Mindsets, attitudes and learning

6. Growth mindset and teacher practices

Teachers have the most impactful role in how students interpret and respond to their learning environment. Educational psychology has long stressed that the quality of teacher–student interactions – e.g. through support, relational closeness and cognitively demanding instruction – shapes students’ motivation, confidence and academic behaviours (Pianta, Hamre and Allen, 2012[29]; Wentzel, 2012[30]).

The way teachers interact with students and organise instruction is central to understanding how beliefs such as growth mindset and other learning dispositions relate to performance. Teacher practices can shape the classroom environment in which students’ mindsets operate and evolve, influencing how strongly growth mindset beliefs translate into academic outcomes. Examining how teacher practices intersect with growth mindset provides insights into the conditions under which students’ self-beliefs align with performance.

For policymakers this is a critical question. Promoting growth mindset at scale is only effective if teacher practices and classroom climates are aligned. Otherwise, there is a risk that students may endorse growth mindset beliefs but fail to see them validated in their day-to-day experiences.

Teacher support and their influence on students as moderators

Across OECD countries, PISA data indicate that teacher support moderates the performance advantage associated with a growth mindset. For instance, in mathematics, the score gap between growth- and fixed-mindset students is 14 points when teacher support is low, but increases to 19 points when teacher support is high (Table B.26). In other words, when teachers provide more guidance and help, growth mindset more effectively translates into higher performance. This suggests that supportive environments do not simply improve performance on their own, but rather strengthen the way in which students’ mindsets are linked to learning outcomes.

In addition, PISA data show that students with a growth mindset are somewhat more likely to be found in classrooms with stronger teacher support (Table B.11). Although the percentage point differences are modest and not observed in every country (five percentage points, on average), they appear consistently across many education systems. This pattern indicates that growth mindset is not evenly distributed, but tends to be more prevalent in learning environments where teacher support is perceived as being stronger.

However, these associations are far from uniform. On average, students with a growth mindset are more likely to report that their teachers take an interest in their learning, provide extra help, and persist until understanding is achieved. Although there is variation across countries/economies, the relationship is positive and statistically significant in the largest number of countries/economies, suggesting a positive symbiose in the classroom between teacher’s attitudes and students’ growth mindset. The relationship is the strongest in Australia*, Ireland*, Korea, Macao (China), New Zealand*, Chinese Taipei and the United States* (Table B.23).

Figure 11. Growth mindset and teacher practices, by level of performance

Change in the following teacher practices when students disagreed or strongly disagreed that "your intelligence is something about you that you can’t change very much", by level of performance in mathematics; OECD average

Notes: On average across the OECD, the percentage of students reporting a growth mindset is 49% among low performers, 61% among mid-performers, and 67% among top performers.

Differences are accounting for students' and schools' socio-economic profile.

The socio-economic profile is measured by the PISA index of economic, social and cultural status (ESCS).

Differences that are statistically significant are shown in a darker tone (see Reader’s Guide).

Source: OECD, PISA 2022 Database, Table B.23.

However, as shown in Figure 11, this relationship can vary by performance group. For example, the figure shows that for the OECD average, the relationship between growth mindset and teacher support is weaker for low-performing students. However, it is important to interpret these findings with caution, as this pattern is not consistent across systems. Among mid-performers, the association is positive in 30 countries/economies and not significant in most others, Among top performers, the link is often weaker or absent, while among low performers, it is frequently absent or negative across systems. These findings suggest that, rather than being a universal pattern, the relationship between teacher support and growth mindset varies widely depending on system context and student proficiency level. Only in a handful of countries, such as Australia*, Chinese Taipei, Norway and the United States*, does the association remain positive across all performance levels.

Similarly, the link between student-teacher relations and growth mindset is not systematic. For mid-performers, the relationship with growth mindset is robustly positive in 29 countries/economies and on average across OECD countries, while in over half, the relationship is not significant (Table B.23). For low performers, however, it is often negative (in 31 countries/economies, and on average), possibly reflecting that struggling students may perceive less support even when teachers provide it. Among top performers, while strong and positive for the OECD average, the relationship is mostly not significant across countries/economies. In no case is the relationship negative for mid and top performers. Only in Ireland*, Germany and Korea it remains positive across performance levels.

While student-teacher relations strengthen the positive association between growth mindset and mathematics performance (12 score points more among students with high-quality student-teacher relations) (Table B.27), this relationship is likely to vary by performance group.

Where associations between growth mindset and perceived teacher support do appear, they could be reflecting compensatory dynamics, i.e. teachers provide stronger relational support to weaker students, but these students, who may already struggle with confidence, do not necessarily develop stronger growth mindsets. In fact, dependence on teacher closeness may coexist with weaker self-belief. For mid-performers, by contrast, relational closeness can align well with growth mindset beliefs, reinforcing both motivation and confidence. For top performers, the association is largely flat, reflecting that their mindset is grounded more in prior success and self-efficacy than in their relationships with teachers.

This same dynamic presents itself in students’ perceptions of mathematics instruction. Figure 11 shows important differences by students’ proficiency level at the OECD average. However, most education systems show no statistically significant association between perceptions of instructional quality and growth mindset (Table B.23). In 24 education countries/economies, low-performing students with a growth mindset report lower perceived quality of instruction. By contrast, for mid-performing students, the association is more often positive (in 23 countries/economies), while for top performers the relationship is mostly non-significant and positive in only six countries/economies.

