The art of summarising

One of the most useful single pages I’ve read over the last couple of years is Dr. Phil’s system for coding social science journal articles (use Amazon’s “Look inside” feature to see the coding summary on page 3). I was reminded of it at the start of the year when I was trying to extract the argument structure from one of the attention papers, and again when reading and writing literature reviews during TMA04.  Dr Phil’s Summary of Previous Literature (SPL) definition describes the art of summarising perfectly:

This process entails a tremendous amount of condensation, taking complex ideas and reducing them into paragraphs, sentences, and if the author is brilliant, one word.

Every journal article you read (especially experimental reports) will be full of SPLs, but an example from the summary of language processing (OU students only) jumped off the page for me.

Homophones can be miscategorized (Van Orden, 1987)

On the one hand I was a bit dismayed that a 15,000 word paper which I covered in about 300 in my literature review last year could be so well summarised in just 4!  On the other, knowing the paper fairly well I appreciated such a concise but accurate summary (although, following the Feynman technique, I would actually add an example)

Homophones can be miscategorized (Van Orden, 1987).  For example, the word ROWS might be miscategorised as A FLOWER as it is phonetically similar to ROSE.

The art and skill is to select the words that save others the extensive time and effort required to find and read the primary source (see TMA03 for an example).  I would actually go one step further and say that using no words is one pole of this continuum i.e. choosing to write nothing because it isn’t relevant to a TMA or exam question is a type of summarising.

Advertisements

A chapter summarising framework

[19/8/2014: EDIT: My thoughts on summarising have changed a lot since writing this, especially after trying to create summary notes for exam preparation.  Whilst there are still some useful ideas for processing the module content I’m not convinced that it’s comprehensive or the optimal approach to summarising in a way that supports answering exam questions.]

It was suffering the consequences of poor summarising that initiated my search for active tasks.  Other people’s summary notes provide useful ideas for summarising techniques and ideas you may have missed, however it’s no substitute for the deep processing that comes from producing your own summary notes.  Focussing on the process rather than the product reduces temptations to passively reading the book or other’s notes.  Furthermore, your summarising skills need to be continuously engaged to stem the tides of information in DD303.  ABC.  Always.  Be.  Condensing!  The good news is, the summarising process itself is an active learning task in its own right.  This approach is a work in progress but it seem to be taking me in the right direction.

First I’ll describe the process I take when reading a book chapter then I’ll get into the specific techniques I use to summarise the content.  As my PDF reader allows me to, I start off by adding bookmarks for each section.  This gives me the broad overview of what to expect from the chapter and allows me to quickly jump from section to section.  Next I create a word processor document for the chapter and scan the PDF chapter for keywords, adding them to the first column of my key terms table.  I get a notebook and selection of coloured pens and then I’m ready to start reading the chapter.  I read each section (or subsection) closely trying not to move on until I’m sure I’ve understood what I’ve read.  This is a slow process, but my aim is to only ever do this once!  Here are the things I do during reading:

  1. Whenever a citation appears I add it to Zotero.  Wherever possible I add citations to journal articles using results from OU Library searches (I revert to Google for journal articles unavailable at the OU).  I keep the printed book open at the section I’m reading and the PDF on the references page so I can copy and paste search terms.  Book citations are added using Google Books.  I usually read article abstracts and sometimes download the full article if it looks very relevant/interesting.  This sounds like a lot of work but using the Zotero browser plugin you this only about 5 minutes per item.  It also has a number of other benefits.  First, I find that slowing down the reading process also helps to ensure I’ve understood what I’ve read and the abstract often clarifies points summarised in the chapter.  Second, it’s provides regular literature searching practice which should help with TMA04.  Third, it progressively builds your Zotero reference database which makes referencing TMAs reasonably painless.  Finally, it’s useful for avoiding plagiarism as you can quickly add citations to your word processor notes to distinguish other people’s ideas from your own.
  2. I pay special attention to key terms in preparation for writing them in my own words.
  3. I summarise the section in hand-written and drawn notes.

Words …

Clearly, one of the major summarising approaches is to replace lots of other people’s words with fewer of your own!  Defining key terms in your own words is an active summarising task that reduces large sections of the chapter to very concise summaries, particularly if you use key terms in the definition of other key terms.  A related technique I use is to copy the summary bullet points from the PDF into my word processor document and summarise them further!  My key term definitions are sometimes good enough that I can remove a bullet point altogether.  If I’ve understood the material well then the remaining bullet points often seem quite verbose so I’ll re-word them using (and/or linking to) key terms and abbreviations (see below).  I occasionally add additional bullet points for things I thought were important or interesting.  This task also acts as a section review.

