España 
Italia 
United States 
Brasil 
Chile
Ecuador
Mexico
Colombia
Perú
Is stressing and worrying about learning and completing a math test a positive thing? The PISA 2022 results have raised alarms about students’ math performance upon completing compulsory schooling. Some support a return to more rote mechanical learning, suggesting that a certain level of stress and worry might be advantageous for students learning math.
However, as was evident in the event ‘Schools: Reflecting the Society of the Future’, organized by the Cercle d’Economia and Innovamat, it is not enough to simply reiterate knowledge, students must know how to apply it in unknown contexts. This was emphasized by Andreas Schleicher, Director of Education at the OECD and creator and coordinator of the PISA tests. At Innovamat, our analysis of the problems released from the PISA 2022 tests led us to the same conclusion.
Andreas Schleicher also stressed that students’ attitudes towards math are crucial for learning. So much so that the PISA 2022 collected this information in a questionnaire that students answered along with the math test. But… What is the connection between math performance and students’ attitudes toward math?
Let’s delve deeper, starting with the results of Spain and the OECD countries in the PISA 2022 tests!
1. PISA 2022 Tests: The Relationship Between Anxiety and Performance in Math
In addition to the math achievement test, students who participated in the PISA tests also answered questions about their attitudes toward math. For example, they were asked to what extent they agreed with statements such as: “I often worry about having difficulties in math class” or “I get very nervous when I do math problems”, statements related to a certain degree of anxiety or stress about math.
The Innovamat research team analyzed the relationship between these anxiety indicators and performance. We define ‘math anxiety’ as the emotional reaction characterized by feelings of fear, tension and apprehension in situations involving a math-related situations. Conversely, in this article, we relate math performance to the results that students achieve in math tests, specifically, the PISA tests.
Below, you can see the results of the students in the math skills test according to their degree of “worry” and “nerves”. Blue bars indicate cases in which students agreed with the statements. That is, these are the students who were most worried or nervous about math. On the other hand, green bars indicate students who did not agree.
Graph 1: Average math score on the PISA 2022 tests in OECD countries, and in Spain, according to the response to the question: “I often worry that I will struggle in math class.”
Graph 2: Average math score on the PISA 2022 tests in OECD countries, and in Spain, according to the response to the question: “I get very nervous when I do math problems.”
As you can see, the students who were more worried or nervous about math performed worse, on average, on the PISA math tests. This difference is statistically significant both in Spain and in the across OECD countries (p < 0.01, in all cases). Therefore, we can conclude that students with higher levels of math anxiety generally obtained poorer results in the PISA tests.
This phenomenon is well-documented in scientific literature. For instance, this article published by Professor Dilara Bakan in 2015 shows a similar relationship between attitudes and math´ performance in other countries, using data from PISA 2012. In England, Hong Kong, Holland, Greece, Turkey and the USA, greater self-efficacy or confidence in math correlates with better results. On the contrary, higher anxiety predicts poorer results, particularly in countries that had low overall performance in math (Greece, USA and Turkey).
2. PISA 2022 Tests: Math Anxiety and Gender Differences
The PISA 2022 results reveal significant gender differences regarding anxiety. Analyzing all the responses to the math anxiety questionnaire, we find that girls reported having much higher levels of anxiety than boys. This is true in Spain and most OECD countries and in both cases it is statistically significant (all p < .001). Furthermore, we found that Spanish students generally showed higher levels of anxiety than the OECD average.
Graph 3: Average anxiety levels in the PISA 2022 questions according to gender of students. The math anxiety index was calculated using responses to six questions about math anxiety. Positive values indicate that students responded to experiencing more anxiety towards math than the average student in OECD countries.
3. Math Anxiety and Performance in Elementary School
At this point, you may be wondering… At what age do students start to showing signs of math anxiety and how does this influence their performance?
During the 2022-2023 school year, the Innovamat research team asked ourselves exactly the same thing. To answer this question, we administered a test on motivation and anxiety in math to 291 third grade students in Catalan schools. In addition, at the end of the school year, the students also participated in a test on math content and procedures: the ConMattests, designed by Innovamat.
