Blackwell et al. (2007): Fixed and Growth Mindsets

Why investigate mindset in real schools?

A student’s mindset is their belief about whether intelligence can change. Dweck distinguished a fixed mindset, in which intelligence is seen as essentially unchangeable, from a growth mindset, in which intelligence can develop through learning and effort. Blackwell, Trzesniewski and Dweck (2007) investigated whether these beliefs were related to achievement during the transition to junior high school in New York City.

The researchers called fixed beliefs entity theories and growth beliefs incremental theories. These describe beliefs about intelligence, not measurements of how intelligent someone actually is.

Their proposed explanation was that beliefs influence how students approach difficulty. A student who thinks ability can develop may see a disappointing result as a reason to practise or change strategies. A student who thinks ability is fixed may interpret it as evidence that improvement is impossible. These different responses could gradually affect achievement.

The paper contained two studies. Study 1 asked whether existing beliefs predicted changes in mathematics achievement over time. Study 2 asked whether teaching a growth mindset could improve motivation and achievement beyond the benefits of a similar academic workshop.

Study 1: following students over time

Study 1 was a longitudinal correlational study. The researchers measured students’ beliefs rather than assigning them a mindset, then followed their mathematics achievement through seventh and eighth grade. Seventh grade is roughly equivalent to Year 8 in England.

The sample comprised 373 students: 198 females and 175 males, from four successive incoming seventh-grade cohorts at one New York City public secondary school. They varied in ethnicity, socioeconomic status and achievement. The project covered five years as successive cohorts entered the school, but each cohort was followed through two school years—not for five years individually.

At the beginning of seventh grade, students completed a motivational questionnaire during lesson time. Rating-scale questions measured their theory of intelligence, learning goals, beliefs about effort and helpless responses to failure. For example, the distinction between learning goals and helpless responses concerns whether students aim to develop understanding or respond to setbacks as though they cannot improve.

The researchers used sixth-grade mathematics test scores as a measure of prior achievement. They then examined school mathematics grades across four assessments: autumn and spring of seventh grade, followed by autumn and spring of eighth grade. Measuring prior achievement helped them investigate whether mindset predicted later progress rather than simply reflecting students’ starting performance.

Study 1: what changed?

Mindset was not significantly related to mathematics achievement at entry to junior high school. However, it became a significant predictor of mathematics grades as students progressed through seventh and eighth grade.

Students with stronger growth-mindset beliefs showed greater improvement over the two years than students with stronger fixed-mindset beliefs. The relationship with later grades remained when the researchers took prior mathematics achievement into account.

The important result is therefore a difference in progress over time, rather than a finding that growth-mindset students were already substantially better at mathematics before the study began. This weakens the explanation that growth beliefs were merely a consequence of earlier success.

Nevertheless, Study 1 cannot establish causation. Students’ beliefs were measured, not manipulated. Other influences, such as support at home or responses from teachers, might contribute to both mindset and progress. The researchers therefore conducted an intervention study.

Study 2: changing beliefs through teaching

Study 2 combined measurement of existing beliefs with an independent-measures experiment. It recruited 99 seventh-grade students from a different New York public secondary school. Again, students came from varied ethnic and socioeconomic backgrounds.

Students first completed the motivational questionnaire. They were then offered an eight-week workshop about the brain and study skills and randomly allocated to an intervention group or a control group. Undergraduates acted as workshop teachers and mentors.

The independent variable was the workshop condition: growth-mindset intervention or control. Outcomes included students’ beliefs, classroom motivation and mathematics achievement.

Both groups received useful teaching. The comparison tested whether the growth-mindset material added a benefit, rather than simply comparing students who received attention with students who received none.

The intervention explained that learning can develop and strengthen connections between nerve cells. Students recalled skills they had initially found difficult but improved through practice and mistakes. They also discussed how labels could discourage people from trying, for fear of looking incapable. These were messages taught in the workshop; the study did not directly measure changes in students’ brains.

How were the outcomes measured?

At the end of the workshops, students completed a multiple-choice quiz testing recall and understanding of the teaching. Some questions covered material taught to both groups; others tested the growth-mindset material. Students were told that the quiz assessed the teaching, would not count towards their grades and would not affect whether they received a certificate.

The theory-of-intelligence questionnaire was repeated three weeks after the final session to assess changes in beliefs.

The mathematics teacher also wrote about students who had shown changes in classroom motivation. These comments were coded to identify positive changes. Crucially, the teacher did not know which condition students had been assigned to, or even that there were two different conditions. This made the teacher a blind assessor.

Mathematics achievement was examined using grades from spring of sixth grade, autumn of seventh grade and spring of seventh grade. This allowed the researchers to compare achievement patterns before and after the intervention.

Study 2: findings and conclusions

The groups did not differ in recall of the general workshop content, but the intervention group performed significantly better on questions about the growth-mindset material. This suggests that the intervention’s distinctive message had been understood, rather than one group simply learning all workshop material better.

The intervention group developed more growth-oriented beliefs. More intervention students were also identified by the blind teacher as showing positive changes in motivation.

Their mathematics achievement showed a more favourable pattern after the workshop than that of the control group. The intervention counteracted a declining achievement trajectory, while decline continued in the control group.

Together, the studies supported Dweck’s theory: beliefs about intelligence predicted progress, and teaching that intelligence could develop improved motivation and achievement. Study 2 provided stronger evidence of a causal influence than Study 1 because it manipulated the teaching and used random allocation.

Evaluating the evidence

A strength is its relevance to everyday education. The research took place in real schools and used actual mathematics grades. This gives it high ecological validity compared with a task that bears little resemblance to classroom learning. Study 1 also followed development over time instead of relying on a single snapshot.

Study 2 had useful controls. Random allocation reduced the likelihood of systematic differences between conditions. Both groups received workshops, helping control for attention, mentoring and general study-skills teaching. Blind teacher assessments reduced the risk that expectations about the intervention would influence reports of motivation.

Different measures strengthened the evidence. Questionnaires assessed beliefs, teacher comments assessed classroom behaviour and grades assessed achievement. However, self-report questionnaires may be affected by students giving answers they think are desirable, particularly after being taught the intervention’s message.

Real schools also make control difficult. Family support, teaching and other experiences can influence grades. Teachers might respond more positively to students who become more engaged, so achievement could involve several connected influences rather than beliefs acting alone. The experiment tested a package of readings, activities and discussions, not a single isolated sentence about intelligence.

Generalisation is limited. Both samples came from New York public schools and focused on seventh-grade students. Their diversity is valuable, but it does not establish that the same effects occur at other ages or in other education systems. Achievement was measured in mathematics, so benefits across all subjects cannot be assumed. The short follow-up in Study 2 also leaves uncertainty about how long the intervention’s effects lasted.

Ethical protections mattered because the participants were children. Students and parents gave informed consent, and students knew they could withdraw. The control group still received useful teaching rather than being denied educational support.

What does this imply about development?

The findings support a nurture influence: educational experiences can change beliefs and behaviour, which can contribute to progress. They do not show that inherited characteristics or prior ability are irrelevant. A balanced conclusion is that mindset is one influence on development, alongside ability, teaching, opportunities and support.

The educational implication is that helping students understand improvement as possible may encourage constructive responses to difficulty. It is not evidence that effort alone guarantees success or that students are personally to blame when progress is limited.