Why the A Levels fiasco is wrong - and betrays students

.

Liberal Democrats will stand up to abuse of power - and of statistics. The use of 'statistics' to adjust the grades of students shows how little the people in power understand numbers, and put their own prejudices above real people.

You cannot use accumulated data about a whole group to make accurate predictions about a single individual within it. You can use statistics to predict what could happen to the average of the group - but that doesn't deal with individual people.

And any algorithm which systematically downgrades results of some groups more than others, and fails to reflect exceptional performers, is dangerous and unfair. It is against everything Liberal Democrats represent. We stand for the rights of every individual to achieve their full potential, not to be judged on their school or the area they live in.

If the effort which has been put into defending a fundamentally flawed results guesstimation system had been put, instead, into online exams, the current mess could have been avoided.

The algorithm seems to have applied teachers' ranking order for students, and then re-applied the results from the previous 1-3 years, to the current cohort based on that ranking. That is a statistical fallacy; there is no reason why the current cohort in, say, a German class would achieve exactly the same results as last year. Weighting was then used for class size, boosting public schools, another statistical fallacy when applied at 'local' level.

Next time you go for a dental check, make sure it's done by a person, not an algorithm. Otherwise you may get a filling in your lower right molar because several people in your road had one last year!

Tony Clayton is a former Divisional Director at Office for National Statistics