Measuring Math Anxiety and Motivation
Measuring Math Performance
Math performance was assessed using the ConMat tests, designed by Innovamat, as benchmark assessments. Like other benchmark tests, such as the TIMSS (Trends in International Mathematics and Science Study), ConMat classifies questions into two cognitive domains: content and processes. Regarding content, ConMat differentiates between concepts (facts, vocabulary, etc.) and procedures (operations, visualization of constructions, use of measuring instruments, reading of graphs, etc.). In the case of processes, the questions are organized into four dimensions: problem solving, reasoning and proof, connections, and communication and representation.
Relationship Between Math Anxiety, Motivation and Performance
Consistent with our hypothesis, we found that higher general motivation was linked to better math performance.
Graph 4: Math performance on ConMat tests (y-axis) according to the level of math motivation (x-axis). Each point represents one student’s result. The straight blue line and gray area represent a linear model with a 95% confidence interval. The R value in the upper left corner represents the Pearson coefficient, the correlation coefficient between motivation and the percentage of correct answers on the ConMat. A value of p < 0.05 indicates that the correlation between these two factors is significant.
Conversely, greater anxiety about math was associated with poorer results.
Graph 5: Math performance on the ConMat tests (y-axis) according to the level of math anxiety (x-axis). Each point represents one student’s result. The straight blue line and gray area represent a linear model with a 95% confidence interval. The R value in the upper right part/section/ corner represents the Pearson coefficient, which is the correlation coefficient between motivation and the percentage of correct answers on the ConMat. A value of p < 0.05 indicates that the correlation between these two factors is significant.
4. Math Anxiety, Motivation and Gender Differences
In the same study we also examined whether boys and girls exhibited different degrees of anxiety. Already in third grade, at 7 or 8 years old, girls reported higher levels of anxiety than boys.
Graph 6: Box plot showing anxiety levels of 3rd grade students according to their gender (Boy or Girl). The left panel shows students’ anxiety about math tests, while the right panel shows students’ general anxiety about math-related activities. Higher values are associated with higher levels of anxiety. The black line inside the box represents the median response (quartile 2 or 50% percentile), while the box indicates quartile 1 (25% percentile) and quartile 3 (75% percentile). The lower whiskers show students who are below quartile 1, while the upper whiskers show those who are above quartile 3. As you can see, girls not only show more anxiety in general, but the interquartile range (Q3-Q1) is also greater. That is, there is more dispersion in their responses, with some girls with very high levels of anxiety (girls’ Q3 is higher than boys’ Q3).
Math anxiety Among Teachers Influence by Gender
Research shows that math anxiety is not exclusive to students; teachers also experience it, affecting their students. For example, a 2010 scientific study from the University of Chicago led by Sian Beilock and Susan Levine, shows that female teachers’ math anxiety negatively influences girls’ math performance, but does not affect boys. The authors explain that in elementary school, where the majority of teachers are women, math anxiety can convey the feeling that women are not good at math, which can convey the belief that women are not good at math, demotivating girls and resulting in poorer performance.
This is particularly concerning and calls for reflection on how we can reverse the effect. This year, our research team is piloting a study in which we evaluate the effect of an intervention we have designed to promote a growth mindset. By fostering a growth mindset and challenging certain stereotypes about math and gender, we aim to reduce students’ math anxiety. This study is inspired by research from various researchers, such as Carol Dweck and Jo Boaler from Stanford University, who have shown that a growth mindset correlates with better math performance in adolescents. Andreas Schleicher in the event organized by the Cercle d’Economia and Innovamat, also stated that those students with a greater growth mentality—those who believe that they can learn what they do not yet know through effort—generally achieve greater academic success.
5. Conclusion
The data analyzed in this article shows that worry, anxiety and stress to not contribute positively to math performance. On the other hand, motivation is a crucial factor that can help students achieve better results. One of the objectives of the Innovamat team is to provide teachers with resources and tools, based on scientific evidence, to motivate students and help them understand that math is exciting and that anyone can learn it.
