Growth Mindset

The terrible truth about reading rates in Australia (and how to fix them)

One in five of all our students fail to achieve minimum levels of reading or maths. That’s shocking. What’s even more shocking is that if you look at the pool of disadvantaged students, that figure skyrockets to one in three, compared to one in ten among advantaged students.

But some disadvantaged students beat the odds and succeed – and that’s we call academic resilience.

These figures are straight from the Programme for International Student Assessment (PISA).  The good news is we can fix this but there is a long way to go. One key difference between the resilient and non-resilient is the growth mindset,  a belief that one’s ability can increase over time, that intelligence is not fixed but changeable. Encouraging students to believe that will be a key driver in any changes and any improvement. 

In addition to scoring lower on tests such as PISA, research has shown that students from socioeconomically disadvantaged backgrounds have poorer educational outcomes than their more affluent peers on a range of measures, including school completion

Yet, despite this association between socioeconomic disadvantage and poorer educational outcomes, a small number of students who come from disadvantaged backgrounds do excel at school. Just over 13 per cent of socioeconomically disadvantaged Australian students, and 11 per cent of students across the OECD on average, show what PISA terms ‘academically resilience’ by scoring in the highest quartile of reading literacy performance. 

This apparent success despite the odds is the focus of a new ACER report that examines what, if any, characteristics these academically resilient students share, why this might be, and what we can learn from this small group that might assist in improving outcomes for all students.

Enjoyment of reading helps

Learning to read is a challenging task that requires persistence and motivation. It has been suggested that enjoyment of reading and motivation to master tasks may be two manifestations of academic resilience. Similarly, Dweck suggests goal-oriented students tend to be academically resilient and exhibit higher levels of confidence than others, and they are likely to seek challenges and be persistent. 

In Australia, and across the OECD on average, academically resilient students tended to enjoy reading more, were willing to work hard to master tasks, and indicated more of an inclination to set and pursue goals than did non-academically resilient students.  

Figure 1.  Differences between resilient and non-resilient students in attitudes and dispositions

While these findings may be as one would expect, there is interesting variation across other participating countries in students’ enjoyment of reading. In Japan and the participating Chinese jurisdictions, both academically resilient and non-academically resilient students scored at or above the OECD average on the enjoyment of reading index. In Denmark, Norway and Sweden neither group appeared to really enjoy reading, although non-resilient students enjoyed it less than resilient students. 

Figure 2. Average scores on Enjoyment of reading index, resilient and non-resilient students, Australia and comparison countries

Gender differences at play

Given that females outperform males in PISA reading literacy in every country, gender is likely to be a major factor in whether a student is academically resilient. Interestingly, while a larger proportion of Australian female students than male students were academically resilient, there was no statistically significant difference between their reading literacy scores. This suggests the resilient males are even more resilient than their female peers.

Table 1. Mean scores PISA 2018 Reading Literacy, by gender

Australian femalesMean (Standard Error)Australian malesMean (Standard Error)
Whole cohort 519 (2.0)487 (2.2)
Non-resilient students373 (4.1)352 (3.5)
Resilient students613 (6.0)617 (7.3)

Whole school influences

Prior research has found that the average socioeconomic profile of a students’ school is strongly associated with their performance on PISA.

While a substantial proportion of academically resilient Australian students attend schools in the lowest socioeconomic group, far more resilient students than non-resilient students attended schools in the highest two quarters of aggregated socioeconomic background, suggesting that attending schools with more advantaged peers may play a role in a student’s chance of being academically resilient. 

Table 2. Distribution of resilient and non-resilient students by school socioeconomic background

School aggregated Socioeconomic backgroundResilient students (%)Non-resilient students (%)
Lowest quarter3957
Second quarter3028
Third quarter2012
Highest quarter114

There are a number of reasons attending schools with more advantaged peers may play a role in a student’s chance of being academically resilient. It may be the influence of peers on students’ motivation for learning, or because the more advantaged schools themselves have better access to resources than less disadvantaged schools, that students attending more advantaged schools receive stronger support from parents or teachers, or perhaps that they were selected to attend these schools on scholarship.

