported that interpretation, but did not define foreground. Plain “Why use this evidence?” would retain the task.
- **“Observation and inference.”** I was unsure whether inference meant a guess. The glossary's contrast between what a source records and the explanation I draw from it fixed this; the library-hours microexample in How to play made it concrete.
- **“January hypothetical-ballot comparison.”** I took this to mean people were asked which candidate they might choose, rather than an actual result. The full RU1 card explained the January question more clearly than the source label. That recovery was enough, although I still underestimated what an early 3% could tell me.
- **“State two plausible explanations… Name evidence that could help separate them.”** I expected somewhere to write both explanations and a research idea. The immediate control was a three-choice reading, followed by a decision-note field. I answered the choice, then used the note, but did not actually supply all parts of this instruction. No visible check flagged that gap. Split these into an explicit short response or label them as discussion prompts.
- **“A proposed overclaim need not be yours” / “bounded account.”** I needed a second reading to understand that the game was testing a supplied claim rather than accusing my paragraph of saying it. “This claim may differ from yours” and “claim with stated limits” would help. The sentence saying no software had read my prose clarified the relationship, but interrupted the hearing atmosphere.
- **“Which connection changes?”** After choosing Narrow a step, I had to map “Possible consequence” back to “What might that help explain?” on the previous page. I managed it and changed that step. Reuse the original question labels and show the current sentence beside the selector.
- **“Similar advantage, different trajectory.”** I guessed trajectory meant what happened next or the eventual result. The follow-up explanation about actions and options helped. I chose Malawi because its opposition won, but remained unsure whether I had demonstrated a *similar* advantage. A concrete comparison-question example would help without supplying an answer.
- **“The capsule suggests MW3 and MW4.”** Capsule was unfamiliar as a label, and “MEC” in the preceding paragraph was unexplained there. Reading Source details revealed Malawi Electoral Commission. “Summary” and an expanded first use would be easier.
- **“Use MW3 in comparison.”** I expected a selected-card marker. Observed behavior: it inserted `MW3` into the comparison text box and displayed “MW3 added to your comparison draft.” The toast explained the action after the fact. “Insert MW3 in draft” would match it.
- **“Evolving working view”** in the final preview still contained my original opening, although the final account reflected my revisions. I could see both, but wondered whether an edit had failed. Distinguish the opening-page working hypothesis from the account being revised, or consolidate them.

## What I actually completed

I built a supplied chain: election commission → refused registration → challenger could not appear on ballot → narrower field. Attaching evidence four times felt repetitive. I attached RU2 to the last step, then later narrowed that step to wording about Nadezhdin's possible campaign. The final preview retained `[RU2]`. This was my mismatched evidence attachment, not evidence of a software reasoning bug; the revision flow did not draw my attention back to it.

On the first hearing question I selected the reading distinguishing ordinary respondents' experiences from opposition organizers' opportunities, then defended my account with RU2. On the second, I initially chose “A January 3% establishes that a later campaign could not have gained support.” My simulated reasoning was that 3% seemed too small to matter. Observed feedback explained that a hypothetical choice with limited recognition did not fix the result of a campaign that never happened. I reopened source records, reconsidered, selected the cautious alternative, and narrowed the consequence. The feedback fixed this particular misunderstanding. My attempt history showed two attempts.

My filed 72-word account was:

> Putin's opponents had fewer ways to challenge him. RU1 shows Nadezhdin being stopped from appearing on the ballot. RU2 shows prison sentences for organizers, which could scare people away from joining. This helped Putin face less competition. But RU4 suggests public support also mattered. The survey only records what people said. We cannot know how many votes Nadezhdin could have gained with a full campaign or how much restrictions changed the result.

I read MW3 and MW4, selected MW3 for the draft, and wrote a 39-word comparison about Malawi's courts reopening competition and challengers joining together. I requested surveys before and after the alliance as next evidence. I added a separate limitation phrase, completed the optional Vale source question, and clicked File this hearing record. Observed result: **“Hearing filed · interpretation awaits discussion.”** Preview showed the account, comparison, decisions, attempts and citations. No completion blocker remained. I did not test downloads or conduct an actual spoken class discussion.

## Priorities and verdict

**Blocking: none observed.** I reached a filed record through the intended stages.

**Major: vocabulary and reading load accumulate at the exact point independent reasoning gets harder.** Malawi's introduction packs candidates, courts, rules, alliances, turnout and tiny reporting discrepancies into one paragraph. Source drawers repeat lengthy metadata. As S1, I would skim those parts and lean heavily on the suggested cards. Use a short sequence of events, expand acronyms, and put precision/audit details in a further disclosure while preserving access.

**Major: the hearing's written demands exceed its obvious response structure.** I could complete the inclusion checks without answering the requested two explanations or proposed distinguishing evidence for the second question. Make those requirements visibly actionable and distinguish them from optional discussion.

**Minor: label changes, insertion controls and duplicate limitation/working-view fields add bookkeeping.** Showing the exact causal sentence being revised would also help catch my lingering RU2 attachment mismatch. Asking again for a limitation after it is already in my paragraph feels like form filling, although the final label makes the limitation easy for a reader to find.

The cream background, strong headings and restrained colors felt calm and readable. The persistent exhibit helped me remember my choices. Smaller metadata and long source paragraphs were less inviting. The enjoyable part was seeing a simple “Putin stopped opponents” explanation become less certain without being told to abandon it. The three slots made choosing evidence meaningful. The multiple-choice extremes also made the cautious answer easy to recognize after feedback; recognizing it was easier than producing the requested reasoning myself.

Verdict: finishable and intellectually worthwhile for this student, with vocabulary support and teacher discussion still doing substantial work. A 55-minute lesson seems plausible with selective reading and guidance; this test cannot establish a human completion time. I would simplify the language and response structure before adding more content.
