Political data literacy / Instructor guide / V2.2

The Complaints Department

Created by Yoav Duman · 10 October 2026

Audit an answer. Give a useful reply. Step into a fictional public office where every reply needs an evidence check. The Complaints Department, created by Yoav Duman, turns earnings, representation, public opinion, and an optional housing source audit into short investigations: pin the values, keep supported findings, question overclaims, and write a useful reply.

Plan the session

Instructor recommendation: use a 50-minute class for the first file, paired discussion, one optional fresh case, and submission. For the full three-file sequence, allow 60-75 minutes including discussion, or split it across two meetings. Add about 10-15 minutes for the optional housing audit. These are planning recommendations; the game’s 35-45-minute core target has not been established by a human student pilot.

Play the website edition · Download PDF guide

Learning goals

1. Identify a display’s measure, population, place, period, numerator or denominator, and units before drawing a conclusion.
2. Separate three judgments: whether the original question has been answered, whether a claim is supported by the available evidence, and what action should happen next.
3. Retain a sound descriptive finding while qualifying an unsupported causal explanation or an unmeasured claim.
4. Propose evidence that addresses a specific unresolved question and explain what that evidence would still leave uncertain.
5. Compare a preserved first reading with a current response, explaining a revision or a justified decision to keep the original reading.
6. Distinguish the fictional office correspondence and invented practice scenarios from the historical Census, IPU, and Pew evidence and the housing source-audit packet.

Before class

1. Course fit: introductory political science, social science methods, quantitative literacy, public policy, gender and politics, democratic representation, or evidence-based writing. Students need subtraction and the idea of a percentage; Help and terms explains medians, percentage points, and evidence limits.
2. Open the current V2.2 game and decide what students should submit: a readable HTML submission report or a PDF made with Print or save PDF. State the assignment location and deadline in your course system.
3. For a first 50-minute class, assign File 01, Earnings, or choose another independent file. Students can choose one file at the desk. The default guided sequence is Earnings → Representation → Congress opinion → optional Housing audit. Class links assign only selected files, in authored order; the first selected file opens directly and unassigned files create no prerequisites.
4. Students enter a name or instructor-approved course identifier at the desk. Housing can be assigned directly as a one-file source audit or added to the default sequence. Completion counts only the assigned files: a Housing-only or Congress-only assignment can finish at 1/1. Housing still preserves the first audit before revealing its matching-source lead.
5. Tell students they can use View data table, View source, Enlarge graph, and the representation scale toggle freely. These access tools do not count as requested hints. Pinning is possible from either the graph or the table.
6. Explain that the app checks factual batches where present, but an instructor assesses written reasoning. A filed record is a completed record, not a guarantee that every judgment or explanation is correct. There is no countdown and no policy preference score.
7. Decide whether students will use optional online saving and optional instructor review. A local checkpoint and a readable report remain useful even when the online service is unavailable.

Run the lesson

Pair with Name the Problem

The Complaints Department asks whether a public reply answers a resident’s question, what the evidence establishes, and what useful response should follow. Name the Problem asks students to compare two views of an issue, choose a political frame, and defend what deserves attention using an explicit value or standard. Use either independently. For a sequence, assign the Earnings file here, then congressional representation in US Government or the democracy case in Comparative Politics during another meeting. Bridge them with two questions: what kind of claim is being made, and what support does the evidence provide? Debrief which judgment carried over and which required a value. Assess accurate comparisons and justified reasoning. Retaining a sound answer, acting, investigating, and monitoring can all be defensible.

Assess the reasoning

Completion records required moves; assess the explanation and evidence yourself.

Discussion prompts

1. Earnings: what does the $22,550 difference between two estimated annual medians establish? Which extra assumptions would turn it into a claim about identical work, every individual, or a particular cause?
2. Earnings: could jobs, hours, promotion, care arrangements, and discrimination be connected pathways rather than competing labels? What question would you need to define before deciding what to hold constant?
3. Representation: why can the original historical enquiry close even though the new India lower-house enquiry remains open? Does a worldwide seat share answer a question about a particular country or all political power?
4. Representation: the share rose from 15.6% to 27.0%. Why is this an 11.4-percentage-point change? In the fresh case, how can the number of women rise while their share falls?
5. Congress: distinguish unfavorable opinion of Congress, support for a specified referendum reform, and the reform’s possible effect on attitudes. Which observation or study would address each?
6. Housing: what can be recorded before the source is found? After the matching chart appears, which verification tasks remain? Why does missing documentation establish neither that the concern is false nor that the estimates are valid?
7. Across files: when is Not established by this evidence the right conclusion? How is it different from Contradicted?
8. Across files: what makes a brief final reply useful to the resident? Can a record be complete while an enquiry, source check, or causal question remains pending?

