Formative assessment #3 — giving more effective feedback

In this third part of my series on formative assessment (FA) in Stile, I'll share some more tips from Dylan Wiliam's 'Embedded Formative Assessment' and how they look in Stile.

Previously, I outlined why FA is quite possibly the most efficient and effective way to boost student achievement, and how important it is to define and communicate clear learning goals.

This post is loosely based on chapter 5 ('Providing feedback that moves learning forward') which kicks off with the following warning:

It seems obvious that feedback to students about their work should help them learn, but it turns out that providing effective feedback is far more difficult than it appears.

Much of the feedback that students get has little to no effect on their learning, and some kinds of feedback are actually counterproductive.
— Dylan Wiliam, 'Embedded Formative Assessment'
wiliam.png

Marks actively hinder learning

Dylan begins by citing a number of studies that effectively confirm that assigning marks or scores to student work gets in the way of learning. If you leave feedback along with marks, students tend to only look at the marks and skip the comments.

We covered this in our post on why you're wasting your time by giving both marks and feedback. Time is teachers' most precious resource, so let's not waste it. Or in the words of Alfie Kohn:

Never grade students as they are still learning
— Alfie Kohn, quoted in Embedded Formative Assessment

Luckily, lessons don't have marks associated with them in Stile by default, so we've got you covered there 😁

To praise or not to praise?

checkmark2.png

Those of you who read my post 'The art of feedback', which discussed quite possibly the most cited paper on feedback of the last 10 years (John Hattie's and Helen Timperley's article 'The Power of Feedback'), may remember the authors' claims that praise is both the most common form of feedback observed in classrooms and the least effective.

Wiliam echoes those findings, citing research that ties praise to ego-involvement rather than behaviour ("I'm (not) smart" vs "I put in a lot of effort") which may dissuade students from wanting to improve.

Apparently, the most effective teachers praise less frequently, and in a "credible, contingent, specific and genuine" manner relating to factors that are within the student's control. 

When it comes to praise, less (and targeted) is more.

Timing feedback

Automated feedback in Stile

Automated feedback in Stile

Being able to deliver instant feedback is often seen as one of the hallmarks of electronic learning. However, getting instant results is not always the most beneficial for learning, as less mental effort is extended if kids can just try again right away until they get it right.

Having to work out answers carefully before clicking 'Submit' (like students do in Stile) is actually improving their mindfulness and, as a result, learning. And with automated feedback, giving personalised as well as perfectly-timed feedback is even easier!

Can feedback lower student achievement?

A meta study found that out of 131 studies on feedback conducted over 90 years, 50 studies (two out of five) found that the feedback provided actually lowered performance – that's a significant chunk!

The researchers looked at the results very carefully and found that in eight possible situations, only two had a positive impact, depending whether they met their learning goal or not:

  1. if a student met or exceeded their learning goal, only changing their goal led to increased aspiration
  2. if a student didn't meet their learning goal, an increase in effort led to better results.

To complicate matters further, results also depend on

  • the individual receiving the feedback
  • the kind of task at hand and
  • the recipient's perception of the person giving feedback

Wiliam also mentions Carol Dweck's work on the topic, specifically whether students attribute their failures and successes to internal, stable factors (ability, intelligence) or external factors (effort, lack of sleep, etc). The conclusion drawn was that the most successful learners personalise their results (i.e. "it's up to me") and believe that they can do something about it. It is this kind of attitude we must help nurture in our students through feedback.

However, the feedback needs to forward-facing: pointing out deficiencies in student work without giving a chance for resubmission demoralises students.

Instead, focus on ways to improve, looking to the future and give students a chance to have another go. 

Therefore, don't be afraid to hit the 'Resubmit' button and get students to refine their responses, hone their skills and improve their knowledge. Or in the words of Wiliam:

Feedback functions formatively only if the information fed back to the learner is used by the learner to improve performance.
— Dylan Wiliam, 'Embedded Formative Assessment'

Also, before you ask for a resubmission, why not throw in a written-response question at the end, asking them how they incorporated your feedback? That encourages them to not only look at your feedback again but also compare their work against the success criteria and steps outlined in the feedback to reach them.

What kind of feedback is best?

In a nutshell: achievable next steps that learners can put into action right away and cause thinking (as opposed to reacting emotionally). This requires, of course, that you know exactly

  1. what the learner already knows or can do (elicited knowledge)
  2. what she doesn't know yet (the stated learning goals)
  3. how to break down the steps to get her her there (an action plan)
  4. how to deliver this information in an accessible way 

Don't just point out what's wrong; give short, helpful tips and specific examples for improvement, or as Wiliams calls it, a recipe for future action' or a series of activities that get learners from the current state to the goal state.

Practical tips:

Here a few practical things you can start tomorrow:

  1. Mark student work with +, - or =: if they've improved on the last piece of work, they get a +, if the quality of their work has decreased they get a - and if it's the same the get a =. 
  2. Always provide three ways to improve for every submission, no matter how good or poor it was overall.
  3. 3 questions: a variation on the above, where you annotate student with 1, 2 and 3, add a written response question at the bottom of the lesson and ask students to explain each annotation in more detail
  4. Tell students how many answers are not up to scratch (e.g. 5 out of 20), then ask them to find and fix them (works well for maths). Finding problems is hugely helpful in advancing learning, so let's make it part of the process.

In my next post in the series, I'll outline how we can find out what that current state is!

Guido Gautsch
VP of Happiness