Writing a Literature Review

Information on research, statistics and publications - tips including how to recruit participants, gain funding, understand your results and get them published.
Post Reply
Posts: 3100
Joined: Sat Mar 24, 2007 11:32 pm
Location: London

Writing a Literature Review

Post by Ruthie »

A literature review in psychology should be a systematic and thorough investigation of all the available literature within a given topic. Depending on what it's for, you may need to pick and choose from some of these tips. For instance, in an undergraduate dissertation on pain in cancer, you do not have the word count to do a full review. Conversely, if you are submitting a lit review to a journal then you need to have been comprehensive in your search, and in theory have read every piece of paper ever written on the subject. A similar level should be required for a PhD.

First off, define the parameters of your search. Say you're looking at theory of mind in autism; are you going to look at Asperger's? High functioning autism? Or just in children? Many fields have a massive literature behind them and you will need to pick your battles to decide how much depth and breadth are required. Consider variations in spelling and naming conventions, as well as abbreviations. MND is ALS in America, MRI is IRM in France, some people call genetic mutations by their short vs long name e.g. D90A or Asp90Ala...

Next, make a plan, and write it in the style of a method section. E.g. "A literature was carried out using the database PsychInfo and the search terms "ALS cognitive", "MND cognitive", "ALS neuropsychology", "MND neuropsychology", "ALS genetics" and "MND genetics".

When you come to write the method section of your literature review, you should provide evidence that you have performed it systematically. e.g. "A total of 193 hits were returned of which 87 were relevant. 6 papers were review articles, 24 were case studies, and the rest were original reports. The author was unable to obtain 15 of these articles. Therefore a total of 72 articles were considered in this review."

Search methods. With the joys of the internet there are a range of tools available, such as
Everyone's got a favourite!

Remember, full access to these articles may rely upon an athens password or institutional subscription. Otherwise you may only get the abstract. Reading only an article's abstract is NOT recommended. Authors often over-exaggerate the significance of their findings in abstracts and with little room for methodological considerations you may be lead down the wrong track. Also, many full-text articles are only available from the recent past, e.g. articles from before the mid-nineties may not be available online. In that case it's off to the real-world library to look for papers in "the stacks".

In both cases, electronic and offline, if you find an article that's relevant, you should keep a copy of the whole thing. Electronically download the PDF and change the filename to something relevant, e.g. Wicks 2006 Cognitive change in subtypes of ALS.pdf so that you'll be able to find it easily if you need to come back to it. If you've got reference manager or endnote you should load the papers into that and consider using each article's unique ID code as your filing system. You should always photocopy the whole of a paper journal because it is that much more difficult to go back to it if you need to. Someone else may have checked it out or it may be waiting to be reshelved. If it's worth reading, it's worth having. Copy all the pages, especially references as you may need to go hunting those down if you come across something of interest.

Weight. How much weight to ascribe to each piece of evidence? First off, consider the type of article. A comprehensive review by a leading expert in the field is worth a lot. A single case report published as a letter in the "Isle of Man journal of psychopathology" is worth a lot less. Although imperfect, a quick method to ascertain utility is to consider the impact factor (http://thomsonreuters.com/products_serv ... ct_factor/) of the journal in which tye're published. It is hard to get published in journals with high impact factors, and articles will have undergone stringent peer-review by the time you've read them. Top ranking journals reject about 85% of everything sent to them, so if a paper doesn't cut the mustard it will bounce further and further down the list of journals within their field. You might want to use tools such as ISI to find out which journals are considered prestigious in your field (see also http://en.wikipedia.org/wiki/Impact_factor)

Right, so you've got a mountain of paper, gigabytes of PDFs, and a blank word document on your screen. Time to start writing? No.

A common mistake that many writers make is to start writing before they know what they're going to say.