Save, recover, and submit

1. Device saving belongs to the current browser. Use Save progress file for a portable JSON checkpoint, especially before leaving, switching devices, or starting a different assignment. Check that the download exists.
2. Restore progress accepts a supported Complaints Department checkpoint and previews it before confirmation. Confirming downloads the outgoing assignment before replacement. Cancelling or rejecting an invalid file keeps the active work. Valid Version 1 work can resume with its original first readings and records preserved.
3. Optional online saving: open Online save & recovery, choose a label and whether to Allow instructor review, then Create my online copy. Download the private recovery card. Later edits save to the connected copy; Save online now offers an explicit save. The recovery key opens that copy, so students should retain it privately.
4. On another device, open Online save & recovery, paste the complete recovery key, and Find online copy. Review the preview, download the online checkpoint if needed, then choose Download this draft and load online copy. Use a separate online copy for a different run.
5. Commit first reading preserves the initial response within this saved run. Later controls edit the current response. Factual attempts and requested help are recorded; unchanged checks do not create extra attempts. The optional fresh case’s first response and submitted second response are separately locked.
6. Filing requires the first reading, the required evidence pair, both question statuses, a next step, a request, a final reply, and a revision or retention explanation. Earnings also requires two different possible explanations and an investigation explanation. Housing requires an investigation explanation. Representation requires the separate linked country enquiry. Investigation prose is optional for Representation and Congress; Congress offers a separate optional second evidence request.
7. Earnings, Representation, and Housing use factual batches. The current batch must be checked and supported, or checked with worked assistance acknowledged, before proceeding. Changed answers or pins need a new check. An assisted record may still contain incorrect factual answers; the report labels that distinction.
8. Closing an enquiry while an original question is not fully answered requires a brief closure rationale; students can instead choose a next step that keeps it pending. Filing records that decision without grading the prose. Changing response fields after filing returns the file to draft; review and file it again.
9. The readable report includes historical evidence and limits, first and current readings, pinned values, requests, writing, the current factual draft, recorded attempts, assistance, and optional fresh responses. A progress JSON restores work; it is not the readable submission. Nothing is submitted automatically to Canvas or another course system.
10. Assess the reasoning rather than prose length, political preference, or willingness to change a sound answer. Use the assistance trail to understand the work, not as a correctness score. Preserved readings document this run; exported local files do not prove that a student never edited a checkpoint or played previously.

Instructor tools

1. Open Instructor access in the footer and enter your instructor code to unlock the studio. Use Lock studio when finished. The code is for teaching controls, not a student submission step.
2. Content & focus lets you edit the opening introduction, teaching prompt, and case titles or prompts, and choose assignment focus. Publish teaching settings creates a saved release. Evidence values, answer definitions, and preserved first-response rules retain their authored structure.
3. Class links lets you enter a title and instructions, choose at least one assigned file, and Create class link. Copy the resulting link for students. Each named class link keeps a fixed teaching-settings snapshot; publishing new general settings does not change that existing link. Only selected files count toward completion, and the first selected file opens directly. A different saved run keeps its original scope: Start assigned files downloads its checkpoint before opening a separate run.
4. Releases & rollback lets you download settings and releases or Publish this earlier snapshot as a new release. Restoring a settings file first creates an editable draft. A rollback keeps release history and does not rewrite student responses.
5. Record review lists only online copies students explicitly allow you to review. Refresh shared records, open an individual record, inspect the readable writing, first/current choices, source limits, assistance, and attempt history, then download a readable report or checkpoint or Print record / save PDF. The overview’s recorded count is participation status, not a quality score.
6. A shared online copy is not an automatic course submission. Tell students separately what to upload to the course system. If a record is absent, check that the student created an online copy and chose instructor review; retain the regular report-submission route.
7. The in-game instructor notes provide the 50-minute plan, eight-point rubric, primary-source links, and evidence limits, with Download instructor notes. Development used AI reviewer perspectives; no human student pilot or measured learning improvement is claimed.

