An Educational Researcher’s Journey


An Educational Researcher’s Journey by Mairi Ann Cullen and Jenny Delasalle

 

This chapter follows a search journey in seven parts carried out by Mairi Ann Cullen, a researcher at the University of Warwick’s Centre for Educational Development, Appraisal and Research (CEDAR). This particular journey took place in April 2011 but the principles employed along the way will be of relevance across the disciplines and across the years. Within each part of the journey, the literature search principles exemplified have been described by Jenny Delasalle, at the time a Librarian at the University of Warwick with a remit to support researchers.

Our researcher’s journey is described in seven parts, and the librarian’s search principles form a list as long as the alphabet, so a kind of A-Z guide to literature searching grows from one individual’s journey. The librarian’s principles are tips that researchers from other disciplines might also find useful, so this chapter works rather like an illustrative travel guide.

Mairi Ann Cullen’s original blog posts and related trackbacks are available on the University of Warwick’s ‘ResearcherLife’ blog.

 

Part one: Where I’ve come from

I listened and learned from Principal Investigators (PIs) and colleagues about creating a literature search template and defining a set of limits. They taught me that it was imperative to keep a running note of how I was going about the process so that I could later describe what I had done. As time and technology went by, I picked up some cursory knowledge about literature searching online. Recently, however, I’ve become more aware of just how cursory this knowledge was – I knew I wasn’t getting the most out of the search tools available and turned to the Library staff for help.

Today, I’m going to begin to put into practice what I’ve learned so far from Chris Bradford and Jenny Delasalle, Academic Support staff in the University of Warwick Library. Most importantly, they taught me that it’s worth learning how to use each specific database that is relevant.

 

Part one: Literature Search Principles

  1. When you’re already familiar with all the relevant articles you find, then it’s time to stop searching.
  2. Follow the reference trail: when you find a good article, look at the papers that it cites.
  3. Take advice from experienced researchers.
  4. Keep learning throughout your career: the search tools available to you will keep evolving.
  5. Library staff will be up to speed with the latest search tools and techniques, so take their advice and tips.
  6. Explore all the databases of relevance to your research.

 

Part two: My literature search launch

Today I’m starting to search for literature to situate a draft article on which I’m joint, but not lead, author. My task at this stage is to place our findings and our argument in the relevant literatures around parenting styles, anti-social behaviour, social class, and social policy. I’m familiar with the literature on parenting styles so I’m going to start with literature on anti-social behaviour.

First stop, of course, the Library homepage – I clicked on Resources, then selected my Subject as ‘Education’ (I’ll start there but I plan to come back later and try again by selecting ‘Social Sciences – all subjects’) and my E-resources as ‘Databases’. These choices take me to the reassuring Education resources page with the photo and contact link for our Academic Support Librarian, Chris Bradford. I know that I can ask her for help if I get myself lost in the literature searching minefield.

I select ‘Key electronic resources for Education’ which takes me to the recommended set of relevant databases. I scan the list and, at this stage decide to ignore the ‘key list’ and go for the wider remit of ASSIA (Applied Social Sciences Index)  as I suspect my search term, ‘anti-social behaviour’ is not Education-specific. The CSA Illumina search page opens up. Having benefited from a conversation about literature searching with Jenny Delasalle, I’m aware of the need to learn a bit about each database as I use it so I take advantage of the option to click on ‘search tips’ before I launch off on my own.

 

Part two: Literature Search Principles

  1. Be clear about the purpose of your literature search before you begin.
    More on this: Setting your own work in context, exploring a new topic, finding literature from other disciplines, carrying out a systematic review all require different approaches, amounts of time and record-keeping practices. You will need to follow an approach suitable to your own discipline and needs.
  2. Be precise about the topic(s) you wish to explore.
    More on this: For ideas of possible ways to refine your search topic, visit the advanced search form on the database you are searching: what elements could be useful to you? Refinement ideas might come to you as you handle the literature that you find. See Part three of this journey, where Mairi Ann explains her topic more precisely; Part four, where she is reminded to focus her research on Britain in particular; and Part five where she recognises the danger of side-tracks.
  3. Find out what resources are recommended for your discipline by your library website or librarian.
  4. Note the difference between the database or source of information (in this case, ASSIA) and the search platform or tool you are using (in this case CSA Illumina).
    More on this: See also Principle M, know your database or source. Some platforms will allow you to search many databases at one time, and some databases are available on more than one platform (eg 30+ interfaces for PubMed to search MedLine). This can make it worth your while investigating which databases are of interest to you, and which platforms you can use to search those databases: one particular platform might save you time by bringing you results from a number of databases. However, you often get the most specific search/refinement options when searching one database at a time, so if you are getting too many results then it might help to search only one database at a time.
    Also see Principle Q: you should know what elements of an article a keywords search will be searching through.
  5. Use the database’s or tool’s own ‘search tips’, or ‘help’, or guides, to learn how to get the best from that particular database.
    More on this: Often, you can find bite-sized explanations and tips behind elements of a search page from links on the search page itself. When you want to use wildcard, truncation or phrase searching (see Principle O in Part four and Principle R in Part five), you need to know how to express those in the characters recognised by that database.