Students’ perceptions of teaching quality could be shaped not only by instructional practices themselves but also by how these align with students’ academic standing. Low-performing students’ belief in improvement may raise expectations; when instruction does not meet these expectations, they may judge it more critically. For mid-performers, the relationship is more often positive, reflecting that growth mindset beliefs and recognition of effective instruction reinforce each other. For top performers, the association is inconsistent, suggesting that across systems, their learning confidence could make them less dependent on perceived quality of instruction.

These patterns indicate two important insights. First, supportive relations and perceived quality of instruction can enlarge the returns of growth mindset for performance, but the direct relationship between growth mindset and perceptions of teachers is highly group-specific and often absent. Second, what may appear as negative associations for low performers could be better understood as compensatory dynamics or perception biases: struggling students with growth mindset beliefs may demand more from their teachers and, as a result, report more critically on teaching or relational support.

Growth mindset and cognitive activation: Engaging students in deeper learning

Beyond teacher support and student-teacher relations, which emphasise emotional closeness and guidance, cognitive activation represents a different dimension of teacher practice. It refers to instructional strategies that actively challenge students to think, integrate knowledge and solve problems, rather than passively receive information. These practices include requiring students to explain their reasoning, connect concepts and evaluate solutions, all of which develop higher-order thinking skills such as critical analysis and decision-making (Lipowsky et al., 2009[31]).

PISA data show that cognitive activation can moderate the relationship between growth mindset and mathematics performance. Students with a growth mindset benefit more strongly from these practices, as cognitively activating tasks give them opportunities to apply their beliefs in effort and persistence. For instance, in mathematics, the interaction between cognitive activation practices, growth mindset and PISA score shows that students with a growth mindset scored 14 points higher than students with a fixed mindset when the index of cognitive activation was low, but scored 20 points higher than students with a fixed mindset when the index of cognitive activation was high, on average across OECD countries – the 6-point difference being statistically significant (Figure 12 and Table B.28). In this sense, cognitive activation is not about reassurance or support, but about providing demanding opportunities that stretch students’ learning capacity. However, in many countries/economies, the relationship between growth mindset and performance is the same for both groups. This limited moderation likely reflects how cognitive activation is distributed: low performers in some systems have limited exposure, while in others activation is applied uniformly, with the strongest associations emerging among students experiencing moderate levels of exposure.

Figure 12. Growth mindset and performance in mathematics, by level of students' cognitive activation

Change in mathematics performance when students disagreed or strongly disagreed that "your intelligence is something about you that you can’t change very much", by the quarter of index of cognitive activation in mathematics: Foster reasoning

Notes: Only countries and economies with available data are shown.

Differences are accounting for students' and schools' socio-economic profile.

The socio-economic profile is measured by the PISA index of economic, social and cultural status (ESCS).

Statistically significant differences between bottom and top quarters are shown in a darker tone (see Reader’s Guide).

Countries and economies are ranked in descending order of the difference between bottom and top quarters (top - bottom).

Source: OECD, PISA 2022 Database, Table B.28.

Yet, it is not because teachers use cognitive activation practices that students report more growth mindsets. Among top and mid-performers, there is no significant correlation – students do not necessarily report higher growth mindset simply because teachers use cognitively activating practices overall (Table B.23). Among low performers, the association can even be negative.

Teachers may deliberately adapt the use of cognitively activating strategies when working with struggling students. Low performers may therefore experience fewer opportunities for challenge, reasoning and responding to feedback. Thus the negative association does not imply that cognitively activating practices undermine mindsets, but rather that students with weaker skills may be less exposed to these practices.

Another finding emerges when examining students’ self-reports of exposure to cognitive activation. Previous PISA 2022 analyses show that, on average across OECD countries, less than half of students frequently engage in self-reflective practices like explaining their reasoning in mathematics. Yet, there are systematic differences by performance group – about 40% of low performers report being exposed to these practices, compared to over half of skilled performers (OECD, 2024[3]). Potential reasons for this include:

Targeting: Teaching practices are adapted for performance level, with stronger students receiving more cognitively activating tasks.

Stratification: Schools may be stratified, so that low- and high-performing students attend different classes or schools with different teaching styles.

Perception: Low-performing students may be less likely to recognise cognitively activating practices, even when exposed to them, because such tasks feel unfamiliar or difficult.

The apparent contradiction between a negative or absent relationship between growth mindset and cognitive activation, and higher self-reported exposure among more skilled students, can be explained by recognising that cognitive activation is neither equally perceived nor equally effective across performance groups. For top and mid-performers, these strategies reinforce learning but do not necessarily build growth mindset, since their belief in growth mindset may already be strong. For low performers, cognitive activation can be both a challenge and a source of frustration; they may perceive these practices as harder or less supportive, leading to weaker or negative associations with growth mindset, even while these strategies still enhance their learning opportunities.

For policymakers, these findings suggest that supporting teachers to provide targeted help, for example, to use cognitively activating practices, and to sustain positive relations can create environments where growth mindset yields stronger returns in learning. At the same time, education systems should recognise that support may be perceived differently by students in different performance groups. Overall, the evidence suggests that the most effective systems are those where teacher practices are aligned with students’ beliefs and needs, reinforcing growth mindset where it can be most effective (among mid-performers), compensating for struggles without fostering dependence (among low performers), and maintaining high expectations for all.