Theory and evidence boundaries

1. The fictional office makes the student’s task concrete: answer what can be answered and leave a useful next enquiry. Letters can contain deliberate overclaims. They are teaching material, not statements endorsed by the game or by the historical source.
2. Earnings - Census 2022 ACS one-year estimates: the relevant population is U.S. adults ages 25-64 with earnings and a bachelor’s degree or higher, grouped by first bachelor’s major and recorded sex. Annual earnings refer to the preceding 12 months in 2022 dollars and include wage/salary and net self-employment income. Economics medians are $107,300 and $84,750, a $22,550 difference. This is not an hourly wage comparison or a matched-work experiment; the table reports margins of error. Adjustment alone does not identify discrimination, and adjusting away jobs, hours, or promotion can remove pathways through which unequal opportunities operate. Source: https://www.census.gov/data/tables/2022/demo/educational-attainment/acs-detailed-tables.html
3. Representation - IPU historical October snapshots: 15.6% in 2004 and 27.0% in 2024 yield an 11.4-percentage-point increase. The worldwide aggregate is seat-weighted across chambers, not an average of country shares. Its denominator is filled seats with recorded gender: 6,445 women among 41,381 MPs in 2004; 11,970 among 44,301 in 2024. Total filled seats, including those without recorded gender, were 42,650 and 44,578 respectively. Earlier points use 31 October; later points use 1 October. The trend does not identify causes, establish every country’s trend, or measure all political power. The linked India lower-house question is a separate historical enquiry. Sources: https://archive.ipu.org/wmn-e/arc/world311004.htm and https://data.ipu.org/women-averages/?date_month=10&date_year=2024
4. Congress - Pew July 10-16, 2023 survey: weighted estimates are 26% favorable, 72% unfavorable, and 2% no answer about Congress. The INSTFAV item was asked of 4,233 Form 1 respondents within a survey of 8,480 adults, targeting noninstitutionalized U.S. adults ages 18 and older. The 2% is not a neutral category; percentages are weighted estimates, not raw counts. Do not automatically apply the full-sample margin of error to the item’s subsample. The item does not measure referendum support. A direct question about a specified proposal addresses support; a credible causal design addresses its effects. Students commit an independent first reading before departmental notes and case-specific help, while the table, general terms, and source information remain available. Source: https://www.pewresearch.org/politics/2023/09/19/how-americans-view-congress-the-president-state-and-local-political-leaders/
5. Housing - source discovery and verification are different: the fictional packet initially lacks its citation, denominator, and methods. Below $50,000 it displays 9% and 38%, a 29-percentage-point difference. After the student commits a first reading, the source area reveals an independently located matching chart from the University of Washington’s Racial Restrictive Covenants Project. All ten displayed percentages match the image, but its estimates have not been independently reproduced. The project describes weighted IPUMS USA samples averaged across 2021 and 2022 ACS data, with household race defined by the reference person and specified race/ethnicity exclusions. Extraction details, income-band rules, year combination, geographic crosswalk, and design-based uncertainty still need checking; the page also mixes year labels. Its explanatory claims need their own evidence. The instructor notes and exported record include this lead. Source: https://depts.washington.edu/covenants/homeownership_king.shtml
6. Optional fresh cases use invented worked data, not new Census, IPU, or Pew findings. Earnings tests group medians versus individual and causal claims; Representation tests count versus share when the denominator changes; Congress separates measured dissatisfaction, support for a remedy, and predicted effects. Housing has no additional fresh case.
7. A selected evidence request is a proposed investigation. The game does not pretend that the requested documents or study have arrived. Not established by this evidence leaves a claim open; it does not mean false. A filed audit can honestly record pending verification.

Practical troubleshooting

1. A later assigned file is unavailable: file the preceding assigned records first. Unassigned files never block access. To start any individual file directly, choose it at the desk or use a one-file class link. Housing-only links enable the audit automatically.
2. Commit first reading is blocked: mark both questions, pin at least two values, choose a provisional next step, and record confidence. Congress additionally requires a supported observation selection and a brief limitation with an evidence request.
3. Continue or filing is blocked after a change: factual feedback applies to the checked batch and pins. Check the current batch again. For a percentage-point difference, enter the number without a percent sign and select Percentage points. Use the filing checklist for missing writing or the required evidence pair.
4. A student is stuck on factual checks: after an incorrect batch, staged help or a worked explanation is available. Acknowledging the worked explanation permits an assisted route once the current batch is checked. It records assistance and does not certify incorrect answers as supported.
5. Congress notes or housing provenance are missing: both are timed to follow the first commitment in the student file. Congress source information and the table are available beforehand; the housing matching-source lead appears after its first reading.
6. A fresh-case answer cannot be edited: its submitted first response is preserved. Use Optional · Try a separate second response after feedback. Once submitted, that second response is also preserved; begin a separate run if a new first response is needed.
7. The record says Draft: confirm the report’s assigned scope. Only those records count. If you opened a different class link while an older run was saved, Start assigned files downloads the old checkpoint before creating the intended run. Online recovery preserves each copy’s scope; create separate copies for different assignments.
8. Another tab changed the save: keep the current tab open and download it. Use the displayed recovery controls to download the available versions before deliberately loading or replacing the browser save. Do not close a tab containing the only copy of a draft.
9. Browser or online saving is unavailable: keep the page open and download a progress checkpoint. Resolve device-saving warnings before replacing work or syncing online. Online conflicts preserve the device draft; review the online copy or create a separate copy rather than forcing a replacement.
10. No instructor record appears: ask the student to confirm an online copy exists and that Allow instructor review is selected, then Refresh shared records. Students can still submit the downloaded report through the course system.
11. A download or print seems incomplete: locate the downloaded submission HTML and open it to review the writing and attempts. Print or save PDF opens recorded attempt details. Confirm the resulting file before submitting; a recovery card or progress JSON is a different artifact.

Earlier online copies belong to their original game website. Download a checkpoint there, restore it here, then create a new online copy on this website if needed. The instructor code is supplied separately.

Investigation directory

Primary source citations and limits are available in the game’s Instructor notes and research cards.

Activity page · Play the website edition

Created by Yoav Duman · V2.2. Timing is a planning recommendation; adjust after a classroom pilot.