Cheating & teaching: more than just an anagram

I always perceived Britain to be incorrupt; I believed incidences of corruption were aberrations.  I always saw teachers as fundamentally honest.  Recently, I’ve rethought both: political corruption seems endemic: British politicians, British people, are no more ethical than anyone else.  Likewise, teachers…

The first revelation came reading Perry Anderson.  “Europe is ill” he began, outlining “a pervasive corruption of the political class” by instancing accusations levelled at Kohl, Chirac, Schroder, Sarkozy, Ahern, Blair.  I’m just old enough to remember cash-for-questions; the clean government of New Labour replaced them with inflated expenses.  We’ve settled down: but, I realised, the next big political corruption scandal will come.  Reflecting on this, on the banks, on the Met, I reimagined Britain: everything seemed to conform to Jonathan Haidt’s argument that “People are trying harder to look right than to be right (The Righteous Mind, p. 89).”

The second revelation started as I read an article which began:

One afternoon in the spring of 2006, Damany Lewis, a math teacher at Parks Middle School, in Atlanta, unlocked the room where standardized tests were kept. It was the week before his students took the Criterion-Referenced Competency Test, which determined whether schools in Georgia had met federal standards of achievement. The tests were wrapped in cellophane and stacked in cardboard boxes. Lewis, a slim twenty-nine-year-old with dreadlocks, contemplated opening the test with scissors, but he thought his cut marks would be too obvious. Instead, he left the school, walked to the corner store, and bought a razor blade. When he returned, he slit open the cellophane and gently pulled a test book from its wrapping. Then he used a lighter to warm the razor, which he wedged under the adhesive sealing the booklet, and peeled back the tab.”

“He photocopied the math, reading, and language-arts sections—the subjects that would determine, under the No Child Left Behind guidelines, whether Parks would be classified as a “school in need of improvement” for the sixth year in a row. Unless fifty-eight per cent of students passed the math portion of the test and sixty-seven per cent passed in language arts, the state could shut down the school. Lewis put on gloves, to prevent oil from his hands from leaving a residue on the plastic, and then used his lighter to melt the edges of the cellophane together, so that it appeared as if the package had never been opened. He gave the reading and language-arts sections to two teachers he trusted and took the math section home.”

A number of features exacerbate the tale.  Lewis was “a star teacher,” the school a “sanctuary” and “safe haven” in an exceptionally deprived district.  The principal, Christopher Waller, felt pressured to ensure it could keep operating while meeting city officials’ data-led demands of Adequate Yearly Progress.  “I need those numbers,” he told teachers, “You need to teach to the test. Do what you’ve got to do.”  The school’s progress made it the subject of laudatory reports: “Even the kids know their data,” one noted.  A team of teachers helped change test answers; those who objected were brushed aside.  The school’s halo drew attention:

Parks attracted so many visitors who were eager to understand the school’s turnaround that teachers had to come up with ways to explain it. At Waller’s direction, they began maintaining what they called ‘standard-based mastery folders,’ an index of all the objectives that each student needed to grasp in order to comprehend a given lesson. Lewis, who was taking night classes at the School of Education at Clark Atlanta University, wrote his master’s thesis on the technique. It was a wonderful system,’ he said. ‘But we only put it in place to hide the fact that we were cheating.'”

Eventually, of course, it all came crashing down. As I understand it, the structure of American tests makes them more susceptible to manipulation than in Britain.  But are things here so different?

• A head teacher, 38 years in the profession, was banned from the classroom for altering SATs to improve the results of “those least likely to do as well as they should” (2014 – BBC).
• A teacher in Hull instructed a Teaching Assistant to complete one student’s exam and, on other occasions, told students the answers to their exams (alongside other misbehaviour) (2013 – NCTL).
• A head teacher in Birmingham altered a number of Year 6 mental maths papers with a colleague (2013 – NCTL).
• Wakefield City Academies Trust removed a school’s entire governing body and suspended eight teachers having broadened an investigation into online GCSE exams this summer (2014 – Academies Week).
• A head teacher was found guilty of accessing SATs papers held in a locked safe and amending them, a few years after taking the school out special measures while battling cancer (2013 – Telegraph).
• A primary school head teacher was found guilty of “extensively and systematically” altering SATs papers (2011 – Kent and Sussex Courier).
• 2013 SAT results were quashed at a school a year after it was named England’s best-performing primary (2013 – Telegraph).