The importance of a growth mindset 

Research has shown that holding a ‘growth mindset’ – a belief that your ability can increase over time – is linked to better academic achievement and can even temper the effects of socioeconomic disadvantage.

Data from PISA supports this theory, as Australia’s academically resilient students were more likely than non-academically resilient students to hold a growth mindset. Eighty per cent of academically resilient students disagreed with the statement ‘your intelligence is something about you that you can’t change very much’, compared to just 41 per cent of non-resilient students and 70 per cent of Australian students on average.

Growth mindset, along with enjoyment of reading, motivation and goal setting, stand out against gender and school profile as areas related to academic resilience that can be readily targeted by education systems to help address socioeconomic disadvantage. By directly addressing these issues in the classroom, we may be able to improve outcomes for the 87 per cent of Australian students who have not overcome their disadvantage.

Dr Sue Thomson is Deputy CEO (Research) at the Australian Council for Educational Research, and the National Manager of the PISA project.

Cover Image by Bruce Matsunaga / Flickr

Personal Best: how setting PB goals can significantly improve student performance

Setting a personal best goal within every-day classroom activities can improve student performance and, significantly, it can be achieved through a very simple change to an existing classroom exercise.

Results of our latest research study on setting personal best goals, by our team from the University of New South Wales and University of Sydney, may have a broad application across all classrooms.

We used classroom-based problem-solving exercises and made just one simple change to instructions given to students. While we used a mathematic exercise in this instance, we believe any teacher of any subject could use this technique in any classroom.

What does setting a personal best goal involve?

Definition of personal best (PB) goals

Scientia Professor and Professor of Educational Psychology at the University of New South Wales, Andrew Martin, has defined personal best (PB) goals as specific, challenging, and competitively self-referenced goals that involve a level of performance that meets or exceeds an individual’s previous best.

Setting goals and our previous studies

Most of the available research on personal best goals has been survey-based. For example, in an earlier study we surveyed 249 high school students twice across a year about their use of PB goals using responses to PB questions (e.g., “When I do my schoolwork I try to do it better than I’ve done before”). We found that the relationships between PB goals and deep learning, academic flow, and positive teacher relationship remained significant beyond their previous scores on these factors. We believe these results support the importance of PB goals in school settings.

More recent investigations have tested the effects of PB goal setting using experimental research methods. In the context of an annual mathematics contest, Professor Martin and professor of psychology at the University of Rochester, Andrew Elliot, invited some students to set a PB goal based on their previous year’s performance, while others were simply reminded of their previous year’s performance. Students in the PB goal condition substantially outperformed the control condition.

Experimental investigations such as this provide strong tests of causal hypotheses (where you can predict what will happen when you change one thing in a relationship). In this case, we could predict that PB goals act to enhance learning processes and/or outcomes for students.

Our latest study

Our latest study aimed to test whether PB goals can enhance mathematics performance in a primary school context, using an existing class activity, SuperSpeed Math, developed by Chris Biffle of Whole Brain Teaching.

In this study, we aimed to boost short-term improvement in arithmetical problem-solving fluency, across a range of arithmetical skills, under time pressure to build number fluency.

Describing how the design of SuperSpeed Math incorporates PB goals, Biffle argues:

Students love to play SuperSpeed Math because they love to strive for goals and to set and break personal records. Players are never competing against each other, but against their own previous best effort. Thus, the learning objective is set at exactly the right level, no matter a player’s ability.

What we did

Our study involved 68 children across Years 5 (35 students) and Year 6 (33 students) who chose to participate. A research assistant worked with each child for approximately 35-40 minutes.