First of all you need to get to grips with the mountain of literature you've obtained, and this means sitting down and actually reading it, cover to cover, taking notes. People have different preferences for how they do this, maybe condensing key facts from each paper on to one sheet of A4 and then merging information later on by themes. Other people might think about paragraphs that will be in their lit reviews e.g. definition, diagnosis, prevalence rates, prognosis and devote a separate bit of A4 to each one, filling in the relevant numbers and points from each paper as they go along. Separate sheets of A4 are much better than writing notes in th margins of papers as you go along. Why? Because it's more space-effective to have 10 sheets of A4 than 30 papers to convey the same amount of information, say if you're revising for an exam.

Now some bad news. Whilst you're reading through this big old mountain, you're going to find more papers that you should have obtained. My advice is to write down these references somewhere else, keep a copy handy, and carry on going. If you interrupt your reading and note-making every time you come across a paper you don't have you will spend all your time doing the physical search, and learning very little. So make a note of it and go get the second round of literature in one big go, later on. During your reading you may come across a review which summarises it comprehensively anyway meaning you might not need it. This is also a key step in preventing procrastination as it can feel like real work tracking down every piece of paper in the universe, but you've already got plenty to be getting on with right now!

OK, so you've been digesting papers for 8 hours now and have got a nice set of notes going. You understand the key concepts. You've seen some papers with great methodology and you've seen some that make you slap your forehead. You've come across some flashes of insight that made you smile, and you've seen some backwards thinking that made you cringe.

Now it's time to start writing... the outline.

I like to do this using the outline bullets in Word, the ones that go 1, 1.1, 1.1.1, etc. The outline is a plan for your essay that will help you get the thing written on time, on track, and on word-limit. So say you've got 2000 words to review stigma in schizophrenia. Is is reasonable to spend the first 750 words defining schizophrenia, it's epidemiology, causes, etc.? Probably not. The question is about stigma, not a description of what schizophrenia is. This can probably be done in about 3 sentences, maybe 2 if you're brief. You can always refer the reader to a more detailed description by referencing it in these sentences (e.g. Schizophrenia is a thought disorder characterised by hallucinations, delusions, and other abnormalities of thought. It typically affects young males and may have an increased incidence amongst migrant Afro-Caribbean populations. For a comprensive review see Kumari et al (2006).)

As well as plotting out how long each section will be, you should then start filling in references from your notes. So I know that in my diagnosis section on ALS/MND I'm going to mention the Brooks et al (1999) revised El Escorial criteria paper, so I'll make a note of that there...

Feel free to use subheadings in your review, particularly if you're looking at things from several perspectives. So for instance if I'm looking at hallucinations in Parkinson's disease I might do one section on their nature, prevalence, and relationship to medication, which are all clinical findings. Then I might have a new subheading "neuroimaging evidence" which I can start off without having to come up with a linking sentence. This can help keep things tight and concise.

Now that you've got an outline, and some seeds have been planted, start filling out on the points you've made with sentences, and start joining them together. Don't worry too much about linking them all up yet, just make sure your point has been made clearly. Introduce abbreviations the first time you use them and define technical terms / jargon.

If you find you've got paragaphs that go: "Using a self-report questionnaire in 82 patients, Hogg et al (1995) suggested the prevalence of depresion in MND was 40%. However, using face-to-face interviews, Smith (1997) felt it was more like 8%. These were both contradicted by Jones (2000) who used a trained army of psychics to work out that it was actually 95%...." then you might want to think about making a table. A table can get across a ton of information very quickly and show, at a glance, the trends observed in the literature. You should have a column for first author (do "et al" to save space) and year, one for methodology used (use abbreviations like BDI in the table and then explain them in the legend BDI: Beck Depression Inventory), one column for the sample size, and then the key point, make sure you're comparing like with like. So the column you're interested in might always be "% mildly-severely depressed". If you can't figure out a consistent outcome variable for all your papers just summarise them; e.g. "85% normal on BDI, 10% mild-moderate, 5% severe", then "only 5% referred to psychiatrist". In other words, don't just say "% depressed" and assume that the numbers are equivalent when they've been gathered with different methodologies.

