Math anxiety and poor numeracy tend to arrive together. Students who freeze on a problem fall behind, and falling behind makes the next problem scarier. Most classroom fixes tackle one or the other.
Researchers at Almuslim University wanted to know what happens when you design for both at the same time. They built a statistics module inside Edcafe, ran it with real high school students, and measured what changed. Here’s what they found.
Disclosure: Edcafe was not involved in the design, funding, or review of this study. It was independently conducted and published by researchers Friantiani Safitri, Siti Khaulah, and Bulan Nuri, with funding from DPPM 2025.
The study at a glance
Almuslim University’s researchers chose a specific starting point: Grade 10 students at SMA Negeri 1 Lhoksukon in Aceh, already midway through a statistics course and already showing signs of math anxiety the study was designed to address.
Friantiani Safitri, Siti Khaulah, and Bulan Nuri from Almuslim University decided to build something different. They didn’t just try a new tool in class and see what happened. They used the ADDIE framework, covering Analysis, Design, Development, Implementation, and Evaluation, to design, validate, and test a full learning module inside Edcafe before students ever opened it.
The validation ran in stages. Academic experts reviewed the design for theoretical soundness. Four practising mathematics teachers evaluated it for real-world usability. A small group of 6 to 10 students piloted it first, then a full class took part in the field trial.
The study was published in the Jurnal Penelitian Pendidikan IPA, Volume 11, Issue 11, in November 2025.
The problem they set out to solve
The researchers weren’t trying to fix two separate things. They frame math anxiety and numeracy gaps as a cycle: anxiety reduces a student’s willingness to attempt problems, and avoiding problems means the underlying skill never develops.
Three structural issues drove that cycle in their context:
- The subject itself. Statistics requires multi-step procedures where one error compounds into the next. Students who already doubt themselves tend to shut down before they start.
- The tools in the room. Most teachers rely on whiteboards, printed books, and projectors. Those tools deliver content but don’t give students a way to try, fail, and get corrected without it feeling high-stakes.
- The timing of feedback. When feedback only arrives after a test is graded and returned, the anxiety from the attempt has already settled in.
That’s the design problem Edcafe was brought in to solve.
What they built inside Edcafe
The researchers designed a specific module, built it inside Edcafe, and specified exactly how each feature was meant to work. Four components made up the full design:
- Slide decks. Statistics content for Grade 10, built directly in Edcafe for students to access at any time.
- Student worksheets with an integrated assessment grader. Students submitted answers and received a score, a correction tied to the achievement standard, and a targeted remedial suggestion, immediately, before moving on.
- A teacher-supervised chatbot. Students could access it independently outside class hours. Teachers controlled what it could discuss, keeping it grounded in the lesson material, so students had somewhere to go when stuck without the chatbot going off-script.
- Self-paced access. The full module was available for students to work through on their own time, not only during class.
The feedback structure was the deliberate core of the design. Every submission returned a correction before the student moved on, closing the gap where math anxiety normally takes hold.
Quiz Practice Mode does the same thing for quizzes. Students get feedback after each question, before moving on. See how different quiz modes can support learning →
Expert and Practitioner Ratings on the Almuslim University Module
Before any student opened the module, two independent groups evaluated it: academic experts in media and subject matter, and four practising mathematics teachers. Here’s how each one scored the design:
| Validator | Score | Category |
|---|---|---|
| Media expert | 91.66% | Very valid |
| Content expert | 88.33% | Very valid |
| Practitioner 1 | 97.22% | Very valid |
| Practitioner 2 | 95.83% | Very valid |
| Practitioner 3 | 94.44% | Very valid |
| Practitioner 4 | 84.72% | Very valid |
The “very valid” category is the highest band in the validation scale used by the study. Media validity assessed whether the app’s design, layout, and interactive elements were appropriate for the target students. Content validity assessed whether the statistics material was accurate, well-structured, and aligned with learning objectives.
