Witness statements are notoriously unreliable: through the fog of bias, perception, relationships and memory, the world is viewed in a highly subjective way.
And herein lies the problem with lesson observations. The heart of performance management in education is essentially a witness statement. Unsurprisingly, this cornerstone of possible pay progression is widely criticised as a result.
Last month in TESS, observations were identified as having an "enormous and unjustifiable" negative impact on morale ("Original thinkers? We don't want your sort", 14 March). Thinktank Policy Exchange has also waded in, commenting that lesson observations by Ofsted are no better than "flipping a coin".
The concern is not new, of course. Comments from worried professionals have been increasingly evident in recent years, with many highlighting the climate of fear and intimidation created by observations. "The more we were told we were useless, the harder it became," one teacher said.
These sentiments were echoed by an experienced and highly regarded professional who told me how welcome it would be during observation feedback to be told that they were doing "just something, anything, good".
Problems arise because lesson observations are not just witness statements but interpersonal assessments. It is no surprise, then, that they are fraught with subjectivity and vulnerable to bias.
Other performance indicators can be very precise and therefore easy to apply. One example might be the application of a new whole-school strategy: unless you adopt it, your wage will not increase. Meeting targets for student progress is another common determiner. Setting aside the philosophical and ideological debates around the issue as a whole, these examples are clear and easily administered. Lesson observations need to be equally straightforward if they are to be trusted.
So how can it be done? Psychologists Ian Rothman and Cary Cooper state in their book Organizational and Work Psychology that most individuals want to learn about themselves. The authors say that performance management can be highly motivational if used well, but harmful if not. It would be useful for us to follow their advice and have a lesson observation system that is:
Practical, so it is easy to administer.
Valid, so it measures only what it is intended to measure.
Standardised, so it is administered by everyone in the same way.
Reliable, so judgements are seen to be accurate.
Objective, to ensure that procedures are clear and address the points above.
These concepts need to be part of an agreed observation policy that is universally understood and accepted. Any teacher being observed should insist on these factors being explicitly adhered to so the process can be seen to be equitable. This can be enforced by a checklist for the observed, to be handed back as an evaluation of the observer's performance.
Although useful, this may not overcome one of the primary problems of observation: bias. John Arnold's book Work Psychology: understanding human behaviour in the workplace identifies subjectivity and bias as natural cognitive shortcuts, used by our brains "in an effort to cope with a complex world". It is bias that undermines many observations and prompts complaints from those being observed. Bias gets in the way of a fair assessment.
Many believe that subjectivity is something we simply have to live with but I would argue otherwise. If we make ourselves aware of the various types of bias, we can begin to take steps to combat them.
Whether you are observing or being observed, take note of the points below. Use them to spark discussion in your school about the best way to tackle bias, so that observations are the source of positive improvement and reward, not negative, unfair and unproductive attacks.
The trick of timing
Primacy and recency bias - the tendency to base judgements on either the earliest or the most recently occurring events - comes about when colleagues are observed at different points in the performance management cycle.
It is easy to react with increasing negativity when we see a mistake being repeated by several colleagues. For example, the first time we come across the overuse of mini-plenaries, we might find it understandable (perhaps it was a previous target: use mini-plenaries to gauge the learning progress of specific groups). But by the time we've seen it in the fifth teacher's lesson, we have had enough. This can make us more critical of the last colleague.
The opposite can occur when a whole-school issue is initially mistaken for a weakness in an individual teacher. In this case, will the first person be dealt with more harshly than the last?
We can guard against such imbalances by insisting that observers adhere to standardisation and that the observed are aware of their position in the cycle and are ready to produce evidence to challenge this type of bias. Good leaders will encourage this dialogue and provide a protocol that is based around peer support during face-to-face meetings.
The "halo and horns" effect is based on initial judgements, where expectations are formed by early encounters. For example, a very good first lesson observation outcome can provide a lasting halo effect, while a poor grading can be difficult to shake off. The negative bias can ultimately make some observers less sympathetic. I have also seen a trainee attract the "horns" label owing to his attitude in the staffroom. The grading of his teaching suffered.
It is therefore crucial that:
Observers dismiss irrelevant thoughts and ensure that they are making judgements against a clear set of agreed criteria. Each observation should start with a clean sheet.
Those being observed talk to colleagues to ensure that there is consistency. These conversations can be informal, but must follow procedures and offer evidence if a grievance is to be lodged.
The similarity effect occurs when you recognise your own traits in a colleague you are observing. This bias can be seen as favouritism.
As an observer, you should not be easier on those who make the mistakes that you once made or could see yourself making. We tend to like people who are similar to us, so we must be aware of positive prejudices. Observers must not be seen to be favouring colleagues they get along with.
As such, it is necessary that:
Observers ignore personality issues.
The observed check if the observers' favourites are being dealt with more sympathetically. If so, they should gather evidence and challenge unfairness.
Stereotyping happens when we simplify the characteristics of a group and attribute them to all of its members. Stereotyping has perhaps the great impact and is more common than we might imagine. John Arnold suggests that "everyone relies on stereotyping, although they might not be aware of it". It can be the toughest bias to combat.
Validity is the key here: observers must be measuring what they have been asked to measure. Comments such as "He's a bloke, so what do you expect?" or "She's an art specialist, so I'm expecting..." have no place in a fair and robust observation framework.
To combat this tendency:
Observers must dismiss from their minds any expectations (positive or negative) gained from previous observations of those from a similar group.
The observed need to talk to people from a similar background and challenge stereotyping if it is apparent. It is easy in this area to trade in mere anecdote but it is important to find evidence, or it could be seen as bias against the observer.
When being observed, challenge the observation system. When observing, challenge yourself. Hopefully these points will help. We do not deserve to be judged by a system that is no better than the "flipping of a coin".
Bill Lowe is a former headteacher and now a senior education lecturer at Newman University, Birmingham
From keeping an open mind to accentuating the positive, this expert guide will help you to give effective feedback.