España 
Italia 
United States 
Brasil 
Chile
Ecuador
Mexico
Colombia
Perú
This independent study by WestEd shows that, after two years of implementation, students who used Innovamat’s resources achieved better results in mathematics than those who did not, with an effect size of 0.24.
In the 2023-2024 school year, Innovamat reached U.S. classrooms for the first time. For us, that step could not be limited to implementing the program in a new country. We also wanted to understand what impact Innovamat could have in this new context.
That is why, from the start, we set out to accompany its arrival in the United States with an independent effectiveness study. We wanted to measure results, but also to learn: what happens when a school begins to transform the way it teaches math? How much time does that change need to become established? What effects are observed in student learning?
Why conduct a study in the United States?
The United States has a long tradition of evaluating educational programs. It also has reference frameworks such as the What Works Clearinghouse (WWC), an initiative of the Institute of Education Sciences that helps identify studies that meet certain evidence standards. In this context, the evidence levels of the ESSA (2015) law are also used to classify the strength of the available evidence on an educational program. The main point is that effectiveness studies do not all have the same rigor and do not allow us to draw conclusions with the same certainty. The level of evidence depends on how the study is designed. These levels range from justifying the scientific or theoretical basis on which a project is based (level IV) to demonstrating the positive impact of the program with a rigorous, well-designed, and well-implemented experimental study (level I).
But the key factor that allowed us to move forward with this study is that the opportunity arose to receive funding from the Jacobs Foundation, a foundation that promotes the use of evidence in education and the development of EdTech solutions. One of the conditions for receiving this funding was to work with an independent institution recommended by the foundation itself. Among the available options, we chose WestEd, a nonprofit organization specialized in educational research and evaluation.
The hypothesis: impact takes time
From the first conversations with WestEd, we shared a clear hypothesis: transforming math instruction is not something immediate.
Innovamat is not just about replacing some materials with others. It means introducing a different way of learning math: more discussion, more student agency, more use of multiple representations, connected sequences, work with concrete materials before moving to the symbolic, and practice designed to consolidate knowledge and develop fluency, etc.
Like any educational change, this shift takes time. The first year is usually a year of adjustment: classroom dynamics, teaching routines, and student participation patterns all change. Our hypothesis was that improvements in math performance would be more noticeable starting in the second year of continued use.
With this idea, WestEd designed a two-year longitudinal study. In other words, student progress was followed for 2 years from the moment they began using Innovamat resources.
How was the study designed?
The study was carried out in New Jersey with 2nd and 3rd grade students during the 2023-2024 and 2024-2025 school years. At that time, only three U.S. districts were using Innovamat. Of those, two used the same external assessment, LinkIt!, which made it possible to have comparable data from the start. This point was key. To know whether a program has an impact, it is not enough to look at how students end up. We also need to know where they started. That is why WestEd used the LinkIt! results from fall 2023 as the starting point, and the fall 2025 results as the main measure after two years of implementation.
LinkIt! is a benchmark assessment that some districts were already using regularly to measure progress in math during the school year. It was not a test designed by Innovamat or administered specifically for the study, but rather an external source that was already part of the districts’ regular operations.
Because it was not possible to randomly assign which schools used Innovamat and which did not, WestEd designed a quasi-experimental study. Put simply: it compared students using Innovamat with similar students from other schools that did not use it.
To build this comparison, WestEd started with an initial list of New Jersey schools that had not adopted Innovamat. From this list, LinkIt! provided anonymous data to WestEd from 1,650 students in 17 districts that also used this assessment. From there, WestEd applied a student-level matching process: for each student in the Innovamat group, it looked for another student as similar as possible in variables such as initial performance, grade, gender, ethnicity, socioeconomic status, or whether they had an individualized education plan. In total, the final study sample included 910 students, of whom 455 used Innovamat.
The logic is easy to understand: if we want to compare results, we must first make sure the groups being compared are as similar as possible before we begin. This type of design, called a quasi-experimental design with matched groups, is recognized in the manuals of effectiveness studies of the What Works Clearinghouse at the ESSA Tier II level. However, in cases where the sample is clustered in a small number of schools and this clustering is not accounted for, the design is considered part of the ESSA Tier III level.
Results: performance improves starting in the second year
As the main result, it was found that after two years of implementation, students who used Innovamat achieved better results in math than the comparison group.
Specifically, WestEd found a statistically significant difference (p = .002) between the two groups, with an effect size of 0.24. Although these terms are technical, the idea is simple: the observed difference does not seem to be explained by chance, and the size of the effect is relevant in the context of educational research. According to WestEd, a difference of this size is approximately equivalent to a student at the 50th percentile moving to the 58th percentile. In other words, if we ranked 100 students according to their results, a student in the middle of the group using Innovamat would move to approximately 58th place. This translation helps give an intuitive sense of the result, although it should be interpreted as an approximation.