OK, now you've got a bunch of well constructed, relevant, concise paragraphs supported by referenced evidence. This is a great foundation to be working from. The next bit is relatively easy, connect the dots! Starting from the beginning, read the whole thing out loud to yourself. Where are the sentences that take leave you out of breath? Cut them in half. Where are the sentences that make no sense or are written in the passive tense? Reorder them. Fill in the linking paragraphs in between: Whilst self-report questionnaires are useful for gathering large amounts of information, they may be inferior to structured interviews. Several studies have used this technique, such as...."

In the final third / quarter of your review you should be in to the nitty-gritty of the discussion, where you examine some of the controversies within the subject. If you're going to critically evaluate something you should be coming down on one side of the fence or the other. It's no good to say "Lomen-Hoerth thinks dementia is present in 50% of MND patients. However Wicks thinks its only present in 5%". You need to critically evaluate these positions; "Lomen-Hoerth's sample were recruited from referrals to a dementia clinic so her estimates may be inflated, whereas as Wicks' patients were excluded if they had a bulbar onset, or poor respiratory fuction, known risk factors for cognitive dysfunction. Although based on a smaller sample than either study, Jones et al's study is superior to either Lomen-Hoerth or Wicks' study as they used the more systematic Neary et al "Manchester group" FTD criteria in a consecutive sample, arriving at an estimated prevalence rate at 15-20%."

To start winding up the essay, consider future avenues for research, especially where they can combine methodologies from several fields e.g. neuropsychology and neuroimaging, pathology and genetics, etc.

Finally, a brief conclusion to summarise the key points you've made, implications for clinical practice, and where you see the field moving in to the future.

Now go back to the beginning, read it out loud again, check for grammar and spelling and ease of reading. Now leave it for two days, come back, print it out, and do the same thing on a hard copy with a pen. Congratulations, you're on your way to a good 2:1 or a first!

Even though it's not written specifically for Psychology, this is a well-written and comprehensive guide to writing:
http://www.dartmouth.edu/~writing/mater ... what.shtml
And this is also helpful: http://www.york.ac.uk/inst/crd/pdf/Syst ... eviews.pdf

Finally, a site that I found better than the rest for relevance of hits and especially ease of use...


This link is 'psychology specific' for lit review:
https://depts.washington.edu/psych/file ... litrev.pdf

And if you really fancy going head first into reading around how to do a good Lit Review, here are some decent ideas:
  • Aveyard, H. (2010). Doing a Literature Review in Health and Social Care. Maidenhead: Open University Press.
      Chalmers, I., & Altman, D.G. (1995). Systematic Reviews. London: BMJ Publishing Group.
        Fink, A. (2010). Conducting Research Literature Reviews: From the Internet to Paper (3rd edn). London: Sage.
          Greenhalgh, T. (2010). How To Read A Paper: The Basics of Evidence Based Medicine (4th edn.). Oxford: Blackwell/BMJ Books.
            Jadad, A., & Enkin, M. (2007). Randomised Controlled Trials: Questions, Answers and Musings (2nd edn). Oxford: Blackwell/BMJ Books.
              Mawson, A., Cohen, K., & Berry, K. (2010). Reviewing evidence for the cognitive model of auditory hallucinations: The relationship between cognitive voice appraisals and distress during psychosis. Clinical Psychology Review, 30, 248-258.
                McKenzie, L.H., Simpson, J., & Stewart, M. (2010). A systematic review of pre-operative predictors of post-operative depression and anxiety in individuals who have undergone coronary artery bypass graft surgery. Psychology, Health & Medicine, 15, 74-93.
                  Osborne, H., Simpson, J., & Stokes, G. (2010). The relationship between pre-morbid personality and challenging behaviour in people with dementia: A systematic review. Aging & Mental Health, 14, 503-515.
                    Smith, A.E.M., Msetfi, R.M., & Golding, L. (2010). Client self rated adult attachment patterns and the therapeutic alliance: A systematic review. Clinical Psychology Review, 30, 326-337.
                  Note: If you have a suggestion about how to improve or add to this wiki please post it here. If you want to discuss this post please post a new thread in the forum.

                  Content added to by Workingmama 24/06/2019
                  Post Reply