 

Part three: CSA Illumina search tips considered

Below, in bold text, are the first three tips on CSA Illumina’s ‘Tips for successful searching’ page (from 2011), along with my responses:

Before building your search, ask yourself the following questions:

1. What information are you looking for? Consider stating this in writing in two or three sentences.

Articles published on anti-social behaviour (at least since 1997 when New Labour government elected); would be good to know when and why this term was coined and what exactly it means so articles defining and/or critiquing the term would be useful; in particular I need to find articles that provide empirical evidence for predicting who is ‘at risk of’ this type of behaviour, and also articles about interventions around anti-social behaviour, including parenting programmes.

2. Identify general concepts about the information that you're looking for, and ask how they might relate to the search terms that you might use.

General concepts around anti-social behaviour? Opposite might be ‘pro-social behaviour’. Similar terms might include ‘conduct disorder’, juvenile delinquency’, ‘offending behaviour’, ‘juvenile offending’.

3. Are you interested in a specific author's work? If so, then that might be a very good place to start. You might use articles by this author to discover other relevant documents to satisfy your research needs.

I used to think this approach was ‘cheating’ and not a ‘proper’ literature search strategy, but here it is recommended on a database site. This approach was also recommended by Jenny Delasalle so, after all these years of thinking I was doing it wrong somehow, I feel vindicated about starting from one known piece of the literature and launching off from there. However, I’m keen to gain a feel for the field beyond what I know already so I’ve noted this tip but it’s not where I want to start today.

 

Part three: Literature Search Principles:

See Part two, Principles G and H on being clear about your purpose and precise about your topic. Also in Part one, we mentioned the reference trail in Principle B.

  1. Consider searching by author, if there is someone well known in this field.

 

Part four: CSA Illumina’s journal coverage and ASSIA

The fourth search tip on CSA Illumina (in 2011) says:

Are there specific journals that publish articles related to your subject? You might want to take a look at the serials source lists attached to each database’s factsheet (the factsheet appears as an ‘?’ button next to each database name).

This sounds interesting. I try following the ‘?’ next to the database name, I click on ‘selected databases’ and there ASSIA comes up as the first one so I click on the ‘?’ beside it. Wow! Now that does bring up useful information, including subject coverage, dates of coverage (from 1987) and, over on the left-hand side, Serials Source List and an ASSIA database guide. I click on the serials source list and up comes a long list of all the journals included in this database. Some, including some that I’ve never heard of before, do look as if they might cover my topic of ‘anti-social behaviour’. Here are a few – and only one of these is familiar to me:

  • Aggression and Violent Behavior, 1359-1789, Core
  • Aggressive Behavior, 0096-140X, Core
  • Behavioral Sciences & the Law, 0735-3936, Core
  • Behavioral Sleep Medicine, 1540-2002, Core
  • Behaviour Research and Therapy, 0005-7967, Core
  • Emotional & Behavioural Difficulties, 1363-2752, Core
  • Environment & Behavior, 0013-9165, Core
  • Ethics & Behavior, 1050-8422, Core

 

As I go down the list, I feel overwhelmed – there could be an awful lot of literature out there on this topic! But just looking at the two ways ‘behaviour’ is spelled has reminded me to refine my search to articles published in Britain, thus excluding much of the American literature on ‘behavior’ at least in the first instance.

Looking at the ASSIA Database Guide document is also very useful. (I printed it off. I’m hard-copy at heart!) It states that it’s designed for the information needs of the caring professions and focuses on the practical problems of society. It covers education, as well as health, social sciences, psychology, sociology, economics, politics and race relations. It sounds spot on for my quest – the draft article is about parenting programmes targeted at parents of 8-13 year olds at risk of anti-social behaviour. I’ll read on and see what else I learn.