The four teachers evaluated it for real-world usability: could this actually be deployed in a classroom, and would it support the learning process? All four scored it in the highest band. The lowest score, 84.72%, still cleared the threshold for “very valid” by a significant margin.
What the validation confirmed: the people most qualified to judge the design, both as academics and as classroom teachers, said it was ready to use.
Student Results: How Almuslim University Students Responded to Edcafe
After the field trial, students completed a 20-item questionnaire covering the module’s language, design, content, usability, and their own engagement. Across all 20 items, not one student strongly disagreed with any statement.
Three findings are worth pausing on.
100% of students said the module made them more curious to learn. Curiosity is exactly what math anxiety kills first: anxious students don’t explore, they avoid. A full agreement rate on this item suggests the design reduced the defensive posture that tends to shut learning down before it starts.
100% said the activities improved their numeracy skills. This matters because the study wasn’t just trying to make students feel better about math. It was trying to build the actual skill. Students reporting improvement on both at once is the closest the study gets to confirming the core hypothesis.
60.71% strongly agreed that Edcafe made them more active in their learning, the highest strongly-agree score of any item in the questionnaire. “More active” here means answering, attempting, and completing tasks rather than sitting back.
The King Fahd University study found that this type of participation (doing the work rather than watching it) was the only engagement type that predicted higher grades.
The two specific features students credited most were the teacher’s ability to embed images in lessons and the supervised chatbot they could access independently. Both reduced the experience of being stuck with nowhere to go.
What Happened to Their Math Anxiety
Anxiety levels were measured before and after the trial using a questionnaire that placed students into five categories. Here’s what changed:
| Anxiety level | Before Edcafe | After Edcafe |
|---|---|---|
| Very high | 3 | 2 |
| High | 11 | 8 |
| Moderate | 7 | 8 |
| Low | 9 | 10 |
| Very low | 1 | 2 |
The direction is consistent: students moved out of the top two categories and into the lower ones. The researchers are careful not to overclaim, and so should we.
The shifts are modest, and this was one trial over one term. What the data does show is that the design moved things in the right direction, and the same tool that improved student engagement also touched their anxiety, without those being separate interventions.
A Few Things to Keep in Mind
The researchers are transparent about the study’s scope, and it’s worth holding these alongside the results.
It was one school, one subject. All participants were grade 10 students at SMA Negeri 1 Lhoksukon studying statistics. The findings may not carry directly to every subject area, year level, or school context.
The math anxiety shifts were modest. The researchers describe the change as real but not dramatic over a single term. A few students moved categories. Long-term effects weren’t tracked.
This wasn’t a controlled comparison. The ADDIE methodology was designed to develop and validate a learning product, not to measure Edcafe against a non-Edcafe group. The study confirms the design is sound and that students responded well. It doesn’t isolate a single cause.
Our take: those limits don’t undermine the findings. They define what the findings actually are. A rigorously validated design that expert panels and classroom teachers rated highly, and that students responded to positively, is a meaningful result, even without a control group.
Three Takeaways for Your Classroom
1. Correct during practice, not only after it.
The design decision that drove the anxiety reduction wasn’t Edcafe itself: it was when the feedback arrived. Students got a correction before they moved on, not a week later on a returned test. If students only find out they got it wrong after the stakes have already been applied, the anxiety from the attempt has nowhere to go. Build correction into the work itself.
2. Anxiety often starts with the classroom design, not the student.
The researchers didn’t treat anxious students as a fixed group to manage. They treated the classroom design as the variable. Change when feedback arrives, give students low-stakes ways to attempt and fail, remove the wait, and anxiety levels shift. That’s a more useful frame than identifying which students “have” math anxiety and which don’t.
3. See who’s struggling before the test.
One of the quieter findings in the study is that the module gave teachers visibility into student progress during the learning, not only after the assessment. When all responses come back in one place, you can see who attempted the work, who skipped it, and where the class got stuck, while there’s still time to act on it.
Related reading: For a look at how Edcafe performed against traditional instruction in a controlled classroom study, see the King Fahd University study on Edcafe and student engagement.