So the analogy of a attention-spotlight can be used to describe my experience of test-enhanced learning retrieval routines in the classroom. Routines that both brighten and focus the spotlight (more spotlight and less floodlight). Which is arguably the aim of most classroom instruction. The aim of the routine is to systemise and make clear to pupils, what it is they are to do, to attend to, and do. The independent directed thinking task – also serving to minimise distractions.
Two questions? Does a bright, narrower spotlight lead to better short term performance and durable learning? Second, are the same spotlight parameters required for both encoding / learning and maintenance, remembering and relearning phases?
Buchin & Mulligan (2019) The Testing Effect Under Divided Attention: Educational Application
This study assessed the effectiveness of restudying (rereading) and retrieval practice (free-recall and cloze questions) when learning foreign language word pairs and prose passages. Following my first reading of the paper, if I understand it correctly, testing / retrieval is resistant to divided attention effects. Which would appear to be good news for teachers. However, before a second reading I needed to get my head around the ‘attention’ terms used in the paper to be able to explain it further.
On attention and the effects of what researchers refer to as ‘Divided Attention’ (DA) and Full Attention (FA) Buchin and Mulligan (2019) highlight to two distinct phases of learning. I point I have also recently shared via this blog. The two distinct phases: a) encoding / learning and b) retrieval (maintaining / remembering or relearning) or testing.
On attention and the effects of what researchers refer to as ‘Divided Attention’ (DA) and Full Attention (FA) Buchin and Mulligan (2019) highlight to two distinct phases I have highlighted teachers need to consider when teaching: are pupils encoding / learning and maintaining / remembering or relearning.
When attention is divided during encoding, later memory accuracy is greatly reduced. However, if attention is divided during the retrieval or testing phase, there is typically little or no decrement in accuracy.
Why is this? Buchin and Mulligan (2019) propose it is due to Divided Attention’s differential effects on encoding, and retrieval, and the limited research with ecological validity with respect to educational application. Hence Buchin and Mulligan (2019) decided to examine these differential effects.
The Testing Effect Under Divided Attention
This is where I got to learn a little more about dividend attention, from both external sources (e.g., the ambient distractions arising in the environment) and internal sources (e.g., mind wandering, and other off-topic internally generated thoughts). New knowledge that I paused to learn and write about in Attention and retrieval (part 1) – to be able to confidently write this part 2.
Of note, tasks that require active responding, such as ‘off-task use of digital technologies’ (eg laptops and mobile phones) while learning are even more disruptive than ambient environmental distraction (David, et al 2015). I very much doubt that teachers need further warning to not under-estimated the extent to which technology distractions learning.
Buchin and Mulligan (2019) Experiment 1
Classic memory study set up. 40 undergraduates. 64 Swahili-English word-pairs.
- Initial learning phase. 64 Swahili-English word pairs.
- Either restudy or retrieval practice presented as 4 blocks – under either FA or DA conditions.
- FA retrieval – participants were shown only the cue word and told to read it to themselves before recalling the target word out loud.
- DA retrieval – participants were instructed to listen to the 1-9 digits and use the keyboard to indicate if the digit was odd or even.
- A cued-recall test two days later. Participants were instructed to recall aloud the English word associated with each Swahili word presented within the 6s.
Overall, retrieval practice led to greater final cued-recall than restudy under both FA and DA conditions, indicating that the testing effect is robust to distraction. With Buchine and Mulligan (2019: 7) adding that this is important for educators to know, as divided attention “is becoming increasingly more common.” The researchers later qualified that statement in saying “multitasking while learning is increasingly common,” Buchine and Mulligan (2019: 13). Perhaps less so in school classrooms.
Buchin and Mulligan (2019) Experiment 2
The distraction tasks now required participants to keep track of the number of odd digits presented during a single trial – what is described as a “cumulative classification task.”
Results of the final recall test were largely unchanged from Experiment 1. Retrieval practice during Phase 2 led to higher recall than did restudy. Interestingly, the performance on the digit classification task continued to suffer to a greater degree when paired with retrieval than with restudy.
Buchin and Mulligan (2019) Experiment 3
Experiment 3 used prose passages and cloze question tests instead of foreign-language word pairs and free recall tests, educationally relevant and more complex. (Retrieving complex information is more difficult than restudying that same information. It may be that the mnemonic benefits of retrieval will be more susceptible to DA).
The mark scheme caught my attention. Specifically, the test awarded 2 points for correct and complete answers, 1 point for correct but not fully complete answers and 0 points for incorrect, very incomplete, or missing answers. For anyone returning to this blog – you can imagine my surprise to see a 3, 2 and 1 mark scheme – very similar to the model I propose for in class self-assessment.
Results were similar to experiments 1 and 2. Retrieval practice enhanced final recall under both FA and DA conditions when compared to restudy.
What do we learn?
Retrieval practice continues to benefit later memory to a greater degree than restudy learning under both full and divided attention.
Buchin and Mulligan (2019: 15)
Not that we would ever condone or advise learning without full attention – it would appear that retrieval is more resilient to distraction.
For the record, DA disrupted the encoding effects of restudy to a greater degree than the encoding effects of retrieval.
Notes for the classroom
My observation from the classroom is that retrieval is even more “resistant to distraction” when retrieval is personalized / adaptive. That personalisation maintains FA conditions, eg time on task. There are also indirect effects – pupils attendance / punctuality improves, pupils report themselves to be calmer, experience less worry about failure and are more confident in the lessons follow.
We continue to seek high ecological validity and of course seek some generalizability – and add that subject knowledge / content is different and each subject lends itself uniquely to test-enhanced learning routines. It would be interesting to see this same procedure with different knowledge content.
Thinking out loud – how different might DA be from too great a demand on cognitive load?