More good stuff in this Guide – I learn that ASSIA content is drawn from journals of which about 46% are published in the UK (43% in USA) so that’s good for my search needs. I also learn that it is recommended to use the database Author Name Index for author searches and to use its Thesaurus for search term Descriptors. I remember now that Chris Bradford recommended that during a session I attended but I’d forgotten until reminded again today. I learn that the Keywords field in ASSIA is used to search the Abstract, Descriptors and Title – that’s useful to know. In fact, the whole Guide (15 pages with permission to copy) is full of exactly the information I need to go ahead with my search, including how to narrow or extend a search. I’m ready to give it a whirl.

 

Part four: Literature Search Principles

M. Know your database or source. Which journals are covered, for what dates?
More on this: Some journals are covered in multiple sources but not always for the same dates. See also Principle J: you should know the difference between a source or database and a search platform that is the tool used to search. Also learn whether the database you are searching is an abstracting and indexing one, or a full text one: this is discussed here in Principle W, in Part five.

  1. Consider searching for or within particular journals, especially if you know titles that are relevant.
    More on this: Note that the lists of journals in a database can also be useful for authors. Journals that are indexed in the places where you search are likely to be good places for publication, to reach researchers like you.
  2. Allow for spelling variations, eg American/British, plurals, hyphenated words: wildcard symbols may help.
    More on this: Spelling variations can help you to either narrow or widen your search. To widen your search, consider using wildcard and truncation symbols. You should read about the appropriate symbols for wildcard or truncation searching to use in the database that you are searching. For example, Proquest has a guide. There is more on this principle in Part five of this journey.
  3. Look for an index or a thesaurus that you can browse, to find the terms most relevant to you.
  4. When you put your keywords into a simple search box, consider what you are searching through: full text or certain parts of a record?
    More on this: In some databases you are searching the records describing journal articles, as with the database that Mairi Ann writes about. In others you might also be searching the full text of an item (as when you search Google Books), or through the references cited by articles that are indexed.

 

Part five: Some results

I decide to try first with Advanced search. I follow advice and look up the database Thesaurus. Good job I did because anti-social behaviour is ‘Not found’ – the form used is spelled without the hyphen. Also I find out that there is a long list of ‘behaviour’ words. I must be very focused and not get side-tracked into a jungle of by-ways, searching everything ever written on ‘behaviour’! I decide to start as simply as possible by searching ‘Anywhere’ for ‘antisocial behaviour’ AND (England, OR Britain OR United Kingdom), limited to ‘journal articles only’ and ‘English only’. This yields 267 scholars and 161 peer-reviewed journal articles. A quick glance down the first page of the list shows that I need to refine my search further. I go back and add in ‘parent*’. This takes it down to a more manageable 44 peer-reviewed articles.

I scan through this list, clicking to select any that seem relevant. Sometimes I click on ‘View Record’ if the short form information on screen is not quite enough to decide whether or not the article is relevant. Then I ‘Return to Results’ to get back to the list. I end up with 34 ‘Marked Records’ and am able to save this and choose what to do with it in another window – but first I clicked on ‘Please log in to My Research’. I sign up and create a profile. I’m hoping this means that, next time I use this database, I’ll be able to find this particular search for reference. Just in case, though, I elect to save my marked records to a folder on my computer.

This next bit takes me a few hours but I don’t know how to shorten the process. I go to my ‘Marked Records’ list and click on ‘View Record’ for each one. This takes me to ‘Record View’. Here I can click on ‘WebBridge’ to get to the full article electronically. I choose to save the ones I’m really interested in as PDFs on my computer to read and refer to later. I notice that there are lots of options available – e.g. to set up e-mail alerts to particular journals as new articles or new issues come out or to set up RSS feeds - but I decide to try to focus on my literature search and to note these possibilities for exploring more another time. One that I do use, though, is the option to look up articles that have cited the article I’m interested in. This doesn’t always lead to relevant articles for me but it does lead to a few that perhaps I wouldn’t have found easily otherwise. I save these too. I notice that it takes quite a bit of concentration to remember which windows I can close (e.g. to get back out of a particular journal webpage, and back out of the WebBridge window) and which I need to take care to use ‘Return to linked references’. I use the ‘Next’ and ‘Previous’ buttons on the screen to navigate through the records.

