UK Statistics Watchdog pans DfE again

Janet Downs's picture

The Department of Education has been ticked off  by the UK Statistics Authority again.  This time it was for a tweet (now withdrawn) which said:

“70% of white working class boys from grammars go to uni vs 54% from comprehensives. What do you think about grammars”

The watchdog said the tweet raised two issues.  First, it isn’t possible to discover whether pupils are ‘working class’.  The data used referred to white male pupils eligible for free school meals (FSM) in Year 11.  It’s wrong to assume such pupils are working class.

Second, the watchdog said comparing selective schools with non-selective ones was not comparing like with like.   Grammars choose only high-ability children whether FSM or not.  Comprehensive schools educate the whole of the ability range.  This distinction is lost on ministers, the DfE and others who compare results from selective schools, grammars or private ones, with results from schools who take all children.

It’s not the first time the statistics watchdog has criticised the DfE.  It expressed concern about the DfE’s use of PISA international test data for the UK in 2000.   It said a link to ministerial comment from official statistics was not impartial and didn’t meet professional standards.  It found data published in the 2014/15 Academies Report were not official statistics.  It criticised former Education Secretary Nicky Morgan (twice) for misleading remarks about Key Stage 2 results.

The watchdog’s warning about misusing 2000 PISA figures didn’t stop Morgan from repeating them to the CBI last year.   When the watchdog intervened the DfE said ‘quality assurance’ set up after the earlier warning had failed.  Quality assurance failed again in August this year when Nick Gibb used the same flawed data in the Commons.  The excuse this time was that he had used the figures ‘inadvertently’.

In its latest criticism of the DfE, the watchdog said the DfE’s statistical head, Iain Bell, has been ‘working closely’ with the DfE’s press office to make sure all data is checked before publication.  But the DfE tweeter hadn’t consulted the DfE stats team before dashing off the inaccurate tweet.  Another quality assurance failure.

The watchdog has asked the DfE for an ‘update’ to avoid future publication of ‘such misleading communications’.  But no amount of quality assurance will prevent deceptive information being broadcast if the user doesn’t bother running it by the DfE stats team first.  Or, worse, if checking removes the dodgy data but it’s still used ‘inadvertently’ afterwards.




Share on Twitter Share on Facebook

Be notified by email of each new post.


Helen Farrell's picture
Sat, 12/11/2016 - 20:49

Weren't there assurances that the consultation was "the beginning" of an open conversation about selection? Now the D of E is posting deliberately misleading statistics in order to elicit a favourable response. This, allied to the biased questions surely discredits the whole exercise. Although I suspect the fact that the D of E has resorted to such methods (did they really think no-one would notice the false stats?) means the consultation is not going their way at the moment.

Janet Downs's picture
Sun, 13/11/2016 - 08:12

Helen - unfortunately the DfE's idea of 'truth' is data used in a misleading way to support government policies and ministers' prejudices.   The most far-reaching of these was the use of faulty statistics from the PISA international education tests for the UK in the year 2000.  Michael Gove ignored the warning from the OECD (the organisations that administers PISA) NOT to use this data for comparison.  But the comparison was used in a DfE press prelease which was churned by most of the media under such headlines as 'Travesty of our stagnating schools' (The Mail).   Nobody in the media (apart from FullFact) bothered to check and kept on churning the false data until the UK Stats Watchdog expressed concern about DfE use of PISA data.  But this was a couple of years after the event.  And the flawed stats are STILL being used by ministers (albeit 'inadvertently', it is claimed.


Add new comment

Already a member? Click here to log in before you comment. Or register with us.