Educational research — information for parents and schools

What research is this?

We study how classroom simulations help students learn and how groups make decisions in classic economics games. Findings help improve teaching and may appear in aggregate form in educational research. No individual child is ever the subject of a report.

It is optional and teacher-controlled

Research inclusion is off by default. A teacher can choose to opt a specific session in; they can turn it off again at any time. Nothing about the lesson changes for the student either way.

Individual play at home

Children often use Prevolio on their own, at home. That play becomes part of the research only by one of two routes, and an adult chooses both of them.

• A teacher assigns a session as homework. The teacher opts that session in; the students then play it individually, and their play is covered by that one decision. Nothing is asked of the student.
• A parent or guardian starts a game at home and ticks the research box themselves. Those sessions are marked as home play, so researchers can tell family use apart from classroom use.

Play that is never part of the research

Anyone who simply opens a simulation and plays it — no session, no adult involved — is never joined to the research dataset. Visits like that are counted only as anonymous totals: how many times an activity was started, how many of those runs finished and how many were left part-way, in which language and on what kind of device, and — for a test with sections — how far people tend to get before stopping. These are counts of visits, not of people: nothing is stored that could tell us whether two visits came from the same person, or link a visit to anything else. A figure is published only once enough separate visits lie behind it.

There are two additions. The first is for two short tests only — the Big Three and the Big Five financial-literacy questions. When someone takes one of them on their own, we also keep, without any identifier, how the answers came out: for each question whether it was right, wrong, “Do not know” or skipped, and the overall number right, together with the language and the day. Only the first attempt in a browser is kept (a browser that cannot store anything locally, such as some private windows, may be counted again after a reload), nothing ties these records to each other or to the person, and they are only ever published as totals once enough attempts lie behind them. A browser that sends a Do-Not-Track or Global Privacy Control signal sends nothing.

The second is The Ombudsman's Desk, a simulation on the Constitution of Ukraine. Before a learner's first shift, and once more after the fourth, it offers ten short questions on the Constitution; answering them is optional, and nothing is sent unless the learner does. We keep, without any identifier, only the number right out of ten, whether it was the check before or after, the language, the day, and whether a law teacher had finished reviewing the desk's content at that moment. That review is still under way: until it is complete, the desk says “Content is not yet fully verified.”, and each result is labelled with whether the review was complete when it arrived, so the two are counted separately. Nothing ties a learner's two checks to each other or to the person, and the results are only ever published as totals once enough checks lie behind them. A browser that sends a Do-Not-Track or Global Privacy Control signal is not offered the questions and sends nothing.

What is collected

• The choices students make inside an activity (for example a contribution, an offer, or a test answer), tied to the session — not to a named child.
• Optionally, general context the teacher provides about the class: country, grade band, class setting, and whether the topic was taught beforehand.
• Optionally, short 2–3 question surveys about how enjoyable or clear the activity was.

What is never collected or shared

• No student names, emails, or accounts — students play with a code, not an identity.
• No student free-text writing in any published dataset.
• No advertising cookies, third-party trackers, or precise location.
• Nothing that could identify an individual child.

How the data is anonymized

Before any research use, data is de-identified: identifiers are removed, session references are replaced with non-reversible keys, times are coarsened, automated participants are filtered out, and any statistic covering fewer than ten students is suppressed. Published results are always group statistics, never individual records.

Your choices and contact

You can ask the teacher not to include a session, or contact us directly to have a session excluded from research or to ask a question. We will respond promptly. The full detail is in our privacy policy.