As I’m going through this process for my 34 Marked Records, I learn more and more as I go. For example, each Record View shows Descriptors for that article and I can see that some of these would be useful if I dare to edit my search any further. For example I could experiment with using Descriptors such as ‘Interventions’ or ‘Risk factors’ or ‘socioeconomic status’ but I decide not to as my consuming fear throughout all of this time is that I will drown in a sea of apparently endless relevant literature. I need to keep remembering that all I am trying to do is to learn as quickly as I can enough about what is out there to ‘place’ our contribution appropriately in the field. 

I learn that some journals carry more than one article that is particularly relevant to my interests and so I note these and plan to search only these journals another time. I also find that one journal, the BMJ (British Medical Journal), seems very relevant but is not available on WebBridge nor in our own Library (which can be accessed using ‘search catalogue’). I note that I need to Google that journal’s own website in order to access all my Marked Records from the BMJ.

 

Part five: Literature Search Principles

Note the example of a truncation symbol in this search: ‘parent*’: see Principle O in Part four, and remember the constant need to apply Principle G (Part two), knowing what you intend to achieve: don’t get side-tracked!

  1. Combine two related concepts with ‘OR’ and two different concepts with ‘AND’.
    More on this: Many databases work like Google, in taking an implied ‘and’ between keywords. Therefore if you mean to search for a phrase, you may have to use quotation marks or similar, around the phrase. Combining words in this way is sometimes called Boolean searching, and one good guide to Boolean searching is available on the MIT Libraries’ site.Note in Part six that ASSIA’s own guide also explains Boolean searching.
  2. Aim for no more than 60 results.
    More on this: Sixty is probably the maximum number of results that you will want to read through and select for further investigation. When you have more results than you can handle, refine your search by adding another element to your query. When you have too few results, take out an element of your search query.
  3. Create an account and save a list of articles to explore later.
    More on this: You may need to create a personal account that you log into separately, after you’ve already logged in through the authentication process. Use personal accounts to store searches and set up e-mail alerts when new items are added to a database that match your search query, as well as to store lists of interesting articles.
  4. Keep records so that you can find your way back again if you get lost amongst the literature.
    More on this: If you definitely want to change your search topic or the purpose of your searching time, then make a record of how far you got on the original topic and then define your new topic thoroughly from the beginning. (NB, in Part seven, Mairi Ann uses a handy template for recording keywords for each topic.) Literature searching is a journey and you need to keep referring to your directions, and to mark where you’ve been on your map! You could create a spreadsheet (PhD students could turn this into a thesis appendix) and record in it some or all of the following elements:

-         name of the resource searched, eg ASSIA, on CSA Illumina

-         date of your search, eg April 2011 (Keeping up to date: date search was re-run? Email alert set up?)

-         Keywords and limits used (for each search)

-         Number of results for that search

-         Login details or clues about them, for your personal account on that database: in addition to your spreadsheet, do make sure that you save any useful searches or search history, on your personal account

-         Notes on how useful you found the database, the kinds of results you found, etc.

-         Tips for yourself on how to search on that database.

Another recording idea is to print out a copy of your search history when you have finished searching: you can scribble any notes, tips or ideas on the print-out.

Some reference management software enables you to record how you found items that you add to your collection.

  1. Back everything up! It takes a long time to perform such searches, so don’t risk losing track of what you’ve done and found.
  2. W.  Click on links to find the full text of articles, or search the library catalogue.
    More on this: WebBridge is the University of Warwick’s OpenURL resolver, and many other libraries use one called ‘SFX’. OpenURL resolvers are not 100% reliable, as in this example where Mairi Ann could not access BMJ content. The University of Warwick’s Digital Access Manager, James Fisher, commented on the original blog post: “As you have found, this tool often helps to link you to the full-text of articles that we subscribe to. However, sometimes it is not able to do this. This may be because we don’t subscribe to the journal. However, it can also sometimes be due to the information that is passed to WebBridge by the database that you are using. WebBridge tries to interpret this data and check whether we have access to an article. If this data is unclear this may result in WebBridge being unable to check our holdings. For this reason, if you find an article that looks really relevant which WebBridge is unable to find it could still be worth checking the Library Catalogue to see if we do subscribe to it.”
  3. To do a literature search properly, allow plenty of time and shut out distractions.
    More on this: For a search eliciting 60 results, ideally you should allow three hours for defining your topic, searching the database, refining your search, recording your searching and investigating articles of interest. You will then need to allow more time to read the articles that you’ve found, as we see in Part seven.
    It is quicker to find the full text when searching in a full-text database than in an abstracting and indexing (A&I) database such as ASSIA, which Maira Ann searched. You could search the full-text databases first, but unless they are specifically relevant to your subject area they might not include key journals for your research, so you’d have to search in the A&I database as well, and then you’d need to spend time eliminating records you already know about from your A&I results - so I would recommend starting with the database most relevant to your research in the first place, whether it is full-text or A&I.  
  4. When navigating within a database, it is often safer to use the database’s ‘next’ and ‘previous’ buttons than your browser buttons.
    More on this: By using the database’s internal navigation buttons your browser won’t have to send your query to the database again, taking time and potentially stalling, and you won’t go back past a stage at which you selected an article for your list, before the database has registered your tick!

