Sim After Class: practicing everyday tradeoffs without predicting a life

Prepared by Education Angel Group (EAG), with AI-assisted research and drafting. Research date: September 19, 2026. EAG is developing Sim After Class and has an interest in its success. This is a focused research report about its proposed purpose, not an independent evaluation or a finding that the product improves financial outcomes.

1. Proposition, audience and the boundary of the claim

Sim After Class is a proposed school-safe decision game about careers, bills, homes and everyday independence. Its purpose is to make a choice understandable: orient to a situation, choose, observe a modeled consequence, continue and reflect. Its tagline is “Play through careers, bills, homes, and the choices that shape your life.” The central promise is practice. A fictional household's balance is not an actual account, and finishing a scenario is not evidence that a student is ready to live independently.

The current candidate is deliberately smaller than a life simulator. A player would review a fictional first-month home budget, compare two transport choices, select one, inspect its time and cost consequences, revise the selection and explain the tradeoff. Exact transport modes, numerical variables, uncertainty, time advancement, retry and save rules remain undecided. This report does not silently substitute a car-versus-bus fixture, set a recommended spending percentage or define a correct personal choice. The candidate comes from EAG's product definition, not from a study of Sim After Class. [A0]

Students are the intended players, but the target age range still needs definition. Parents and educators would need to understand the topic, model limits, accessibility and privacy before deciding whether the activity fits their setting. Scenario authors would need to keep assumptions intelligible and correctable. These audiences have different needs: a player needs a usable comparison; an adult needs an honest account of what participation can establish; an author needs a model that can be checked.

Overview. Research provides a reason to investigate active financial decision practice, but does not establish the effectiveness of this particular game. Existing public resources already teach budgeting, transport decisions and independent-living topics. Broad financial-education findings are heterogeneous and contested, and beneficial learning results can coexist with undesirable borrowing behavior. The defensible starting claim is therefore that Sim could offer a bounded opportunity to explain a tradeoff inside a disclosed model. Added value over a readable worksheet, persistence of learning and transfer to unfamiliar choices remain questions.

2. Questions that determine whether the idea is useful

The review asks six connected questions. First, can a student distinguish a fictional input from a fact about the real world? Second, can the student compare two choices across both money and time, rather than chase a single favorable number? Third, does revising a choice improve the explanation of its consequences? Fourth, does an interactive presentation add enough value over a static table to justify its complexity? Fifth, does understanding survive a delay and a different fictional scenario? Finally, could the experience produce overconfidence, shame or an impression that circumstances are entirely under personal control?

These questions place understanding ahead of completion. Someone who clicks through every screen but cannot explain the assumptions has not demonstrated the intended learning. Someone who chooses a more expensive option and explains the importance of available time may reason well. A system that treats the least expensive option as universally correct would measure obedience to its hidden preference rather than comparison of tradeoffs.

Three evidence categories remain separate throughout. Source findings describe what a study or institution reports. EAG interpretations explain why those findings matter to this proposed activity. Product hypotheses state what might work and must be tested. A randomized study can support a claim about its own intervention without supporting a claim about EAG's design. A public worksheet can establish that an alternative exists without proving that it changes behavior. A plausible learning mechanism remains a hypothesis until evaluated.

The most consequential unanswered question concerns transfer. The immediate task is to understand a modeled choice, not to optimize a real household. Applying a comparison method to another fictional problem is already a stronger claim than remembering the first result. Applying it months later amid real financial constraints is stronger again. The report treats those as different outcomes, not stages that automatically follow one another.

3. Need and context: financial decisions are not just arithmetic

The proposed educational need is an opportunity to rehearse unfamiliar decisions with consequences that can be inspected and reversed. A first-month home scenario can bring several considerations into view: an expense competes with other commitments, a lower cost can require more time, and the available choices can depend on context. Those are EAG's reasons for selecting a scenario. No survey or local interview in this review establishes how many students need this exact activity or would choose to use it.

The financial-education literature provides background, not a market-size estimate. Kaiser, Lusardi, Menkhoff and Urban's 2020 working paper synthesized 76 randomized experiments and reported positive average effects on knowledge and behavior. Interventions, populations and follow-up periods varied. Its long-run evidence was limited, and the authors did not establish durable effects for all programs. The accessible working paper is the version appraised here; it should not be treated as identical to the 2022 final publication. S1

For Sim, the relevant interpretation is that education is a plausible contributor to understanding, while the format and outcome need specification. A brief individual game differs from a sustained curriculum, a facilitated adult workshop and access to a useful financial service. No pooled effect becomes a predicted gain from this transport-choice loop. There is also no basis here for estimating financial return per player, reductions in future debt or an effect on independence.

