In this study we continue the application of Data Envelopment Analysis (DEA) to assess the efficiency and effectiveness of
the R&D effort of European countries. We use GDP, active population and R&D expenditure as inputs, and publications and patents
as outputs. Being effective means that, in order to obtain a maximum efficiency score countries are forced to perform on every
output goal. A discussion of each country's performance and a comparison with May'sScience results concludes our analysis.
It is shown that Data Envelopment Analysis (DEA) ca be used to construct relative scientific and technological indicators.
The method is explained and illustrated using countries as objects of study; GDP, active population and R&D expenditure as
inputs, and publications and patents as outputs. Using these parameters the efficiency of countries is assessed.
This paper gives a mathematical technique to study influences, using citations. Taking into account both the publications that have a direct influence and those that have an indirect influence, we obtain the total influence measure on a fixed paper.
We give an upper and a lower bound for the slope, on a semi logarithmic scale, of the cumulative graph of a data set, such as a bibliography, originating from the disjoint merging of two similar data sets.
Egghe's continuous information production processes (in short IPP's) are described using category theory. Therefore, we first review the main ingredients of this mathematical theory, introduced byEilenberg andMac Lane more than four decades ago. Then we show how the notion of duality, as used byEgghe, can be placed in the abstract framework of categorical duality. This leads to a natural isomorphism involving the identity functor on a category of continuous IPP's. This natural isomorphism is completely similar to the well-known natural isomorphism between a finite-dimensional vector space and its double dual. We further show that to develop Egghe's theory on IPP's one needs no other intervals than the unit interval.
Using the artificial example of perfectly stratified samples, we have shown the effect different sampling designs have on the determination of concentration values. More concretely, we have studied the following four cases: sampling of items in the case the number of sources is known (we have further considered the cases when there are many items in every source and when this is not so); sampling of items in the case the number of sources is unknown, and finally, sampling of sources.
It is argued that Leydesdorff's theory of citations mixes the ideal or pure case with complicating factors. Ideally, citations
are used as shorthand and for ethical reasons. The social network between scientists should be seen as a second-order correction
on the basic model or, sometimes, even as noise. Metaphorically speaking Leydesdorff's theory is not a theory about ideal
gases, but about polluted air.
Temporal differences in self-citing and self-cited rates of journals are studied. It is concluded that the citation curve
of a journal is composed of two curves with different characteristics: a self-citation (or self-cited) curve and a curve representing