In Part 1 of this series I reviewed some of the evidence on what makes for effective interviews, and in Part 2 I looked specifically at creating a less biased, more structured formal interview. In this post I’m going to lay out my thoughts on the usefulness of the interview lesson.

One of the peculiarities of teaching is that teaching a sample lesson has become a ubiquitous part of the interview process. The received wisdom is that we can work out a lot of what we want to know about a prospective employee’s teaching ability by watching them teach a class of children whom they have never met before. Even though most people now accept the impossibility of reliably grading lessons, we still believe that we can find out a lot about a teacher’s personality: their warmth, their ability to maintain order and their ability to judge what will interest and motivate children. This belief is based on a powerful intuition about how good we are at evaluating a teacher’s performance and drawing meaningful inferences. This intuition may well be wrong.

It may interest some younger teachers to know that it wasn’t always thus. When I embarked on my teaching career there was no expectation that candidates would have to teach a lesson. I’m not sure precisely how or when the practice became widespread but certainly I got my first job without anyone at the school having seen me teach. By the time I went for my second job, a few years later, it seemed to be firmly embedded; as natural as mothering and apple pie.

Before I start in on the criticisms of interview lessons, I want to be clear that they’re not useless and I probably wouldn’t advocate getting rid of them entirely. I think you can get some sense of how a teacher might perform when teaching their own classes, and some of this information might even have some limited use. But – and this is a big but – I think the way most interview lessons are typically run produces far more noise than signal. As such I’m pretty sure that observing potential employees teach leads to some excellent teachers not being appointed and some less good teachers being preferred. 

The first problem is this: everyone knows what they like. There’s a very natural human tendency to prefer people who are similar to us. This is the familiarity effect. When we see another teacher go about their business it’s almost impossible not to think things like, “I wouldn’t have done that,” or, “Yes, that’s exactly the sort of thing I do.” These judgements are automatic and, often, unconscious, but they influence our opinions in powerful ways. If we consider a teacher to be like us, then we’re predisposed to look favourably and charitably on what happens in their lesson. If something doesn’t quite go right we wince sympathetically, and when things turn out well, we’re delighted on their behalf. Conversely, when we see another teacher as being cut from a different cloth, our biases might means we don’t even recognise their successes – everything they do is coloured negatively. This familiarity effect also works on aspects of other people we’d feel ashamed to acknowledge. A white observer is more likely to identify with a white teacher, and female observers are more likely to prefer female teachers. We’re also programmed to form judgements on irrelevancies such as how attractive we find the person. Physical appearance creates a halo effect where we are predisposed to be impressed by those who are tall, slim, well-dressed and attractive. Likewise, we have negative stereotypes at play where if we’re watching a teacher we consider to be, say, overweight, we might make unconscious judgements about what they’re like in other aspects of their lives. I once worked for head teacher who said, after meeting all the candidates for a position, that he didn’t mind who I appointed as long as it wasn’t the fat one! These biases are probably inevitable but knowing about them means we can design process to limit their effect.

Other problems with observing interview lessons include the following:

  • Learning is invisible – all we can see are poor proxies for learning. That said, as long we acknowledge this then it should be too much of problem; we can still learn something by observing proxies.
  • Interview lessons are a one off show piece. Being able to plan and teach a stand alone lesson tells us very little about day-to-day practice. Teaching is a marathon rather than a sprint and the showboater doesn’t necessarily make for an effective teacher.
  • ‘Good practice’ is often trendy practice. We’re often victims of the cult of the new and are disproportionately impressed by fashions and fads. Interview lesson are often hugely over prepared with a staggering array of bells and whistles. Ask yourself, is this sustainable? Is it what you actually want teachers doing day in, day out?
  • Context matters. Just because the lesson you watched didn’t work in this case, doesn’t mean it wouldn’t work in another. In fact maybe the students are used to the ways of a particular teacher and just weren’t in the right mindset to meet a new teacher’s expectations. The only way a teacher can make a lesson work with unfamiliar kids is to pitch it very low.
  • Observed lessons are high stakes. Some people are blessed with the sort of personality that responds well to pressure, others aren’t. Some of the very best teachers I’ve known don’t always perform well when watched. Add high stakes to a highly unnatural scenario and you can expect misfires, surprises and bad luck.
  • The role of luck. Sometimes what we planned works because circumstances beyond our control fell into alignment. Like most teachers, I’ve taught a lesson which works brilliant with one set of pupils only for it to fall flat with another.

Perhaps the biggest problem is that may interviewers believe they are really good at observing. They think they can spot things no one else can. This is not supported by the evidence. Part of the MET project was to show videos of teachers to experienced professionals and ask them to identify which ones were responsible for great results and which ones were failing. The observers got it right less than 50% of the time. Their ability to spot a good teacher was less than that of simply tossing a coin. But, unless faced with such unambiguous proof, we tend to maintain a belief in our infallibility. I often hear school leaders boasting of their ability to spot good teachers from interview lessons and, when challenged, they point to their track record saying, “Look at all these fantastic appointments I’ve made.” This is the survivorship bias – we only take data from the teachers we appointed, not the ones we didn’t. There’s research in a wide range of fields from radiology, to wine tasting, to clinical psychology all showing that expert intuitive judgement does not develop where feedback is biased. feedback is automatically biased when it comes to judging interview lessons; we never find out if our intuitions are accurate because we never check what became of the teachers we turned down. Instead, we develop a sunk cost belief in our intuitive expertise and fall victim to confirmation bias.

So, what can we do to guard against these problems? Here are some ideas you might find helpful:

  1. Make sure the final decision maker does not observe candidates. Better to ask someone who will not be involved in later decisions. Of course, this person will be just as probe to bias as anyone else, but they will be less able to influence later decisions because of a strong emotional reaction produced by the observation.
  2. Be clear on what you’re looking for. Looking for evidence of learning or progress is ridiculous in a 20 minute lesson so instead focus the observer on characteristic you can get some useful information about: warmth, reactions to disruption, ability to explain clearly etc.
  3. Once you decided on what you want to see, get the observer to score candidates on a 1-5 scale (as suggested with interview questions). While I’m clear that you cannot reliably grade lessons, you can grade proxies. Scoring will still be subject to bias but it should lessen its effects.
  4. Don’t make teachers guess what you want them to do. If you’re a trendy progressive and want to see lots of group work, say so. If you’re more traditionally minded and want to see children working in silence, be clear about it. You should be absolutely explicit in advance of what you’re looking for (see point 2) and, ideally, you should give all teachers the exact same lesson plan and watch how they interpret it.
  5. Preparing for an interview is stressful enough without having to plan something your observer might end up disliking. If you think it’s important to see how teachers plan, give them a separate planning task.
  6. Don’t give teachers feedback on their lesson. This is not only arrogantly presumptuous, it’s an utter waste of everyone’s time. Your feelings about another teacher’s teaching are just that: your feelings. The only time your feedback could ever be useful is if you were planning to watch them give another interview lesson. Much better to allow candidates to tell you what they noticed about the lesson. If it was a shambles and they can explain why and talk through what they’d do differently this is worth knowing. It also allows you to see if they’re honest, thoughtful and willing to improve.
  7. Consider doing without an interview lesson. What could you do instead? What information would be missing that you couldn’t establish through other means?

These are suggestions based on the research into limiting unconscious biases. If you’re sceptical about them you might be right, but you might also be fooling yourself. If you would like to suggest other helpful ideas for improving interview lesson, please leave them in the comments below.