Students began by completing a short paper-based questionnaire to gauge

(a) their prior ability in mathematics,

(b) their use of and the classroom emphasis on Personal Best, Mastery, and Performance (competitive) goals in learning mathematics;

(c) their mathematics self-concept; and

(e) students’ valuing of and interest in mathematics.

Students were then randomly assigned to the Personal Best condition (compete against your own previous Personal Best score on mental maths questions), and a No Goal comparison condition. In the first round of a set of mental maths, students would follow these instructions:

“We’re now going to do some mental maths questions. This is how we’ll work together.

I’m going to give you some sheets of paper with mental maths questions on them. For each sheet of questions, your task is to answer as many questions as you can in 60 seconds.

I’ll give you some feedback on how many you got correct, then we’ll go through that sheet of questions again. Please start with the questions on the top row, going from left to right, then the second row from left to right, and so on.”

For the PB condition, after each initial 1-minute attempt at a set of initial test questions (e.g., on addition), the researcher gave the following instructions:

“OK, your score on these addition questions was [X]. This is your Personal Best score. Now we’re going to do these questions again, and I would like you to set a goal where you aim to do better on these questions than you did before.”

For the students in the comparison condition, the only difference to the experimental (PB) condition was that students in the condition were informed,

“OK, your score on these [addition] questions was [X]. Now we’re going to do these questions again”.

All students completed 10 sets of two rounds of arithmetic problem-solving, that is 20 rounds in total.


As expected, there were no statistically reliable differences between the two groups on the “pre-experiment” prior mathematics ability test or self-reports on Personal Best, Mastery, and Performance goals.

Our primary research question focused on the number of mental mathematics questions solved by the two groups. We found a small but statistically reliable difference between the two groups in favour of the Personal Best goals group.

Our study shows that setting a Personal Best goal during classroom-based problem-solving exercises improves performance.

What does this mean for teaching and learning?

We believe these results may have broad application in classrooms. Although the effect of setting a PB goal was relatively small, it was achieved through a very simple change to an existing classroom exercise.

John Hattie has argued small effects may still be practically important when considered in context. In our study, we aimed to boost short-term improvement in arithmetical problem-solving fluency, across a range of arithmetical skills, under time pressure to build number fluency.

A large change to such a complex skill-set probably won’t happen through a ‘one-shot’ intervention. Instead, students will engage in such fluency-building activities many times over the course of weeks or even months. We believe the small effect found in our study can be the start of a “virtuous cycle” of enhanced learning and performance. Beyond exercises like SuperSpeed Math, PB goals can potentially play a number of roles in students’ goal-setting and school reporting.

Our study focused on Personal Bests in mental maths, but there are many opportunities to set and strive for Personal Bests in other subjects and skill-sets. For example, students may strive to spell more words correctly in a forthcoming spelling test than they had in the previous test; they may aim to read an extra book or resource for an assignment than they had done before; before a test they may do some revision over the weekend when they previously had done none; or, they may aim to ask a teacher for help when they had previously been reluctant to do so. In all such cases, students are their own benchmark and their own point of reference for self-improvement.

More broadly, these findings are consistent with a “growth mindset”, which holds that every child is capable of learning given the right opportunities, access to productive strategies, effort, and support.


Here is our paper in the Australian Education Researcher Personal Best (PB) goal-setting enhances arithmetical problem-solving 

If you would like to know more about our Growth Mindset work please visit the Growth Mindset site



Paul Ginns is Associate Professor of Educational Psychology in the School of Education and Social Work at the University of Sydney. Paul uses numerous research methodologies (for example, experimental and survey-based research) and analytic methods, including general linear models, exploratory and confirmatory factor analysis, structural modelling and meta-analysis, to investigate student learning.




Professor Andrew Martin, PhD, is Scientia Professor, Professor of Educational Psychology, and Co-Chair of the Educational Psychology Research Group in the School of Education at the University of New South Wales, Australia. He specialises in student motivation, engagement, achievement, and quantitative research methods.