Publish or perish? But always with impact
There's more to judging impact than the size of your impact factor, argues Nicola Beesley.
We’re all told our research must have impact, and as a veterinary researcher early in my career this is something that already plays on my mind. But how is the impact of research measured? Are these measures fair?
When considering academic impact (rather than economic or
societal), publications are often used as a measure of success - and the higher
the journal’s impact factor the better, right? I for one am not so sure.
Bigger isn’t always better
When choosing where to submit papers for publication, I have found that
impact factor plays some part in your decision, but I feel picking a journal
where a paper will be read by the appropriate audience is equally as important.
For example, most vets in clinical practice tend to read veterinary specific
journals, e.g. The Veterinary Record (2012 impact factor 1.803 compared to
Nature’s 38.597). Even with the advent of open access journals, from personal
experience, these are rarely read in veterinary practice.
Let’s consider how impact factors are calculated. In any given year, the impact factor of a journal is the average number of citations received per paper published in that journal during the previous two years. So an impact factor of 1 means that on average articles published up to two years ago have been cited once. Or does it?
In 2005 a self-evaluation by Nature [1]
discovered that 25% of papers contributed to 89% of its impact factor for 2003.
And in 2004 the majority of Nature's papers received fewer than 20
citations. Clearly some papers will be cited numerous times: including
publications of genomes, new experimental techniques and software programs. Do
these outliers make a difference? Perhaps not, they may just cause a “blip” in
the impact factor of a journal for a year or two. A further evaluation of a sample of 100
journals (including physical, chemical, biological, earth sciences and
engineering journals) found an r2 value of 0.94 between impact
factor and five-year median (i.e. a measure that is robust to outliers) of
citations [2].
Let’s consider how impact factors are calculated. In any given year, the impact factor of a journal is the average number of citations received per paper published in that journal during the previous two years. So an impact factor of 1 means that on average articles published up to two years ago have been cited once. Or does it?
It’s worth also remembering, that retracted papers still
contribute to impact factors [3].
Indeed, I can see why examples of poor or controversial research might be cited
to emphasise the value of one’s own research.
The 2005 Nature self-evaluation further revealed a
difference in citation rates between disciplines with papers relating to cancer
and molecular and cell biology cited far more than those relating to physics,
for example. This probably extends to sub-disciplines as well.
Finally, it’s interesting to note that when “independent”
parties have tried to replicate impact factors, they have found discrepancies
between their values and the published values [4].
Proceed with caution
The European Association of Science Editors released a statement relating to
impact factors in 2007 and recommends that they are used:
only – and cautiously – for measuring and comparing the influence of entire journals….not for the assessment of researchers or research programmes' [5]
Eugene Garfield, who first mentioned the idea of an impact
factor in 1955, discusses many of the issues I’ve talked about in an article
from 2006 [6].
None of his arguments particularly convince me that the positives of impact
factors outweigh the negatives (but you may feel differently!) In his
conclusion he quotes Hoeffel who states that:
impact factor is not a perfect tool to measure the quality of articles, but there is nothing better, and it has the advantage of already being in existence and is therefore a good technique for scientific evaluation [7].
This statement doesn’t really fit with the philosophy I have
as a scientist – surely we should be striving towards alternatives. Alternative citation measures which can be
used to rank scientists against each other include h-index, m-index, and the
g-index. Indeed these are just a few
examples, but as a report by the Joint Committee on Quantitative Assessment of
Research discusses these too are by no means perfect [8].
Wider impact
There are, of course, other impacts
research can have apart from publication. The University of Liverpool is one of
only a handful of veterinary schools in the UK and the majority of teaching is
given by staff involved in cutting edge research. This means that new graduates
(105 this summer) go out into practice armed with up-to-date knowledge. There
is a steep learning curve for new graduates, but many “older” vets appreciate
them as a source of new knowledge.
Vet students also carry out research with us, or work in the
various diagnostic laboratories and animal hospitals during their extramural
studies (EMS). Veterinary surgeons in practice come back to the University for
CPD and postgraduate qualifications. We also actively engage with the public to
talk about our research, such as as food safety outreach events and art exhibitions.
Indeed, the new(ish) concept of “Altmetrics” uses other measures, not just
citation counts, to measure impact. These include number of views and downloads
of articles, discussion in the news or on social media, and recommendations of
the articles to others [9].
Perhaps these, or a variation of these, will be a better way of measuring all
the different ways research can have impact.
This is by no means an exhaustive review of the papers and
opinions that exist on impact. However, in true scientific spirit, I am optimistic
that my research will be judged not only on the impact factor of the journal it
is (hopefully!) published in, but also the alternative means I have at my
disposable to generate IMPACT.

Lots of citations do not necessarily equal quality work. According to Google, the retracted article by Wakefield et al. on the link between MMR vaccine and autism has 1,066 citations! Most of them of course say things like: “This contradicts an earlier publication by Wakefield et al.”. I suppose it certainly had impact, just not in a good way…
ReplyDelete