SFR30/2015 ‘National curriculum assessments at key stage 2 in England, 2015 (provisional)’ was published on 27 August 2015.

This is the final year of operation for separate level 6 tests. From 2016 there will be single tests for all KS2 learners capable of taking them.

Headline percentages were unchanged from 2014, with 9% of entries awarded L6 in the KS2 maths test, 4% in the grammar, punctuation and spelling (GPS) test and 0% in the reading test.

The underlying numbers tell a slightly different story. The graph below compares the number of learners achieving L6 in each test in each year since 2012. (The GPS test was introduced in 2013.)


L6 graph 1


Following rapid increases up to last year, the number of learners succeeding at L6 has almost levelled off in both maths and GPS.

In fact, the success rate has fallen slightly in maths and risen slightly for GPS. If this year’s numbers are expressed as percentages of KS2 test entries, rounded to two decimal places:

  • 8.80% were successful in maths, compared with 8.93% in 2014
  • 3.87% were successful in GPS compared with 3.77% in 2014.

However performance in both L6 maths and GPS tests continues to far exceed performance in the reading test.

In 2012 the number of successful learners was a mere 900, but this more than doubled in 2013, only to fall almost back to its starting point in 2014. The reasons for this collapse remain unexplained.

According to these provisional figures for 2015, the number of successful learners has increased by some 56% compared with 2014, but the total of 1,460 still falls far short of the high point achieved in 2013.

The percentage of total KS2 test entries resulting in the award of L6 has increased from 0.17% to 0.25%. (Unlike previous years no information about the number of entries to the L6 tests alone has so far been released.)

The SFR tables also reveal that the gulf between test and teacher assessment results for L6 reading continues unabated, though there has been some improvement.

Whereas numbers achieving L6 in the maths test and via TA are broadly aligned for each year (apart from a relatively small disagreement in 2014), this is far from the case for reading.


L6 graph 2


In 2015 ten times as many learners were awarded L6 by TA as achieved it on the test, returning almost to the level of discrepancy evident in 2012 and 2013. (In 2014 TA success exceeded test success by a factor of 19, further highlighting the unexplained issues with the test.)

This huge difference between test and TA outcomes is atypical. It does not exist for L5 reading: in 2015 48.1% of all reading test entries were awarded L5, compared with 49.7% of all teacher assessments.

It would suggest fundamental disagreement between schools and test developers over the level of performance required to achieve L6. But what efforts have been made to rectify this, if any, and why have they proved so resoundingly unsuccessful?

As schools prepare for the new regime, what are we to conclude from this evidence about the success or otherwise of the L6 tests?

It seems to me that:

  • The maths tests have been hugely successful and the GPS tests, introduced only in 2013, have shown similar promise.
  • By contrast the reading test never reached critical mass, suffering particularly from its 2014 setback. It has been poor value for money.
  • Entries held up perhaps better than expected in 2015. There has been no major collapse and even a small upturn in reading (compared with the serious setback in 2014).
  • Most schools seemingly have little grasp of the standard required to achieve success in the L6 reading test, or are very poor at preparing learners, or quite possibly both. The recommendations of the DfE-commissioned ‘Investigation of Key Stage 2 Level 6 Tests’ (2013) have had no impact. The imminent change in the assessment regime has no doubt led all parties to consider such effort redundant this year.
  • But the capacity of the new single tests to differentiate performance effectively at the equivalent of L6 remains unproven. The sample tests published in July 2015 are not entirely reassuring. Reports of trials have not been published. It may be hard to take a further backward step in the top-end assessment of reading, given the very limited impact of the L6 test, but maths and GPS are potentially more vulnerable. It would be a real shame if the baby was ejected with the bathwater.



August 2015