Posted in High-impact journals, Web 2.0
Surviving in science these days is all about high impact. How is this impact being measured? Managers, deans, operators, science editors and grant officers, to mention just a few non-active scientists, know the answer exactly. They judge the scientist by the:
- number of papers published in refereed journals
- number of papers in high-impact journals
- number of citations, and more specifically by the h-index
To remind you: if the h-index of a scientist is 20 the scientist has coauthored 20 papers with at least 20 citations.
The majority of scientists that have a AAA-rating according to the above criteria find these criteria good indicators. The majority of scientists that have a poor rating on this scale will find all type of arguments why the above rating says nothing about the quality of a scientist. In my opinion the criteria only test marginal quality, that is to say scientists that have an exceptionally low rating on all three of the list, are not performing well.
Of course there are a number of possible comments to make on the criteria, like neglect of age of scientist, neglect of number of coauthors, difference in social status of scientist (director of a hierarchically organized institute), difference in culture of scientific fields. The purpose of this post is however, we have better get the numbers right, before we discuss their relevance.
ISI Web of Knowledge
The ISI Web of Knowledge had a monopoly on supplying the above three numbers. However, there are a lot of problems with the database of the ISI:
- Many people interested in science and scientists do not have access to the Web of Knowledge, as one has to pay for this access. Scientists working in universities in western countries tend to forget that even a lot scientists, notably those not working in western countries, do not have access to these data.
- The ISI data is of poor quality as it is quite difficult to disentangle authors that have common names. Otto wrote a great article on the ReseracherID as a solution to this problem.
- The web site is if poor quality, awkward to use and with strange capitalization of names.
As a result the h-index of a scientist inferred from the Web of Knowledge is often polluted, partly caused by the fact that the person who extracts that number does not know how to correct for theses flaws.
A number of commercial enterprises is to compete with the ISI and were until recently not very successfully with it. Years ago I checked Google Scholar and discovered that my own publication list was very poorly represented. That impression stuck in my mind until a few weeks ago. A colleague of mine stated that Google Scholar and in particular the Citations part was excellent. I did not agree, but it made me check Google Scholar again and I was deeply impressed. Google’s algorithms are so good that they know that “de Vries” and “DeVries” can be the same author. ISI, without ResearcherID, cannot make any connection between them.
The quality of Google’s representation of my publication list is extremely good and better than the ISI. In addition the formatting is much more pleasant.
Google Scholar’s privacy
Google is fighting worldwide with authorities about privacy violations, but in the case of Google Scholar the giant from California is unfortunately restrictive: Scientists have to make their Google Scholar profile public if they want other people to be able to have read access to it.
So my advice to scientists: make your Google Scholar profile public. Many journalists and many applicants for junior science positions will love it. If Google Scholar Citations will be a success ISI will have to reconsider its business model and become free, open access as well.
I will be honest about my secret motive. If the h-index becomes so readily available many more scientist will play the card of h-index optimization (HIO) up to a point where the h-index becomes meaningless.
You want an example? Just check my Google Scholar Citations profile