Seven anecdotes prove little; what is the national picture?  Increasing numbers of primary schools have had SAT results quashed.*  In 2012 Ofqual reported the highest number of penalties for malpractice in exams ever; this rose by two thirds in 2013 (2014 figures have not yet been released).  The total numbers are low though: only five primary schools had all their SATs results quashed in 2012; just 135 penalties were handed to schools by Ofqual in 2013 and only one school was stripped of the right to hold exams.  I suspect this only scratches the surface (a belief shared by at least one invigilator).

The depth of the issue shows more clearly on national datasets, like this graph of the results of the Phonics test in 2012:

This suggests that nationwide, thousands of teachers have interpreted student responses just charitably enough to ensure they reach the desired level.  This pattern repeated itself in 2013; this year, the pass mark was withheld; (in the pilot, in which the pass mark was unknown, no such spike occurred).

I wonder if this reflects more accurately the pattern of cheating which is taking place.  Seriously unethical behaviour tends not to be a dramatic choice, but the result of a number of small, opportunistic decisions.

The clearest example, and perhaps the worst culprit, is surely controlled assessment (itself introduced to replace discredited coursework).  I suspect almost every secondary teacher has heard of ways to aid students which exceed the rules but are untraceable: teachers writing comments on students’ drafts on separate pieces of paper, which students can hold alongside their own work, is one favourite.  Or see this thread on ICT; similar concerns can be raised in almost every subject.

Most teachers are unlikely to consider a little additional help as cheating: rather, they are giving students the chance to do their best.  The cumulative effect is toxic however.  A London school has been cited to me as an example of great practice more than once in the last year: “look at their results.”  Yet I’d heard as gospel that those results relied on cheating in borderline students’ English controlled assessments.  What matters, more than this one school’s actions, is the culture it creates: if (as many teachers believe) the rival school down the road is practically writing students’ controlled assessments for them, what harm is the little extra help I’m giving my group?

Let me be clear: I’m not accusing every teacher of cheating.  (For the record, I’m not accusing anyone of anything).  I’m arguing that a culture has developed in which a lot of shady things are going on, or are perceived to be going on, and that it’s to the detriment of all (notably teachers, who usually derive little personal benefit from the results).  At every level, teachers and leaders are under immense pressure to get the grades needed: the situation described in the Atlanta school is akin to that in many schools which ‘Require Improvement’ or just want to keep their current Ofsted grade.  Teachers disdain much of the data which dominates their work: the incentive to cheat is, perhaps, clear.

What is to be done?

Let me start with two obvious solutions which I don’t believe would help.  It would be naive to advocate getting rid of tests or limiting their importance; they’re not going away (although point 4 below could lessen their impact).  Likewise, increasing the number of random inspections of schools solves nothing: it’s like relying on a few more traffic police to deal with endemic bad driving on dangerous roads.  In both cases, we need to redesign the system entirely.  We might:

1) Measure students’ long-term success.  As I’ve argued previously, schools should be developing students in ways which aren’t obvious at sixteen.  Assessing where students are at 25 may be too slow for the political or leadership cycle, but it would tell us whether education really worked: are students in work, in prison, in loving relationships, contributing to society, happy?**  Such assessments are much harder to game.

2) Make public exams public.  Hold exams in centralised, independent sites, not schools.  This would allow economies of scale on logistics and security (and therefore on verification of these procedures by Ofqual).  Pressure would be reduced on schools entering a handful of students for exams.  PE departments could have their gyms back for the summer term.  If not, at least have students swap schools.  (This may sound implausible, but in India, students must sit exams at another school.  Why not here?)  It would make the suspicions and problems described here a thing of the past.

3) Avoid linking pay to student outcomes.  In the past, teachers were indirectly and weakly incentivised to cheat: student success meant more autonomy (perhaps) and kudos, less time explaining yourself to the head.  Performance-related pay tied to student outcomes, as it is in some schools, is one of the clearest examples of moral hazard one could hope to find.

4) Separate purposes deserve separate assessments.  The least original point, and the most important.  Using one exam to assess national educational standards, teacher quality, school effectiveness, student achievement and student aptitude for further education or employment is neither reliable nor valid.  These separate purposes deserve separate tests: most teachers would then be motivated only to manipulate those relating to her own quality or the school’s effectiveness – not ideal, but a start.

*  The most recent Standards and Teaching Agency report dates from 2012; it’s not entirely clear what’s happened since.

** Tim Leunig at the DfE mentioned that it will soon be possible to link National Pupil Database numbers with HMRC tax revenues.  Earning power is no way to measure the value of a person, but it will provide an interesting reflection on how our former students are doing.