Context matters even in fiction. Two transport choices may not both exist in a rural area. A seemingly convenient route may be unusable for a disabled person. A schedule may conflict with caregiving, shift work or school. A model cannot incorporate every circumstance, but it can state what it omits. Asking which missing factor might change a decision would make simplification part of the lesson, rather than hide it behind a realistic room or character.

A narrow activity may be valuable without covering financial capability comprehensively. Career exploration, housing access, consumer protection and getting appropriate help involve questions that a two-choice budget cannot answer. The scope should remain intelligible enough that a learner can say both what was practiced and what was not. Expansion into a larger simulated life should require evidence that additional complexity helps learning, not merely that more systems can be built.

4. Existing alternatives and the gap Sim would have to earn

These equivalents accompany the reader figures. Source IDs refer to the report references. A0 identifies EAG's product definition reviewed for the original report, not an external study or proof of availability. Other source IDs resolve to this report's reference and access record.

Figure 1. Existing alternatives already cover decision practice

A mixed authority and provision comparison, not an effectiveness ranking. FDIC access limits apply.

CFPB worksheet | S4: Calculate and reflect. Fictional transport-cost activity. Full worksheet read; impact not evaluated here.

FDIC materials | S5-S7: Games, curricula and Reality Fairs. Official indexed descriptions; direct retrieval restricted and gameplay untested.

Sim candidate | A0: Inspect and revise. Proposed two-choice home-budget loop. Added value over a static table remains untested.

Sources: A0, S4, S5, S6, S7.

Figure 1. This comparison describes available forms of instruction and EAG's proposed difference. Provision is not effectiveness evidence; FDIC descriptions were available through indexed official text, with direct page access restricted. Text and data alternative.

The CFPB worksheet Deciding which car and car loan you can afford is a particularly close comparator. Its four pages present a fictional situation, calculations, vehicle options and reflection. It uses simplified simple-interest arithmetic. Its figures are teaching assumptions, not current prices or a model for actual loan advice. The full worksheet was accessible. S4

That comparator makes a claim that “students have no place to practice transport costs” untenable. Sim's possible contribution is narrower: consequences that remain visible when a choice is revised, an understandable record of what changed, and an equivalent route through text and tables. Those features need comparison with a well-presented worksheet containing the same fictional information. Merely adding a scene, character or animation would not establish educational value.

FDIC's official indexed Teach Money Smart page describes interactive games and a guide for organizing Reality Fairs, where youth rehearse adult financial decisions. Its youth materials offer grade-specific instruction and caregiver resources. Direct retrieval of these pages failed in this review, so these are institutional provision descriptions, not verified product interactions. The game site opened to an informational or JavaScript-dependent shell, and no game session was tested. S5, S6, S7

These alternatives answer different needs. A teacher who needs a structured lesson may prefer existing curriculum materials. A facilitated fair can support discussion and multiple roles but requires organization. A learner seeking an independent exercise may prefer a short digital activity. Those are comparison criteria, not findings that one option is superior. Age fit, reading demand, access, preparation and explanation quality need to be assessed in the intended setting.

A useful research comparison would hold the fictional situation constant. One version could show both transport options and consequences in a static table; another could let the learner select, inspect and revise. Both would use the same wording, numbers and reflection prompt. Any difference would then be less confounded by better content in one condition. If the static version proves as understandable with lower burden, that is a substantive reason to retain it or simplify Sim. Existing alternatives should be treated as possible solutions, not as weak competitors constructed to justify a new product.

5. Evidence, counterevidence and limits of transfer

Figure 2. Learning evidence does not establish simulation transfer

Qualitative evidence map. No shared effect size, product forecast or assertion that the critique is a formal correction.

S1 and S8 | synthesis and critique: Average effects remain debated. Broad positive synthesis versus provisional publication-selection critique; neither tests Sim.

S2 | adult Uganda experiment: Mixed behavioral outcomes. Facilitated active learning differs from a child digital game; adverse findings retained.

S3 and S3a | adolescent Brazil evidence: A sustained bundled curriculum. Benefits and undesirable borrowing coexist; final and working versions stay separate.

Sources: S1, S2, S3, S3a, S8.

