What's it all about?
In a nutshell, a bibliometric analysis is essentially a citation analysis with added value. A list of authors was provided, all I had to do was find out how many times each of the publications were cited and provide a list of highly cited publications and a list of highly cited publications in high-impact journals. The h-index for each author is then calulated which averages out the number of cited works across all publications (this prevents a skew from an author who may have had one single article cited hundreds of times, giving a fairer representation of how popular the author is).
How is it done?
I used Web of Science (WoS) which had the added feature that it calculates the h-index for each author and I could dip into the Journal Citation Reports database for information on journal impact factors. Other databases are available - Scopus, CiteSeer and even Google Scholar can assist. Which one you use depends upon subject coverage and access. For more background info see:
- Kloda LA. Use Google Scholar, Scopus and Web of Science for Comprehensive Citation Tracking Evidence Based Library and Information Practice 2007 vol 2 (3) 87-90.
How difficult is it?
It isn't difficult at all, that's the beauty of it. However, it can be time consuming if you are not familiar with the database (luckily a lovely librarian from lis-medical took me through the Web of Science on a mini-tutorial over the phone with back-up from Steve at Christie!). Even if you are aware of the database, finding the authors is the easy bit, you still need to be aware of things like:
What type of publications are they wanting to include? - references to meeting abstracts are included in WoS (but not usually cited at al by others - this would drag the h-index down).
Each reference had to be checked - that it was the correct author, nightmare if you have an author with a common name like Smith! In WoS however, you can cross-reference with institutions, addresses, topic, etc.
How do you define highly cited?
How do they want the results presented? - I found it impossible to email the references in the format I required as I had to add in additional data (journal impact factor) and PubMed IDs. I exported the refs into a text file and then imported the file into excel (again a bit of a pain at first as I don't use excel in this way at all and had to familiarise myself with it all).
Benefit for users
For my user it provided him with a list of all the highly cited publications and publications in high impact factor journals from all the researchers involved in a project to submit an application to become a Centre of Excellence. It provided information on numbers of publications, where people were publishing their research, whether the work was being cited and the trend of publication from this particular research group (WoS provided nice little graphs which indicated an upward trend of cited works for the authors).
Overall impression
Given the time frame I was working in, it was a steep learning curve in how to undertake a bibliometric analysis, however, it's certainly not a difficult thing to do, just time consuming, and can help provide impact evidence for researchers, authors, clinicians on cited publications.
Aintree University Hospitals is a Foundation Trust, therefore the potential is there for this type of impact evaluation work. I'm now going to actively promote this type of work as a service I will offer, targeting people in research groups and those looking to submit for funds for various projects.
Did I look myself up as an author?
Yes of course I did!!! But I don't think HILJ has made it into WoS...yet...that's my excuse anyway!!
