1. From cognitive
psychology to learning
design
Dr Chris Atherton
University of Central Lancashire
@FiniteAttention

2. Overview
• Tufte & Death By Powerpoint
• The limits of Working Memory
• Schema theory and learning design
• Cognitive Load Theory
• The use of sparse slides
• You’re really reading this, aren’t you.
• Look — reading is almost certainly
compromising your ability to listen.
Really. I’ll get into that shortly.
The audience started laughing before I got to the last points,
illustrating nicely that they read faster than I can talk.

3. niwer
So let’s start again, before the bullet-points.
(I think this is kinda pretty.)

4. Tufte has done some awesome stuff (e.g. the info-to-ink-ratio)
but he’s still a bit “shoot the messenger” for my taste.

5. death by PowerPoint
• Tufte & Death By Powerpoint
• The limits of Working Memory
• Schema theory and learning design
• Cognitive Load Theory
• The use of sparse slides
• You’re really reading this, aren’t you.
• Look — reading is almost certainly
compromising your ability to listen.
Really. I’ll get into that shortly.
Here I talk about “Death By PowerPoint” a bit. Google it if it’s
new to you, though chances are the sentiment is familiar.

6. (Wordle.net)
Barack Obama’s State of the Union speech, rendered using
wordle.net. Word size reflects frequency of use.

7. (NPR.org / Wordle.net)
NPR asked 4,000 people for three words they remembered
from the State of the Union. Salience trumps repetition.

8. death by PowerPoint
• Tufte & Death By Powerpoint
• The limits of Working Memory
• Schema theory and learning design
• Cognitive Load Theory
• The use of sparse slides
• You’re really reading this, aren’t you.
• Look — reading is almost certainly
compromising your ability to listen.
Really. I’ll get into that shortly.
When we gripe that PowerPoint is intrinsically unmemorable,
remember that memory is also intrinsically unreliable.

9. Probably not evil.
In other words, don’t shoot the messenger. PowerPoint can
be used well, or badly. We’ll come back to this a bit later.

10. Experiment time. Here I told the audience they would have ten
seconds to memorise the information on the next slide.

11. WTMHYELABPP
(10 seconds elapses before I take this slide away)

12. “Okay, now turn to the person next to you and figure out:
How much of the previous slide can you remember?”

13. (Miller, 1956)
magic number 7
A lot of people recalled around 7 items,
probably in part because we tested recall right away.

14. (Cowan, 2001)
magic number 4
In fact, it’s easy to demonstrate that memory for new
information might be as small as 4(ish) new, unrelated pieces.

15. WTMHYELABPP
But that’s partly because these 11 pieces of information
don’t seem to make any sense.

16. WTMHYELABPP
that’s amore
But when I add a clue, suddenly it all makes sense.
(I didn’t mean to sing this part; it just came out. Oops.)

17. WTMHYELABPP
that’s amore
Now these 11 pieces of information have become one piece,
which illustrates the idea of ‘chunking’ information

18. “The procedure is quite simple. First, you
arrange items into different groups. Of
course one pile may be sufficient depending
on how much there is to do. If you have to
go somewhere else due to lack of facilities
that is the next step; otherwise, you are
pretty well set. It is important not to
overdo things. That is, it is better to do too
few things at once than too many. In the
short run this may not seem important but
complications can easily arise.”
It’s demonstrably hard to recall or even process this
information, because it’s just so vague. What is it about?

19. doing the laundry
“The procedure is quite simple. First, you
arrange items into different groups. Of
course one pile may be sufficient depending
on how much there is to do. If you have to
go somewhere else due to lack of facilities
that is the next step; otherwise, you are
pretty well set. It is important not to
overdo things. That is, it is better to do too
few things at once than too many. In the
short run this may not seem important but
complications can easily arise.”
Oh, I see.
(Bransford & Johnson, 1972)

20. (e.g., Anderson, 1977)
schema
Schema = story or rule of thumb. Having activated our laundry
schema, it’s much easier to understand and recall the text.

21. hanging information
up
Good learning design bridges the gap between people’s
previous knowledge (the tree) and new info (the decorations)

22. (e.g., Sweller, 1994)
cognitive load
Cognitive load theory is a useful framework for thinking about
where working memory is being ‘spent’ during learning.

23. Who likes Japanese food? Even if you love sushi, you can only
have so much; there is limited room in the bento.

24. ... unless you ask for no salmon and extra sushi. Bear with me.

25. working
memory
Can you tell where this slightly laboured metaphor is going?

