assess the impact of researchers, their h-index and the sum of the number
of citations of their three most cited publications. The data show that the
h-indices are normally distributed, but the total numbers of citations seem to
show more of a power law distribution, with a few authors or publications
attracting large numbers of citations, such as Rod Ellis’s (1994)‘The study of
second language acquisition’reaching a impressive 9571 citations. The high-
est total number was found for Stephen Krashen (21,396). Rod Ellis had the
highest h-index score (73).
To what extent the citation data reflect real impact as perceived by the
field, becomes clear when the data on“leaders”are compared to the citation
data. The correlation is moderately high: leaders have above average citation
scores, though the interviews also show that the definition of leader is not
uniform, as mentioned in Chapter 4. Clearly, academic status is an important
component.
In contrast to Garfield’s (1994) concern that editors of journals may prefer
review articles because they attract more citations, the data show that there
are only a few review studies that attract many citations. Numbers of citations
for overview articles versus data-driven articles do not differ significantly.
In AL, we rarelyfind articles with more than three authors, and there
seems to be a preference for single-authored publications. This is probably a
remnant of the old humanities tradition, in which scholars worked on their
own. This is in contrast to the situation in the hard sciences, where project
based research is more common. The analyses showed, however, that single-
authored publications are cited more frequently than multi-authored pub-
lications, but the difference is small and cannot lead to the conclusion that,
in general, publishing in isolation is more effective.
When comparing publishing in books or journal articles, the question is
whether, in terms of citation scores and h-indices, the investment in a whole
book is warranted. It could be argued that the time investment for a book is
roughly the same as forfive journal articles. A simple calculation shows that
books are not citedfive times more often than articles.
Contrary to expectations, older publications are not cited more than more
recent publications. The correlation between year of publication and
number of citations is not significant. This is partly explained by the tendency
to refer to more recent publications.
Over the last 15 years, impact factors of journals have become important
in decisions on where to publish and on what counts. In the medical faculty
of my university, researchers are encouraged to publish only in the top 25
percent journals of their specialization, and discouraged to publish in lower
ranking journals, in the sense that their resources for research will increase
with publications in the top 25 percent and decline with publications in the
lowest quartile. This harsh regime does not seem to apply equally in AL,
though the pressure to publish in good journals, often at the expense of
publishing books, is mentioned frequently by the informants. As Robert
DeKeyser remarked:“Impact factors may lead to undesirable publishing
The citation game 119