Blog / Study explainer · 2026-10-27 · By The Evidence Signal

Read a study with us: why a five-year study of 111 schools is rated Level 2

Evidence level: Level 2: Correlational (automatic rating from our published rules, not yet checked by a person). The study follows schools over time and finds that some things go together. It does not compare groups, so it cannot show what caused what. What the levels mean

Source: The Longitudinal Relationship between Schools' Academic Optimism and Academic Achievement in Elementary Schools: Evidence from a School Panel Study, Large-scale Assessments in Education, 2026 (listed in ERIC as peer reviewed). Part of our Research beyond AI coverage. We read the abstract, not the full paper.

Why we read this one together

A headline could easily turn this study into "Schools with confident staff have better readers." The abstract does report that link for one measure. Our rating is Level 2 out of 4, and the details matter. Here is how we got there, one question at a time. You can ask the same questions of any study with our free headline checker.

Step 1: Is there a study, and who took part?

Yes. The study used data from school principals and fourth-grade students in 111 elementary schools in Germany, taking part in a school panel study in 2016 and 2021. These are real schools and students, so our method rates it on Levels 1 to 4. Studies that use only simulated students are rated T (Technical test) instead.

Step 2: Was there a fair comparison?

This is the step that decides the level. The abstract does not describe putting schools or students into groups. The researchers measured each school's "academic optimism" in 2016 and looked at whether it went with fourth graders' reading scores in 2021. The abstract names three parts of academic optimism: academic emphasis, collective efficacy and faculty trust.

The researchers took some differences between schools into account, such as socioeconomic status and migration background. That makes the link more believable, but it is still a link. Schools with higher collective efficacy may differ in other ways that also help reading. Our method rates studies like this, which find that two things go together, at Level 2.

Step 3: What was measured?

Students' reading achievement. That is a learning outcome, so the cap for studies that only measure feelings or engagement does not apply.

Step 4: Was it peer reviewed?

Yes. It appeared in a journal after expert review, so the preprint cap does not apply. Peer review does not raise a level on its own.

Step 5: How many, and for how long?

111 schools over five years. Our method lowers a level when fewer than 100 learners are in the results and the study gives no range showing how precise the result is. The abstract reports 111 schools but not how many students. With fourth graders from 111 schools the count is likely above 100, but a person should confirm it in the full paper. A separate drop applies when whole classes or schools are put into groups with only one in a group. No schools were put into groups here, so it does not come up.

Step 6: Who paid for it?

The abstract does not mention funding. OpenAlex, a public research database, lists Germany's Federal Ministry of Education and Research as a funder. This is not a study of a product, so our hold on company-linked studies does not come up. We have not read the paper's funding or conflict of interest statements.

What this means for you

Read past the headline. Only one of the three parts, collective efficacy, went with later reading scores. Academic emphasis and faculty trust showed no statistically significant link.
Look at the second finding too. The abstract says none of the three parts was related to the change in reading scores from 2016 to 2021. A link with later scores is not the same as a link with improvement.
It does not show that building collective efficacy will raise reading scores. A Level 2 study can point to something worth trying. It cannot show that it works.
If your school is working on collective efficacy, this is a reason to keep watching your own data, not proof of a payoff.

What we have not checked

We have not read the full paper. We do not know how academic optimism was measured, how many students took the reading tests, how many schools dropped out of the panel between 2016 and 2021, or how large the links were. Of these, only the number of students in the results, or a problem with the reading test, could change the rating. The rest still matter for how much weight to give the findings.

How the rating would change

It cannot move up on design. A study that finds links without comparing groups stays at Level 2 or below in our method, however large or long it is.
It could drop to Level 1 if a person reading the full paper finds a problem our method counts against a study, such as fewer than 100 students in the results with no range showing precision, or a reading test made by the researchers with no evidence that it is reliable.