Figure 2. The retained studies and critique answer different questions. This qualitative map contains no expected Sim effect and does not combine different study populations or publication versions. Text and data alternative.

The broad synthesis in section 3 is not the end of the debate. Fernandes, Lynch and Kim's author-uploaded November 2025 manuscript argues that allowing publication selection to depend on both statistical significance and direction yields smaller behavioral estimates than the earlier synthesis. The retrieved manuscript is provisional counterevidence, not a formal correction or retraction. Its methods were not independently reproduced for this report. S8

That disagreement changes the confidence appropriate to claims about typical effect size. It does not settle whether a specific, well-designed exercise can help learners explain a choice. The sensible response is to avoid borrowing a headline effect, compare the actual candidate against a credible alternative, and disclose uncertainty. Selecting only the favorable synthesis would overstate the case; treating a later critique as proof that all education fails would also go beyond its scope.

Kaiser and Menkhoff's Uganda experiment offers a closer methodological comparison of teaching approaches. Adult retailers received active learning, lectures or a control condition. The working manuscript reports benefits on some savings and investment outcomes, but also adverse late-payment findings. Long-term follow-up had substantial attrition and uncertainty. This was facilitated adult education, not a child digital game, and the active treatment bundled several activities. S2

The EAG interpretation is to test active comparison while retaining the possibility of mixed outcomes. That study does not isolate choice revision, reflection, digital presentation or a character guide. The term “active” should describe a meaningful cognitive task, such as explaining why a result changes, rather than serve as a label for tapping a button. More interaction is not necessarily more reasoning.

The final abstract of Bruhn and colleagues' Brazil study reports improved financial proficiency and saving or budgeting alongside increased expensive consumer credit. Its sample is described as approximately 25,000 students in 892 schools. The accessible 2013 World Bank working paper describes an earlier sample and a sustained, bundled school intervention. Those versions must not be combined into one denominator or effect table. S3, S3a

The adverse finding is directly relevant to how Sim should be evaluated: greater knowledge or confidence cannot stand in for safer behavior or improved welfare. A student might learn the interface's arithmetic and still overgeneralize its advice. The first study of Sim should therefore examine misunderstanding and confidence calibration as well as correct comparisons. None of these sources tests the proposed EAG slice. There is no demonstrated transfer from this game to real household decisions, and no claim here that a student will save more, borrow less or make a better career choice.

6. Proposed mechanism: expose a choice, then make revision intelligible

Figure 3. The proposed practice loop stops before real-life claims

EAG conceptual mechanism, not tested causality. No actual account value or financial recommendation.

Inside the fictional model. Review home budget and assumptions → Compare two transport choices → Choose and inspect time/cost effects → Revise and explain the tradeoff. Exact modes, amounts, uncertainty and time rules remain undecided.

A separate research question. Try a new fictional scenario → Assess understanding after a delay. Generalization and retention need evidence. Neither establishes real financial readiness.

Sources: A0.

Figure 3. EAG's conceptual mechanism starts with a disclosed fictional home budget and two transport choices. All links remain hypotheses. No numerical fixture, probability distribution or actual financial recommendation is implied. Text and data alternative.

The model would begin by telling the learner whose fictional situation is being considered, what period it represents and which variables are fixed. Money and time should use explicit units and the same period. The learner would then compare exactly two transport options and choose one. Consequence feedback would show the changed quantities and the rule responsible. Revision would return to the same decision point so that the learner can examine the other option without accidentally advancing the month twice. These are proposed design requirements derived from the current product definition. [A0]

The hypothesized learning mechanism is comparison with explanation. The learner should be able to connect an input to a result, see the tradeoff between dimensions and describe why a different priority could change the selection. Reflection might ask what mattered most, what changed after revision and which omitted circumstance would make the model less useful. A response can be tentative. The point is to make reasoning inspectable, not require a polished justification for an adult audience.

Model truth needs several distinctions. An authored value is a fictional assumption. A calculation is an output of a stated rule. A range can represent deliberate scenario variation only if its meaning is explained. A random event is not an empirically estimated probability simply because software draws it randomly. Without verified data, numerical uncertainty would be an authored teaching device and must be labeled that way. This report therefore uses conceptual figures instead of inventing a budget or probability distribution.

The model should also expose what it holds constant. If the same income and fixed expenses apply to both options, that makes comparison easier but does not imply those quantities are stable in life. If access, weather or family responsibilities are omitted, the consequence explanation should identify those omissions when relevant. Adding an unexpected setback merely to make the game dramatic could teach the wrong lesson if its cause or frequency is arbitrary.

