New Mathematics Anxiety Questionnaire Developed and Validated for Primary School Students

A New Questionnaire Could Reveal Math Anxiety Before It Derails a Child’s Learning

For many children, mathematics anxiety does not begin with a high-stakes examination or an advanced algebra problem. It can appear much earlier, as a racing heartbeat when a worksheet arrives, silence when a teacher asks a question, or a sudden desire for the lesson to end. A study of 302 primary-school students has now produced and tested a questionnaire designed to detect these emotional and behavioural signals together. The Student’s Mathematics Anxiety Questionnaire, or SMAQ, is intended to give teachers and researchers a developmentally appropriate way to identify how anxiety affects children’s participation in mathematics, rather than measuring only whether they feel nervous. The instrument’s 20 items cover anxiety itself as well as children’s choices, effort, persistence and interest during mathematics learning.

Mathematics anxiety is more than ordinary dislike of a difficult subject. Researchers generally define it as tension, worry, apprehension or fear that interferes with performance in mathematical situations. Those reactions can consume mental resources that would otherwise support attention, reasoning and working memory. A child may understand how to solve a problem but struggle to retrieve basic facts or follow a sequence of steps when worry occupies part of the available cognitive capacity. Previous research has linked mathematics anxiety with poorer arithmetic fluency, reduced motivation and avoidance of challenging tasks, and has shown that its effects can emerge during the primary-school years. Once avoidance reduces practice, weaker performance may reinforce the original fear, creating a self-perpetuating cycle.

Existing tools have helped researchers study mathematics anxiety, but many were designed for older students or focus mainly on internal experiences such as fear, nervousness and negative thoughts about mathematical ability. The researchers behind the SMAQ argued that this approach can miss the visible ways anxiety appears in younger classrooms. A child who does not volunteer, waits for classmates to answer, delays starting work or gives up quickly may be communicating distress without explicitly naming it as anxiety. The new questionnaire therefore treats behaviour as part of the measurement framework. Its five proposed dimensions are anxiety, choice of action, effort, persistence and interest. Together, these dimensions are meant to show not only what children feel, but also how those feelings shape their willingness to engage with mathematics.

The questionnaire was developed through a multistep scale-design process. The researchers first defined mathematics anxiety for primary-school students as an emotional response involving tension, worry, nervousness and fear during activities such as lessons, homework, tests and problem solving. They then wrote items in simple language based on situations children were likely to recognize. Two experts in psychological assessment and mathematics education reviewed the initial 20 statements for clarity and relevance, and a pilot test with 10 elementary-school students was used to identify wording that might be confusing. The final response format used three choices—“Never,” “Sometimes” and “Always”—rather than a longer numerical scale. The simplified format was selected because younger children may find it difficult to distinguish reliably among five or more subtle response categories.

The study included students in Grades IV, V and VI, with an average age ranging from 9.88 to 11.53 years across the grades. Of the 302 participants, 132 were boys and 170 were girls; 70 students came from Grade IV, 112 from Grade V and 120 from Grade VI. The sample was recruited through convenience sampling from schools that granted permission to participate, so it does not necessarily represent all primary-school children. About 50 students completed the questionnaire during supervised classroom sessions, while 252 completed it at home under parental supervision using Google Forms. Parents or legal guardians provided consent, and the research protocol was approved by the Research Ethics Committee of Universitas Negeri Malang in Indonesia.

To test whether the questionnaire measured the intended psychological structure, the researchers used confirmatory factor analysis, a statistical method that compares observed response patterns with a pre-specified model. In this case, the model predicted that the 20 items would cluster into five related but distinguishable latent factors. Each item’s factor loading estimates how strongly it reflects its assigned construct; loadings above about 0.50 are commonly regarded as evidence that an item is a useful indicator. In the SMAQ analysis, all 20 items loaded significantly on their intended factors, with standardized loadings between 0.545 and 0.971. The effort dimension showed particularly strong loadings, while anxiety, choice, persistence and interest also met the researchers’ stated standards.

The overall statistical fit was acceptable, although not ideal by every conventional benchmark. The model produced χ²(160) = 347, with a probability value below .001, a comparative fit index of 0.896 and a Tucker–Lewis index of 0.877. Both CFI and TLI are designed to compare the proposed model with a baseline model, and values near or above 0.90 are often considered acceptable, with 0.95 representing a stronger fit. The root mean square error of approximation, or RMSEA, was 0.062, with a 90 percent confidence interval from 0.053 to 0.071. RMSEA values below 0.08 are generally interpreted as acceptable and values near 0.06 as good. The results therefore suggest that the five-factor structure is a reasonable description of the data, while also indicating that future studies should continue testing and refining the model.

The questionnaire also showed strong internal consistency, with a Cronbach’s alpha of 0.90 for the overall scale. Alpha estimates how consistently items in a scale move together, although a high value does not by itself prove that an instrument measures exactly what it claims to measure. That is why the researchers combined reliability analysis with expert review and confirmatory factor analysis. Correlations among the underlying dimensions were generally moderate to high. The article reports associations ranging from 0.43 to 0.84 in one analysis, including especially strong links between anxiety and interest, anxiety and persistence, and anxiety and choice. Because the reported correlations are positive even though the questionnaire describes anxiety-related behaviours negatively, interpretation requires care: the findings primarily show that the dimensions are closely connected, not that anxiety improves motivation or persistence. The direction may reflect item coding and the way the constructs were operationalized.

The practical appeal of the SMAQ lies in its simplicity and its focus on early identification. A teacher could use responses to notice patterns that might otherwise be mistaken for laziness, low ability or lack of preparation. A student who repeatedly reports postponing mathematics tasks, stopping when problems become difficult, remaining silent during questioning or feeling anxious before class may need support with confidence and classroom safety rather than more pressure. The questionnaire is not a diagnostic medical test, and a high score would not establish a clinical anxiety disorder. Instead, it could serve as a screening and research instrument, helping educators investigate whether instructional methods, time pressure, fear of mistakes or classroom climate are contributing to disengagement. Used alongside classroom observations and conversations with children, it could help schools evaluate interventions intended to build confidence, persistence and interest.

The study’s Indonesian setting is also important. Psychological questionnaires cannot automatically be transferred from one language, culture or education system to another simply by translating the words. Children’s interpretations of being called on, making mistakes, seeking help or responding to difficult work can depend on classroom norms, teacher–student relationships and family expectations. The SMAQ was developed for Indonesian primary-school students, with wording and response options shaped around their developmental level and educational context. That gives the instrument local relevance, but it also limits how confidently its performance can be generalized to children elsewhere. Independent studies in other countries would need to examine whether the same five dimensions hold together, whether items work similarly across age groups and cultures, and whether scores predict observable behaviour and mathematics achievement over time.

The researchers say the questionnaire fills a gap by combining emotional and behavioural dimensions of mathematics anxiety in a single 20-item measure. Its validation provides initial evidence that the instrument is reliable and that its proposed structure is statistically plausible, but it does not prove that the tool will prevent anxiety or raise mathematics scores. The sample was relatively accessible rather than randomly selected, the data were collected at one point in time, and the study relied heavily on children’s self-reports. Future research could test the SMAQ with larger and more diverse populations, compare responses with teacher observations and performance records, and assess whether scores change after classroom interventions. For now, the study reinforces a message with implications far beyond one questionnaire: when a child turns away from mathematics, the behaviour may be an emotional signal as much as an academic one.