There’s a wealth of research supporting the use of formative assessment; as synthesised by Hattie (2012). No research has been more pivotal in education circles than that of Black and Wiliam (1998); who together defined formative assessment as:
“all those activities undertaken by teachers, and/or by their students, which provide information to be used as feedback to modify the teaching and learning activities in which they are engaged”
Whilst overall positive, there’s also some criticism of formative assessment in the research. Dunn and Mulvenon (2009) explore the research used in Black and Wiliam’s ‘Inside the Black Box’, revealing some interesting problems with some of the claims. Furthermore, David Didau makes some valid points in his article highlighting some of the limitations of formative assessment, largely about the difference between learning and performance, the value of success criteria and the limitations of feedback. Wiliam’s response addresses some of these concerns, whilst also recognising the limitations and is well worth a read in juxtaposition with Didau’s.
I’m firmly in favour of formative assessment, but believe it should be well-considered and used meaningfully to support learners. Wiliam (2011) identifies five pillars to successful formative assessment.
In summary of the 5 stages, I’d like to discuss my opinions of each stage and how teachers may wish to consider their own use of formative assessment.
Step 1. Sharing Learning Expectations.
I believe it’s important to give learners clear and challenging learning intentions. If not, it’s like them going on a journey with no idea of where they’re going. Under typical circumstances, this surely wouldn’t be of benefit. I can see that it may be useful on occasion to not tell learners what the intention is in order to make them curious, or as not to ruin an element of surprise in a lesson. However, on the whole, I think it’s an important thing to be doing to ensure that there is no confusion as to the purpose of the lesson.
What I’m not so sure of is giving the learners success criteria. Ensuring that learners are clear with what it means to be successful may be important in some, but not all instances. My problem with this is that ‘success’ doesn’t always fit neatly into a checklist/rubric of expectations. Moreover, some lessons can easily take a complete change in direction and this isn’t necessarily a bad thing – learning should happen over time, not in singular lessons or in a linear fashion. Now my last point may well be a contradiction to my first statement about intentions, but having a plan is better than winging it right?
Step 2. Eliciting evidence
We can’t assess learning in a single session, but we can check and clarify initial understanding and correct any initial misconceptions. This doesn’t meant that we are gathering evidence of learning though – make sure we understand that from the outset; it’s more performance at that moment in time that we can gather information on. In order to do this, we need to build a picture of what learners know. Asking a few learners questions is not representative of the whole group, so we need to use methods that allow for a bigger picture to be painted. In intervening where there are issues, we can ensure that we ‘steer’ learners back on track and guide them towards the destination.
One approach that Wiliam (2011) suggests is ‘hinge questions’. Although he proposes these as diagnostic questions that should only take a minute to take and analyse results from the whole class before moving on, I believe that they are a really effective way of assessing understanding and generating discussion, particularly if the questions are well thought through. Harry Fletcher-Wood has provided an accessible, really useful summary of hinge questions which should help anyone in trying them. Be warned – it is a real skill to develop high quality hinge questions.
Further to this, a great way to engineer classroom discussions (which I have being trying to develop myself) was shown by Wiliam at a recent training event, whereby there is a ‘no further questions’ approach to getting the class to discuss their views. How is this done I hear you ask and what is the point? Well, this is expertly explained by Toby French here and here.
Step 3. Feedback
Kluger and DeNisi (1996) carried out an extensive review of the research on feedback. In summary, 38% of studies on feedback reported negative results. This is very concerning in light of the fact that it is so over-prescribed in our education establishments. Learning is a highly complex beast because we are dealing with the thoughts and emotions of individuals. Giving feedback in one way may benefit some, but may hinder others.
I like the whole feedback being like the driving a car analogy. If you constantly look in the rear view mirror, you’re unlikely to get anywhere fast, but if we acknowledge the mirror and focus our attention on the windscreen (i.e. what’s in front of us), we’re more likely to succeed in getting to our destination (there’s a common ‘journey’ link in this post). So if we try to focus our feedback to learners in a way that focusses on the future, how you do this is entirely up to you and down to the individual learner that you know so well. Forget the infamous ‘sh*t sandwich’ and any other approach that is suggested. Some learners may thrive from a lot of criticism, with others needing more of a pat on the head. Know your learners and do what’s best for them – but do so looking forward and not dwelling too much on the past.
4. Activating learners as resources for one another.
We know the problem with feedback and to a large extent, this is the problem with peer assessment too. Nuthall (2007) found in his extensive study (The Hidden Lives of Learners) that 80% of the feedback that learners get is from their peers, with a whopping 80% of that being wrong! So by my calculations, if 80% of feedback is wrong from learners to peers and 38% of the remaining 20% is likely to have a negative effect, that means that only 12.4% of learner to learner feedback is any good. Alright, so that is completely lacking in any evidence base or statistical reasoning, but you get my point I hope – this is dangerous. I can actually resonate with bad peer assessment. I’ve observed and delivered many lessons where a learner with limited understanding of a concept has ‘confidently’ explained it incorrectly to a peer and then I’ve frantically tried to amend the mistake they’ve made, but not really known if they’ve remembered what I’ve said or what their mate said.
Group work can be useful in some instances though and I’m not completely ruling it out. Learners need to develop the ability to work with one another, there will be many occasions in the workplace they’ll need to complete tasks in teams. But mindless planning of group work is another thing to be aware of. See my post here for more information on this.
5. Activating learners as owners of their own learning
This involves learners self-assessing against a set of criteria or of their understanding. Self-assessment gets a massive effect size of 1.44 according to Hattie (2012) as it encourages greater autonomy and helps learners to engage better with feedback according to Black and Wiliam (1998). Of course, there is a lot more to it than that, but I worry that it is regularly used thoughtlessly and often alongside target setting (which I have written about previously here). I think it may have merits, but with so many cognitive biases, am concerned about the accuracy of it and hope to explore my reservations about this in a coming blog post. Let me give a basic example: I ask how well learners understand a concept by asking them to hold up green (good), orange (average), or red (poor) cards. Firstly, how honest are learners likely to be with this? Secondly, what is a ‘good’ understanding (is this where I need my success criteria? – but what if the lesson changed course? Even with success criteria, surely it can be interpreted differently?).
In summary, formative assessment should underpin every lesson in my opinion, but the approaches adopted need to be done so carefully and with a critical mind. Just because a learner is demonstrating that they understand something at a moment in time, this does not necessarily mean that this is the case, nor does it mean that they’ve learnt anything. I quite like this slide from the Dylan Wiliam training event that I attended. It nicely sums it up, but each step needs to be considered carefully.
As always, comments are welcomed.