The mechanism ends at an explanation within the model. A later unfamiliar task would test generalization; a delayed task would test retention. Real behavior would require a different study with a much more demanding ethical and contextual design. These are not automatic outputs of completing the first loop. The product should not convert a saved scenario into a financial-readiness badge, school assessment, eligibility rule or social comparison.

7. Design concepts worth investigating

The first concept is a decision sheet that remains readable throughout the interaction. It would keep the premise, fixed assumptions and two options close enough to compare without memory-heavy navigation. A player should be able to inspect both time and cost before selecting. The visual route and the text/table route would carry the same quantities, consequence rules and opportunities to revise. This is an accessibility and model-consistency requirement, not a finding that a specific layout has been tested.

The second concept is a consequence explanation organized around the change. It would distinguish the starting state, the selected action and the resulting state. A small explanation such as “this value changed because this rule used this input” is a candidate for testing; it must ultimately use approved plain-language content. The result should not be labeled a winning life plan. If a learner revises, the comparison should show how the revised result differs from the prior result under the same assumptions.

The third concept is a reflection that can be private, brief and skipped. A learner might identify a priority, note a missing factor or choose to leave without recording an explanation. Educators may facilitate discussion, but the game should not require disclosure of household income, debt, transport insecurity or family conflict. The fictional situation gives sufficient material for the intended task. Real personal data would add privacy and emotional burden without being necessary to show the practice loop.

A fixed-camera room or guide is an optional presentation candidate in the product definition. Neither should be a prerequisite for choosing or understanding. If assets fail or a device cannot render a scene comfortably, the useful decision route must remain available. Reduced motion should preserve the full explanation. Visual realism should be assessed against comprehension, loading burden and the risk that players mistake a detailed setting for a validated model.

The proposal also requires honest continuity. Pausing or backgrounding should not silently advance modeled time. A returning learner should understand whether the prior choice and reflection were saved, unavailable or intentionally reset. If sign-in becomes necessary, Runstr is the designated identity authority, with product-specific save permissions. Anonymous practice need not force account creation. Shared identity does not authorize sharing simulation state, scores or personal information across products. These boundaries come from EAG's definition and remain requirements awaiting implementation evidence. [A0]

8. Risks, equity and the work the experience creates

A central risk is confusing resource constraint with personal failure. If every setback is explained as a poor choice, the activity can imply that real hardship is a character defect. A fictional premise should allow mixed outcomes and reasonable disagreement. Both transport options can be defensible under different priorities, and some real situations offer no acceptable option. The model should help learners recognize that limitation rather than award moral rank.

A second risk is false confidence. Immediate, exact feedback is useful for understanding a calculation, but can create an impression of certainty that real decisions do not have. Attractive presentation, repeated success and fluent explanations may amplify that impression. The evaluation should check whether learners can state an important limitation after playing and whether confidence tracks demonstrated understanding. A disclaimer at entry is insufficient if the result screen communicates certainty or personal endorsement.

Access has practical dimensions beyond a nominally available alternative. Small-screen comparisons must keep units and labels readable. Keyboard navigation must permit the whole choice-and-revision cycle. Text resizing, slower reading and interruptions should not be punished through timers. A non-3D route must retain the learning task rather than reduce it to a summary. Actual device and assistive-technology evidence would be needed before claiming access; this report provides none.

Adults and operators also bear work. Teachers may need to explain unfamiliar terms, resolve discomfort and help students separate fiction from advice. Authors must review assumptions when scenarios change and preserve earlier explanations when a saved state belongs to a different model version. Support time is part of the cost of the design. A game that shifts extensive interpretation to adults may still be useful, but should not be marketed as effortless independent learning.

Finally, product boundaries matter. No actual account value, credit transaction, financial recommendation or professional guidance belongs in this practice slice. No leaderboard should rank simulated wealth or readiness. A student should be able to exit safely, decline reflection or start again without humiliation. These are product constraints and foreseeable design concerns, not evidence that harm has occurred or that the proposed safeguards have been validated.

9. Evaluation and decisions that could change the direction

Figure 4. Different claims require different evidence

Proposed evaluation framework. No participant results, validated rubric or completed comparison.

Immediate comprehension: Explain the model. Identify assumptions and compare cost/time. Correct clicking alone is insufficient.

