Level 4 review: how a study earns our top rating
Levels go from 1 to 4, and higher is stronger. Level 4 is the strongest evidence we give, and only a person can award it after reading the whole study.
What Level 4 means
A Level 4 study is a fair test with real students, done carefully enough that a person has checked it from start to finish. It is either a randomized trial (students or classes were assigned to use the tool or not by chance) or a review of many controlled studies (studies that compared a group using the tool with a similar group that did not) with real students that searched for studies carefully and checked their quality.
What a study needs to reach Level 4
- It is a randomized trial, or a careful review of many controlled studies.
- Real learners took part, and it measured learning, not just how people felt about a tool.
- It was published after review by other experts. Preprints (papers posted online before expert review) stop at Level 3.
- Students used the tool for more than one session and for at least a week.
- If a company with a stake in the product is linked to the study, a person checked who paid for it and whether the team was independent.
- No serious problem lowered its level, such as a group too small to be precise or whole classes assigned but students analyzed one by one. See the Labels page.
- For a review of many studies: it is rated no higher than the studies it includes.
- It reports how big the effect was, with a margin of error, and the result is precise enough to rely on.
- A person checked that students dropping out could not skew the result and that the groups started out alike.
- A person checked that at least one test was not made only by the researchers, or that the test has evidence it is reliable and measures what it claims to.
- A person checked whether the study said in advance what it would measure, and noted any change in what it reported.
Why only a person can award it
Our automatic first pass reads a study's summary. A summary usually does not show who dropped out, whether the groups started out alike, or how the test was built. Those checks need someone to read the full paper, so the software can never award Level 4 on its own. A study that might qualify is shown at Level 3 with a Level-4 candidate tag and flagged for a person while it waits.
What Level 4 does not mean
A level tells you how much to trust a study's result, not whether the result was good. A strong study can find that a tool did not help. Level 4 is not an endorsement or an official rating, and you should still check that the students and setting look like yours.
Where things stand
As of 2026-10-26, no item in our tracker holds Level 4. When one does, it will be listed in the tracker. For every level, see the Labels page. For the full rules, see the Method page.