Discussion Thread: AI Research Tools and Proper Use AI Research Tools
PowerPoint 5 Searching and screening
Chapter 5 from
How to do your literature review
This is PowerPoint 5 for Chapter 5 …
We will look at
Using academic search engines
Using AI
Snowballing
Using databases
Using keywords
Screening and selection
All of this is discussed in much more detail in Chapter 5 of How to Do Your Literature Review
Using academic search engines and AI
Google Scholar is the most commonly used academic search engine, though there are others.
It searches for literature based on your query.
The query may be a name, a topic, or an article’s title.
It will categorise and store references for you, as well as searching.
It is not as comprehensive as subject-specific databases.
Authors’ names can be used for snowballing
Full versions of an article often are available here.
‘Cited by’ is a good indication of how much impact the article has had.
If you click on this, it will take you to all of the works that have cited it.
Snowballing
If a search from a search engine comes up with a really interesting ‘find’ that is spot-on your area of interest, you can snowball from it.
This involves either …
finding material that the authors themselves drew on (backward snowballing), or …
going forward in time to material where other authors have cited your key reference (forward snowballing).
To forward snowball
Click on ‘Cited by’ under the article in Google Scholar, which will give a list of all the other works that have cited this article.
Look through these articles that have cited your key article.
Do any look especially interesting?
Have any been highly cited?
To backward snowball
Find the article online (you can do this from your search engine search)
Now, examine the article’s reference list to see if there are articles or books that look to be of particular interest to you.
Follow these up with a separate search.
Other features of snowballing
Borrow the keywords of your key article for further searches.
Look to see which journals are carrying papers on the topic you’re interested in.
Organising your work with Scholar
Each Scholar ‘find’ has a little star under it. If the reference is useful to you in some way click on the star and it will save the reference to your personal Library.
Google Scholar creates a ‘Library’ for you as soon as you click on the little star
Click on ‘My Library’ at the top right or top left of the screen and you’ll see the reference for which you clicked the star is now installed there.
Repeat for each reference you want to save, and you’ll accumulate a library of references.
Organising the Library
Open your Library (click My Library, top right or left of screen)
Each reference you have put in there has a little icon underneath that looks like a tie-on suitcase label
Click on the icon, and it will give you the option to ‘Create new’. You can now label the reference in any way you wish. You may want to label this with, for example:
a label for topics (e.g., media, films, newspapers, social media)
a label for closeness (e.g., general, close, specific)
Using AI
Elicit.com will find relevant papers and summarise them
Semantic Scholar is similar to Google Scholar, but allows you to target finds according to fields of study and whether the found articles have a pdf.
Consensus.app is similar to Elicit.com.
ChatGPT will help you to brainstorm research questions and recommend databases.
Bing AI is similar to ChatGPT.
Connected Papers
… will draw a diagram of its finds and show how they relate to each other
There are positives and negatives to AI
Positives to AI
AI machines are good at finding relevant papers and those that are maybe tucked away in a remote corner of the literature somewhere.
They are great for snowballing backwards and forwards.
They review a range of databases (eg PubMed, ERIC, PsycInfo) saving you the trouble of going to each individually.
Negatives to AI
They may miss some obvious finds and connections.
They may latch on to a word in a search query or title that isn’t centrally relevant and go over the top on offering ‘finds’ based on that word.
They are not good at telling a story. Even when they claim to be offering some kind of digest, this is more like a list – albeit a list that is made to look like prose – than a narrative. They are bad at connecting ideas.
Relying on them too closely can mean that the form, structure and priorities of the existing literature can be privileged over a thoughtful review of what is important for the issue that you are exploring.
Because they seem to be doing all the work for you, they may discourage you from reading around and getting an impression, a ‘feel’, of what the literature is saying.
Databases
A library database contains huge amounts of information, organised so you can find it easily.
The information has been organised by fields – keywords, authors, titles, etc. – and these fields of information can be searched.
You’ll find databases under ‘Databases’ in your library website.
Commonly used ones are
ProQuest
Web of Science
Scopus
PubMed
British Nursing Index
Business Source Premier
ERIC
Sociological abstracts
CINAHL
PsycINFO
OpenGrey
Library databases will find much more targeted information on your topic of interest than a search engine is likely to find.
A page from PubMed where the search query was ‘effects of aspirin on heart disease’
How to use keywords in a database search
Create a list of centrally important words for your question.
Be aware of synonyms or related terms that may be used in the literature.
If necessary, adopt the ‘controlled vocabulary’ used in most databases. The database will provide these or subject headings (e.g., MeSH terms in PubMed).
If necessary, use Boolean operators (AND, OR, NOT), and wildcards to combine keywords and refine your search.
Enclose a phrase in quotation marks to search for an exact phrase. For example, use “child carers” to search for this exact phrase.
Record your search strategy, including the keywords used, so that you can replicate or modify the search in future if needed.
Screening and selection – thinning down your trawl of finds
You may apply different eligibility criteria, sometimes called ‘inclusion and exclusion criteria’ for your finds.
For example, you may wish to include only those articles published within a particular timeframe.
Or to restrict your review to pieces reporting on work using a particular methodology.
Or to look for the most highly cited articles.
Screening and selection is an iterative process
You go backwards and forwards, refining and revising as you go.
Search
Refine with eligibility criteria
Summary
There are many invaluable software and online resources that help you to find, store and organise material.
Google Scholar helps to find material and to organise it in your own library.
Artificial Intelligence (AI) tools such as Elicit.com, Connected Papers and ChatGPT can help in the identification and summary of literature sources but will not satisfactorily integrate the various ‘stories’ in the literature.
Library databases and other library resources are invaluable and should not be downplayed with the advent of AI, since they can provide a far more targeted and deep-reaching search.
Always feel free to ask a librarian – online or in person. They want you to ask questions. They know everything, and more, about databases.
Keywords and search limiters form an essential part of screening the large number of finds you will most likely make from database searches. It’s worth learning how to use them effectively.
Searching is not a one-off process. It’s iterative: what you learn from your first trawl through the literature informs the next bout of searching, and the next.
Activity
Open your library webpage and login.
Find the heading that says ‘Database search’ (or something similar) and click on it.
Type in the name of your subject or area of interest for a database.
Find a database that appeals to you, on any topic that grabs your interest.
Enter a search term; you will probably be taken to a more specialised or relevant database.
Now enter various keywords and search terms of your own choosing to see what happens and how much you discover.
Note interesting findings or anomalies and discuss with others.
image1.jpeg
image2.jpeg
image3.png
image4.svg
.MsftOfcThm_Accent1_Fill_v2 { fill:#4472C4; } .MsftOfcThm_Accent1_Stroke_v2 { stroke:#4472C4; }
image5.png
image6.jpeg
image7.jpeg
image8.jpeg
image9.png
image10.svg