A second skill that really helps with summarising is deciding what’s relevant and at what level of detail.  There seems to be something of a knack to effectively capturing what’s relevant and eliminating what isn’t but the payoff in terms of avoiding re-reading irrelevant material is enormous.  Eliminating big chunks tends to be easier with primary sources (I got to page 4 of Henson (2005) before I though anything was relevant enough to note for TMA03) whereas reducing the level of detail seems more appropriate for the module textbook.  By combining written and visual summarising my chapter summaries seem to run to about 4 sides of A4 paper.  The written summarising relies on a (growing) list of abbreviations which I use to write concisely.  Some of these are standard (psychological) abbreviations I’ve noted when reading books and articles and some are my own.  I started to use abbreviations methodically when I realised how helpful they are when testing your recall during revision and for writing exam answer plans.  Here’s a selection that’ll I’ll try to update as and when.

Abbreviation Meaning Example
Ps Participants
P Probability
Is Individuals
Tn Task 1-n
Cn Condition 1-n
R Response
RT Response Time
R-K Remember-Know (LTM)
Sn Stimuli 1-n
x bar (mean) average
proportional to frequency ∝ field strength
f() Function of Remember/know judgements f(recall test)

… and Pictures

In addition to summarising in words, my other big gun is to represent knowledge visually.

Summary of DD303 Cognitive Psychology, Chapter 6 (Long Term Memory) Section 4 (Retrieval)

This first example covers most of the visual summarising techniques I use but requires some commentary.  Whilst I can’t quite get a picture to paint a thousand words, the 2 pictures in this example do paint 700 words (I counted).  I’ve tried to standardise on colours and icons that allow me to see particular themes at a glance.  I haven’t come across much prescriptive advice on how to use these visual approaches effectively and I’ve seen some horrible uses that literally made module content more confusing.  I think you need to plan the way you use visual approach quite carefully to maximise their impact.  I looked at exam questions when I was coming up with ideas for visual note-taking and thought it would be useful at revision time to be able to quickly scan notes to look at them from different perspectives (methods, populations etc.)  Currently I’ve gone for orange to highlight researchers and years and a green circle around an icon indicating the type of research (experiment, model, theory, neuroimaging, neuropsychological, etc.).  I also try to represent the population to highlight whether the study relates to individual differences, developmental, clinical issues etc.  I write key terms that I use in blue.  Second, in general and particularly with experimental studies I try to come up with a way of representing the whole study graphically.  This is a really active task that forces me stop and really think about how to represent the study and findings.  Initially this made reading a really slow process, but the more of these I do, the quicker I get for a couple of reasons.  First, I’m building a library of ways to visually summarise (little correlation, interaction graphs etc.) a whole study in a small space.  Second, after a while the experimental design starts to jump off of the page and it because easy to represent results visually (see 2 x 2 design from Morris et al. (1997) in the example).  With these little graphics in place, arrows are useful to then relate one study to another.  In this example you can see that I’ve linked encoding specificity and Transfer Appropriate Encoding as complementary findings supporting respectively information and processing links between encoding and retrieval.  I attach green icons to arrows to indicate complementary findings and red to indicate conflicts.

This next example is a re-working of a table (towards the bottom of the image) I used last year to represent the explanation (p164-165) for the puzzling findings of the Word Length Effect and the Phonetic Similarity Effect. I had to work hard to understand that short section and didn’t want to repeat that each time I needed to get my head around the explanation. It might give you some general ideas for how to visually compare the findings from two studies.

Graphical comparison of Word Length Effect and Word Similarity Effect

The Word Length Effect vs. the Word Similarity Effect

Incidentally, I tried to help a student on Facebook this year who was also confused by this section.  I think it would have taken me even longer to formulate this response if I hadn’t worked out the tabular approach last year!

I don’t think it’s “visual recall” that’s the issue here. Recall was (had to be!) written (rather than spoken as in previous studies) as there was articulatory suppression (AS) taking place during recall. The WLE goes because AS prevents rehearsal, on which it depends. The PSE remains as it’s not affected by rehearsal. I hope that’s accurate as it took me about an hour to think about!

Finally, here’s an example of a full page of my visual notes and  a summary of working memory research in Alan Baddeley’s own words.

Full page of visual notes for DD303 Cognitive Psychology Chapter 5 (Working Memory), Section 2 (The structure of working memory).

 

Resources

References

Henson, R. (2005). What can functional neuroimaging tell the experimental psychologist? Quarterly Journal of Experimental Psychology: Section A, 58(2), 193–233.

Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16(5), 519–533.

Treisman, A. (1998). Feature Binding, Attention and Object Perception. Philosophical Transactions of the Royal Society: Biological Sciences, 353(1373), 1295–1306.