Note: the “Treatment” group shows the normalized average results of the students using Innovamat, while the “Comparison” group shows the results of students with other educational proposals.
What also matters is what happened after the first year: no significant differences were observed between the Innovamat group and the comparison group. For us, this result is just as important as the previous one, because it fits the initial hypothesis. Deep changes in instruction need time to become established.
The study also explored whether the impact varied across different subgroups. The most notable result was that the effect seemed to be greater for students who started the study in 3rd grade than for those who started in 2nd. These analyses, however, were exploratory and should be interpreted with caution.
Teacher perception: more student-centered instruction.
In addition to analyzing academic results, WestEd interviewed teachers who had implemented Innovamat. Their responses help us better understand what was happening in classrooms.
Teachers described a shift toward a more student-centered instruction, with more math discussion, more discussion of strategies, and more opportunities to participate. They also highlighted an increase in student confidence and engagement, especially among students who had previously felt more distant from math. It is important to note that these observations are very consistent with the testimonials collected in another district in New Jersey, Upper Freehold. For example, the WestEd report includes comments from teachers such as the following:
“You have to hand it over to the kids. You have to step back a little bit, have a little patience, let them figure things out, let them talk amongst themselves, and that’s the switch.”
“… the conversations that kids can hold about math are phenomenal. They talk their way through it. They’re getting to these understandings that I don’t think I could have gotten to as a second grader because they’re just constantly having to prove why. ”
” What I love the most is the whole class enjoyed math. … Those children who came in in the beginning of the year who didn’t like math, who struggled with math in your math lesson, all of the children were able to participate…. I feel like a lot of confidence was gained through their lessons. … a month or two in all the children were participating, which was unusual. A lot of times you had those children who just turned off to math.”
At the same time, they identified challenges typical of the first years of implementation: the teacher learning curve, the preparation of materials, alignment with local standards, assessment tools, and some technical aspects. These challenges remind us of something important: the quality of an educational proposal depends not only on its materials, but also on the support and the implementation process
In this sense, one of the aspects teachers valued positively was the support received from Innovamat advisors, which helped them face the shift in approach, adapt the materials to their reality, and bring the program into the classroom
“…The support could not have been better… they were v eryreceptive to our needs.”
How to interpret these results?
This study provides promising evidence of the impact of the Innovamat program, according to ESSA levels of evidence (Tier III). In this context, and after two years of use, Innovamat is associated with a significant improvement in mathematics results.
However, it is important to be precise. This is not a randomized experiment, but a quasi-experimental study with matched groups and a relatively small sample. This design improves the comparison between groups, but it does not completely eliminate the possibility of unobserved differences between schools that adopted Innovamat and those that did not. For this reason, WestEd indicates that the results should be interpreted with causal caution.
It is also important to put the results in context. Evaluating the impact of a curricular program in real conditions is not easy. As Kraft (2023) indicates, a high percentage of studies in education do not find any relevant impact. This is consistent with the interpretation shared with us by the study’s principal investigator, Mingyu Feng, when she presented the results: “These results are very positive. In most education evaluation studies we do at WestEd we don’t see significant results. Moreover, if improvements are detected, effect sizes are usually smaller.”
On the other hand, as indicated in Bellwether’s “Rounding Up” report on mathematics curriculum efficacy studies in the United States, most programs are evaluated through correlational designs (ESSA Tier III) or quasi-experimental designs (ESSA Tier II). It is not common to use experimental designs with random assignment of who implements the program to be evaluated, as they have a very high cost and are very difficult to implement and coordinate. Some of the most relevant studies in the United States on mathematics programs can be found on the Evidence for ESSA page. If you do a bit of research on studies of mathematics curriculum programs involving the whole class group, you will see that effect sizes are around 0.1, and rarely exceed 0.15.
Furthermore, these results are consistent with other evidence we have been collecting in different contexts. In Spain, for example, our internal studies pointed to a positive relationship between accumulated experience with Innovamat and results in mathematics tests. We had also observed that teachers especially value the impact of the program on student participation, motivation, and learning, as well as on their own professional development.
Overall, this first independent study in the United States points in the same direction as other previous evidence. It encourages us to continue researching, learning, and evaluating with increasingly larger samples and more rigorous designs. The final goal remains the same: to better accompany schools and teachers to improve the teaching and learning of mathematics.
References:
Gold, T., Carroll, K., Newby, L.D.T., and Steel King, M. (March 2023). “Rounding Up: An Analysis of Math Curriculum Effectiveness
Studies.” BellwetherKraft, M. A. (2023). The Effect-Size Benchmark That Matters Most: Education Interventions Often Fail. Educational Researcher, 52(3), 183-187.
