Our leaning tower of PISA

Jennifer Buckingham

07 December 2016 | Australian Financial Review

teacher school mathsTwo sets of independent international test results released in the past week show Australia’s education system has serious deficiencies.

The results of the Trends in International Maths and Science Study (TIMSS) 2015 released last week showed no change in maths and science scores for Australian students since 1995 while other countries improved, leading to a slide in our international rankings.

The Program for International Student Assessment (PISA) 2015 results released yesterday are even worse — the performance of Australian students in reading, maths and science has significantly decreased over the past 15 years.

There has been a corresponding slide in our international rankings because other countries have either maintained their performance or improved.

While the country rankings are interesting and appeal to our competitive side, educational assessment is not the World Cup. The trend in Australia’s own performance over time is far more important than that — and unfortunately the trend is very clearly downward. The decline in reading and maths mean scores is the equivalent of one year of learning, according to the reports from the Australian Council for Educational Research (ACER).

Mean scores are useful for comparisons over time, but they don’t tell us much about student capabilities in real terms. This year, ACER has defined a proficiency standard for Australian students which is described as a “challenging but reasonable” level of achievement.

In PISA 2015, 61% of students achieved the National Proficient Standard in reading, 55% of students achieved the National Proficient Standard in maths and 61% of students achieved the National Proficient Standard in science.


This sounds bad because it is. The deterioriation in Australia’s performance is because we now have more low performing students and fewer high performing students. Students in the low performing group have achievement levels “too low to enable them to participate effectively and productively in life.”

In reading, the proportion of low performers grew from 12% in 2000 to 18% in 2015, while the proportion of high performers fell from 18% in 2000 to 11% in 2015. In maths, the proportion of high performers almost halved from 2003 to 2015, from 20% to 11%. Singapore had the highest proportion of high performers in all domains, with between 25% and 35% of students achieving in the top two proficiency levels.

Other countries — most notably the much-vaunted Finland — have also experienced significant declines in the last two rounds of PISA, showing how cautious we must be about making simplistic country comparisons to guide policy. PISA data is very useful but it must be evaluated carefully, preferably looking at the relationships between policies, practices, and performance using multiple country comparisons over multiple years.

TIMSS and PISA are slightly different in nature and purpose: TIMSS assesses curriculum content knowledge in Years 4 and 8, while PISA assesses the ability to apply this knowledge to realistic problems among 15 year olds. The concurrent implementation of TIMSS and PISA indicates that not only are students not improving their knowledge of the maths and science curriculum, they are becoming less capable at applying this knowledge.

It would be a mistake to jump to the conclusion that students need more ‘problem-based learning’. One finding that stands out in a multiple country analysis of PISA results is that teacher-directed instruction is related to higher mean scores in science, while inquiry-based learning approaches are associated with lower mean scores. Furthermore, students in countries with more teacher-directed instruction were more likely to express an expectation to work in a science-related occupation.

These findings suggest Australia’s fixation with discovery or inquiry-based learning approaches to improve achievement with maths and science by developing student ‘engagement’ ― often at the expense of learning facts and concepts ― is misguided and detrimental. Confidence in these domains is the result of competence, not the reverse.

The same is true in reading instruction. The majority of children will not learn to read through whole language-based approaches that operate on the false assumption that encouraging enjoyment of reading among children is more important than actually teaching them to read through explicit instruction in the foundational elements of written language. Children cannot enjoy reading if they cannot read.

The Productivity Commission’s report on evidence in education published earlier this year noted that research on effective educational practice is disconnected from classrooms. The TIMSS and PISA results demonstrate the consequences of this problem for Australian students.

Dr Jennifer Buckingham is a senior research fellow at the Centre for Independent Studies and director of the FIVE from FIVE reading project.

Print Friendly