 

Part six: Next steps in my literature search

Having started searching, I refer back to the advice on the CSA Illumina Search tips page: their text appears in bold here.

5. After you've located a couple of relevant documents, consider using the descriptors attached to the record to locate similar records.

I considered this: see Part five. This would be a useful way of extending a search if the initial search came up with a very small number. With my 30-plus Marked Records, I have more than enough on my plate!

Need further assistance? Talk to your institution's reference librarian or take a look at our quick reference card and other search related articles: 

  • Quick Search
  • Advanced Search
  • Boolean Operators (including proximity searching)
  • Wildcard Characters

All these topics are covered in the ASSIA Guide, which I printed off and read.

Now I need to stop searching and read some articles. As I do that, I’ll be trying to piece together a rough shape for the literature in this field – what are the key concerns? How do they relate to each other? Once I have a rough sense of that, I’ll decide where next to take my search. I’ll keep posting as I go.

 

Part six: Literature Search Principles

  1. If you find only a few results, use subject headings, keywords or descriptors in an article’s record to explore wider or related topics. More on this: Subject headings, keywords or descriptors in the records of the results are often hyperlinked searches, so can be quick to investigate.

 

Part seven: Complex search strings!

Yesterday, I did some reading and followed up some of the references in these articles to add to my electronic pile. I’m beginning to create a rough draft of a mental ‘knowledge object’ in my head. I even had a go at drawing it out on the blackboard in my office.

Today, I’ve decided to search for some more literature about parenting programmes as social policy. I go back to ASSIA and create a new search. To help me, I make use of the Search Strategy template given to me by Jenny Delasalle which helps me to think of cognate concepts for ‘parenting programmes’ and ‘social policy’. I decide on (social OR public OR education*) AND (parent*) AND (program?s OR intervention*), limited to (UK OR United Kingdom OR Britain).

That came up with no journal articles so I review that search and decide to take out the AND (program?s or intervention*) line and try again. I try various edits of this search but I get no journal articles coming up. Clearly, I’m doing something silly.

I try again. I go for (social OR public) AND (policy) AND (parent*) AND (England OR United Kingdom). This time, I get 213 peer-reviewed articles. I need to narrow this down to those that are about parenting programmes so I edit the search by adding AND (training OR program?? OR intervention). This produces 55 peer-reviewed articles. I scan through these, clicking the tick box for those I think are relevant. Only 11 are.

I think I’ve still not got the search terms as neatly defined as I’d like so I try again, this time with (social policy) OR (public policy) AND (parent* program??) OR (parent* training) OR (parent* intervention) AND (England OR United Kingdom). This time I get one solitary article and it’s one of the previous 11. I’m perplexed – I know this can’t be right. Ah, lightbulb moment. I go back to the articles I was reading yesterday and look at the keywords they used. Instead of ‘social policy‘ they have been really specific and used ‘parenting orders’; instead of ‘parent programmes’, ‘training’ or ‘interventions’, one article has used ‘parenting support’. I try again ...

 

The Librarian’s concluding comments

Now we are reminded of the very first principle in Part one of this journey: when you already know the results you are finding, then it’s time to stop searching. Perhaps our researcher could move on to try her search queries on other databases (Principle F in Part one). We can also see how literature searching is not linear but made of many circles, as the articles that you find suggest ways to define your topic or refine your search further.

What this final part of the search journey shows is that our researcher is tenacious, patient and resourceful, which will no doubt be good qualities for other aspects of research beyond literature searching!

 

The Researcher’s concluding comments

This journey exploring online tools for literature searching has taught me useful principles that I have applied in all my searches since. The article on which I was working will be published in the British Educational Research Journal, volume 39, number 6, in December 2013. It is available online – DOI: 10.1002/berj.3020