I’ve written about mindset before (here), but a recent publication caught my eye, and I thought it was worth sharing.
Earlier this year, the OECD produced a report on its 2018 PISA assessments. This was significant because it was the first time that the OECD had attempted to measure mindsets and correlate them to academic achievements. Surveying some 600,000 15-year-old students in 78 countries and economies, it is, to date, the biggest, most global attempt to study the question. Before going any further, a caveat is in order. The main focus of PISA 2018 was on reading, so any correlations that are found between mindsets and achievement can only be interpreted in the context of gains in reading skills. This is important to bear in mind, as previous research into mindsets indicates that mindsets may have different impacts on different school subjects.
There has been much debate about how best to measure mindsets and, indeed, whether they can be measured at all. The OECD approached the question by asking students to respond to the statement ‘Your intelligence is something about you that you can’t change very much’ by choosing “strongly disagree”, “disagree”, “agree”, or “strongly agree”. Disagreeing with the statement was considered a precursor of a growth mindset, as it is more likely that someone who thinks intelligence can change will challenge him/herself to improve it. Across the sample, almost two-thirds of students showed a growth mindset, but there were big differences between countries, with students in Estonia, Denmark, and Germany being much more growth-oriented than those in Greece, Mexico or Poland (among OECD countries) and the Philippines, Panama, Indonesia or Kosovo (among the non-OECD countries). In line with previous research, students from socio-economically advantaged backgrounds presented a growth mindset more often than those from socio-economically disadvantaged backgrounds.
I have my problems with the research methodology. A 15-year-old from a wealthy country is much more likely than peers in other countries to have experienced mindset interventions in school: motivational we-can-do-it posters, workshops on neuroplasticity, biographical explorations of success stories and the like. In some places, some students have been so exposed to this kind of thing that school leaders have realised that growth mindset interventions should be much more subtle, avoiding the kind of crude, explicit proselytising that simply makes many students roll their eyes. In contexts such as these, most students now know what they are supposed to believe concerning the malleability of intelligence, irrespective of what they actually believe. Therefore, asking them, in a formal context, to respond to statements which are obviously digging at mindsets is an invitation to provide what they know is the ‘correct response’. Others, who have not been so fortunate in receiving mindset training, are less likely to know the correct answer. Therefore, the research results probably tell us as much about educational practices as they do about mindsets. There are other issues with the chosen measurement tool, discussed in the report, including acquiescent bias and the fact that the cognitive load required by the question increases the likelihood of a random response. Still, let’s move on.
The report found that growth mindsets correlated with academic achievement in some (typically wealthier) countries, but not in others. Wisely, the report cautions that the findings do not establish cause-and-effect relations. This is wise because a growth mindset may, to some extent, be the result of academic success, rather than the cause. As the report observes, students performing well may associate their success to internal characteristics of effort and perseverance, while those performing poorly may attribute it to immutable characteristics to preserve their self-esteem.
However, the report does list the ways in which a growth mindset can lead to better achievement. These include valuing school more, setting more ambitious learning goals, higher levels of self-efficacy, higher levels of motivation and lower levels of fear of failure. This is a very circular kind of logic. These attributes are the attributes of growth mindset, but are they the results of a growth mindset or simply the constituent parts of it? Incidentally, they were measured in the same way as the measurement of mindset, by asking students to respond to statements like “I find satisfaction in working as hard as I can” or “My goal is to learn as much as possible”. The questions are so loaded that we need to be very sceptical about the meaning of the results. The concluding remarks to this section of the report clearly indicate the bias of the research. The question that is asked is not “Can growth mindset lead to better results?” but “How can growth mindset lead to better results?”
Astonishingly, the research did not investigate the impact of growth mindset interventions in schools on growth mindset. Perhaps, this is too hard to do in any reliable way. After all, what counts as a growth mindset intervention? A little homily from the teacher about how we can all learn from our mistakes or some nice posters on the walls? Or a more full-blooded workshop about neural plasticity with follow-up tasks? Instead, the research investigated more general teaching practices. The results were interesting. The greatest impacts on growth mindset come when students perceive their teachers as being supportive in a safe learning environment, and when teachers adapt their teaching to the needs of the class, as opposed to simply following a fixed syllabus. The findings about teacher feedback were less clear: “Whether teacher feedback influences students’ growth mindset development or the other way around, further research is required to investigate this relationship, and why it could differ according to students’ proficiency in reading”.
The final chapter of this report does not include any references to data from the PISA 2018 exercise. Instead, it repeats, in a very selective way, previous research findings such as:
- Growth mindset interventions yield modest average treatment effects, but larger effects for specific subgroups.
- Growth-mindset interventions fare well in both scalability and cost-effectiveness dimensions.
It ignores any discussion about whether we should be bothering with growth mindsets at all. It tells us something we already know (about the importance of teacher support and adapting teaching to the needs of the class), but somehow concludes that “growth mindset interventions […] can be cost-effective ways to raise students’ outcomes on a large scale”. It is, to my mind, a classic example, of ‘research’ that is looking to prove a point, rather than critically investigate a phenomenon. In that sense, it is the very opposite of science.
OECD (2021) Sky’s the Limit: Growth Mindset, Students, and Schools in PISA. https://www.oecd.org/pisa/growth-mindset.pdf
Adaptive Learning in ELT 