Delayed generalization: Explain a different case. Test a new fictional scenario after a justified delay; compare an equivalent static activity.

Real-world behavior: Requires a separate study. Do not infer financial benefit, life prediction or readiness from simulated results.

Sources: A0, S1, S2, S3, S8.

Figure 4. This is a proposed evaluation framework, not completed research. Immediate comprehension, delayed generalization and real-world behavior require separate measures; no fictional result is a readiness score. Text and data alternative.

Initial verification should establish that the activity behaves as described. A disclosed fixture and its rules should yield identical results in the visual and text/table routes. Revising should not double-count a payment or time advance. Interruption and resume should preserve or honestly report the state of the choice and reflection. These checks address model and interaction correctness. They do not establish that students learned anything.

A first usability study would ask participants to explain the fictional premise, identify relevant inputs, compare both dimensions and revise a choice. Observers would distinguish arithmetic errors, confusing labels, navigation problems and misunderstandings of the model's status. Adult help should be recorded because unaided and facilitated performance answer different questions. Participation methods, age range, consent and privacy would need separate authorization before recruitment.

A subsequent comparative learning study could use the equivalent static table described earlier. It should specify the primary outcome before collecting results: for example, the quality of a tradeoff explanation on a new fictional scenario. Enjoyment, completion, confidence and time on task would be additional measures with their own interpretations. A delayed assessment would test whether understanding persists. The delay, sample size and scoring rubric should be justified for the selected population rather than copied from unrelated studies.

The rubric should reward identification of assumptions, coherent comparison and recognition of omitted factors. It should not reward choosing the author's favored transport mode. Scoring examples could include different defensible choices. If evaluators know which condition a student received, expectation can affect judgment; independent scoring with condition hidden where feasible would strengthen interpretation. Missing follow-up responses must be reported rather than treated as unchanged or successful outcomes.

Several findings would change the proposed direction. If learners consistently treat the output as advice, simplify or rewrite feedback before expanding content. If the table performs as well with less burden, prefer the simpler presentation. If visual elements impede access or distract from explanation, remove or reduce them. If comparison improves only within the original fixture, describe in-model learning and investigate transfer without claiming it. If confidence rises while explanations remain weak, treat that as a warning, not a success metric. A positive pilot would justify only the next bounded inquiry, not broad financial-outcome claims.

10. Methods, source access, references and corrections

This focused review follows the question in EAG's September 19, 2026 entity brief: when can a modeled everyday decision help students understand tradeoffs, and how limited is transfer? It is not a systematic review. Searches followed original financial-education studies, an active-learning experiment, adolescent evidence, public alternatives and contrary findings. Selection favored original research and institutional materials with identifiable versions and relevant outcome definitions. No participants were interviewed and no Sim runtime, household dataset, current price series or real financial account was examined.

Working manuscripts were reopened from IZA, CESifo and the World Bank. The final Brazil publication was available as an abstract, so its final sample description and outcomes were kept separate from the earlier full working paper. The CFPB worksheet was accessible in full. FDIC alternative descriptions came from indexed official pages because direct retrieval failed; no live accessibility, saving or gameplay behavior was verified. A newer methodological critique was available as author-uploaded manuscript text, while SSRN retrieval returned an access error. These access differences constrain the strength of the report.

The correction search surfaced substantive disagreement about publication-bias adjustment, not a verified retraction of the Kaiser synthesis. No statistical replication was performed. The review does not establish that all later responses, errata or relevant simulation-transfer studies were located. Its strongest limitation is the absence of direct evidence for this exact age-specific, two-choice digital activity. Conceptual diagrams describe questions and mechanisms; they contain no measured product effects. None of the institutional references implies endorsement of EAG.

References and source locators. All external materials were checked September 19, 2026.

Corrections should identify the source, version, affected statement and reason for revision. A later approved fixture, age definition or direct evaluation could materially change the recommendations. Until then, the report supports investigating transparent decision practice and measuring its limits. It does not establish readiness, real-life prediction or financial benefit.

Figure text and data alternatives

Complete semantic alternatives accompany each figure in this reader and its text download. Source IDs refer to the references in this report. A0 identifies EAG's product definition reviewed for the original report, not an external study or proof of availability. Other source IDs resolve to this report's reference and access record.

Corrections and entity context

Readers identifying an error can contact EAG with the section, disputed statement and supporting source. Sim After Class provides the public entity context, not an external citation for EAG's product claims.