26. Just as the bento has limited space for your lunch, so working
memory has limited capacity for the different parts of learning.

27. Intrinsic
Difficulty
Some of your working memory is taken up by how difficult the
task is. Nothing we can do about this; complex is complex.

28. Intrinsic
Delivery
Difficulty
Some of working memory is also consumed by
the manner and mode of delivery.

29. Intrinsic
Delivery
Difficulty
Schema
Building
Some of working memory is used up on ‘good work’, attaching
new information to the ‘tree’ of previous knowledge.

30. Intrinsic
Delivery
Difficulty
day- Schema
dreaming Building
If none of these is too onerous, maybe you’ll have
some space left over for other things.

31. Delivery
Intrinsic
Difficulty
Schema
Building
We know that during hard tasks, we have to concentrate
more; less room for daydreaming or other trivia.

32. Intrinsic
Delivery
Difficulty
Schema
Delivery
Building
Sometimes, a lot of working memory can be taken up
by the information-delivery process.

33. Intrinsic
Delivery
Difficulty
Delivery
Schema
Building
... sometimes maybe at the cost of actually linking the new stuff
with your previous knowledge of the subject.

34. Intrinsic
Delivery
Difficulty
Schema
Building
Ideally, we want as much working memory as possible spare, to
help us learn effectively by elaborating on existing schemas.

35. E
x
t
r
Intrinsic
a
n
e
o
u
s
Germane
Cognitive load theory uses these terms to refer to these
different impositions on working memory.

36. reducing extraneous
load
So surely, one of the aims of good learning design should be to
reduce extraneous load and promote germane load?

37. 1. video
Our first experiment (Atherton, Morley & Pitchford, in prep.)
involved making a video with slides and an audio track.

38. traditional
The audio track was consistent for all presentations; the slides
varied. Traditional slides were bulleted and info-dense.

39. sparse
Sparse slides transitioned more frequently;
they each contained less text.

40. sparse with diagrams
The last condition also had sparse slides, but these also
contained diagrams where appropriate.

41. themes
4
3
2
trad. sparse sparse
text text &
pics
People who watched the video with sparser slides
recalled significantly more themes afterwards.

42. 2. lecture
That’s okay for video, but what about in-person learning?
Here we tried this again, but ‘live’ in the lecture theatre.

43. traditional
Again, we had a ‘traditional’ condition where participants saw a
set of slides with bullet-points and lots of information.

44. sparse
The other condition used sparse slides (again, more slides with
faster transitions) and tried to keep slide-info to a minimum.

45. themes *
7
6
5
4
3
2
1
traditional
trad. sparse
sparse
slides text
slides
Again, we found that those in the sparse text group recalled
significantly more themes from the presentation.

46. E
xt
ra
Intrinsic
n
eo
us
Germane
So how does this fit in with our model of cognitive load?

47. Ex
tr
a ne
ou
Intrinsic s
Difficulty
Germane
One interpretation is that sparser slides reduce extraneous
load, leaving more working memory for schema building.

48. WTMHYELABPP
Let’s come back to Dean Martin for a moment.

49. WTMHYEL
Sure, working memory is limited, but the phonological and
visual parts of the brain each have their own WM store.

50. WTMHYEL
In other words, learners can potentially receive
more information if you feed both centres.

51. death by PowerPoint
• Tufte & Death By Powerpoint
• The limits of Working Memory
• Schema theory and learning design
• Cognitive Load Theory
• The use of sparse slides
• You’re really reading this, aren’t you.
• Look — reading is almost certainly
compromising your ability to listen.
Really. I’ll get into that shortly.
So if we think about Death By PowerPoint again ...

52. death by PowerPoint
Bored
Overloaded
... visual working memory has very little to do, while
auditory working memory has to read and listen!

53. The part of the brain
that lights up when it
sees pictures of the
brain.
Mea culpa: forgot attribution here. Sorry!
http://flickr.com/photos/quinn/4252155172

54. The part of the brain
that lights up when it
sees pictures of the
brain.
Question everything! There’s evidence that even when
neuroscience-related terminology doesn’t directly
support an argument, we’re still more likely to accept the
argument if it is present (Skolnick Weisberg et al., 2008)

55. Let’s talk!
@FiniteAttention
FiniteAttentionSpan
.Wordpress.com
CJAtherton@UCLan.ac.uk
Thanks to my colleagues Andy Morley and Mel Pitchford,
and to Learning Technologies for the invitation to speak :)