analiza a a cercetarii stiintifice intr-o univ din italia - en

8/8/2019 Analiza a a Cercetarii Stiintifice Intr-o Univ Din Italia - En

http://slidepdf.com/reader/full/analiza-a-a-cercetarii-stiintifice-intr-o-univ-din-italia-en 1/10

FEATURE ARTICLE

Scientometric analysis of national university research

performance in analytical chemistry on the basis

of academic publications: Italy as case study

Anna Annibaldi & Cristina Truzzi & Silvia Illuminati &

Giuseppe Scarponi

Published online: 26 May 2010# Springer-Verlag 2010

Introduction

Citation-based scientometric indicators, for example impact

factor (IF) of the Journal Citation Reports ( JCR) from the

Institute for Scientific Information (ISI) database of Thom-

son Reuters [1], have become common (though not

uncontroversial) tools for evaluation of the research perfor-

mance of individual scientists, departments, faculties, and

full universities, other research institutions, or even nations

and continents [2 – 8]. Although their crude use as sole

indicators of research performance has been questioned [9 –

12], particularly when original IF values across disciplines

are considered without some normalization procedure [4,

13 – 22], it is generally acknowledged that citations and

impact factors indicate (are correlated with) the scientific

quality of journals [23] and that they can in fact be used,

together with other indicators, for informed peer review; and

the practice is indeed expanding because of the increasing

frequency of requests from governments [10, 17, 22, 24].

At an international level, to quote only one example,

citation and co-citation analyses were introduced long ago

by the Organization for Economic Co-operation and

Development (OECD) [25, 26] and now they are always

used for biannual evaluation of national performance in

several fields of research [25 – 27].

Several scientometric papers have been published

concerning: national and regional research performance

in analytical chemistry [28 – 33], analytical chemistry in

the European Union (EU) [34 – 36], the evolution of

quality in analytical chemistry journals [37], the evolution

of automation in spectroscopy [38], hot topics in global

analytical chemistry research (100 most cited papers) [39],

and mapping the world of analytical chemistry [6, 40 – 42].

However, there has been no evaluation of the academic

performance of a nation in analytical chemistry through

scientometric analysis of the publications of university

professors.

Stimulated by the request of the Italian Minister of

Education, Universities, and Research to the Italian Na-

tional University Council (CUN) to suggest criteria for

minimum requirements in the evaluation of the scientific

activity of Italian university professors, to be used, e.g., as a

guide in academic competitions [43 – 45], a census was

carried out, using the SciFinder database [46], of the papers

published by (full) professors ( professori di prima fascia) in

the scientific sector of analytical chemistry in Italian

universities.

Evaluation of the quality (or more correctly the impact)

of publications has been carried out using a proxy required

by the CUN, i.e. the journal impact factor reported in the

JCR [1] after proper normalization of all IFs to the scale of

the ISI category of “Chemistry, Analytical” (CA). The

distribution of papers within journals according to publica-

tion frequency and between the various research subject

areas is reported, together with summary statistics on IF

distributions on the basis both of all the publications

retrieved and distinguishing between senior and junior

professors. The scientific output of each professor has been

characterized by reference to his/her h-index according to

Hirsch [47], which was computed from citations of papers

Electronic supplementary material The online version of this article

(doi:10.1007/s00216-010-3804-7) contains supplementary material,

which is available to authorized users.

A. Annibaldi : C. Truzzi (*) : S. Illuminati : G. Scarponi

Department of Marine Science,

Polytechnic University of Marche – Ancona,

Via Brecce Bianche,

60131 Ancona, Italy

e-mail: [email protected]

Anal Bioanal Chem (2010) 398:17 – 26

DOI 10.1007/s00216-010-3804-7

http://dx.doi.org/10.1007/s00216-010-3804-7

http://dx.doi.org/10.1007/s00216-010-3804-7



by individual professors, obtained from the Web of Science

(WoS) database of Thomson Reuters [48]. Also in this case

results are summarized and discussed considering senior,

junior, and all professors together.

Although the admittedly national objective of this study

is to evaluate Italian academic performance in analytical

chemistry, this work should be of more widespread interest

as it provides a method for judging the scientific results of other countries in the same or other disciplines. In any case

it is to be stressed that we have not attempted to extend the

work to wider regions (for example Europe or even part of

the Continent) for two reasons. The first concerns the

difference in academic structure between nations, which

would have made it difficult to produce comparable data.

The other is that the amount of data to be selected, retrieved

and analysed would have been discouragingly great. Indeed

here we propose, for the first time, a general method for

evaluating national academic research in analytical chem-

istry and we apply it, in this instance, to Italy, for which we

have knowledge and competence.

Data source and processing

Databases

In February 2009 we scrutinized all the publications

produced during their professional lifetimes by the 80

Italian university (full) professors (in Italy called professori

di prima fascia, the highest permanent university teaching

position) of analytical chemistry (the scientific sector is

called Chim/01 in Italy). Sixty-five of the group were

senior (full) professors (in Italy called professori ordinari),

and fifteen were junior (full) professors (in Italy professori

straordinari, this is the first position, which generally

extends for three years after appointment). The research

was carried out through the SciFinder [46] bibliographic

search engine, which sifts through two databases, i.e.

CAplus (from 1907 to the present) [49] and MEDLINE

(from 1950 to the present) [50]. After deleting duplicates,

homonyms, conference proceedings, patents, and local

publications of minor relevance, the research papers were

arranged in their journals of provenance, ranked in

decreasing order of publication frequency.

For all ISI journals, i.e. the great majority, the JCR

impact factor (published by Thomson Reuters for 2007, the

latest year available in February 2009) was also retrieved as

a proxy of the impact of publications. In case of a journal

discontinued but merged with another, the papers published

in the former were added to those of the latter and assigned

the same IF. For non-ISI journals, calculations were carried

out in two ways, firstly by assigning them IF=0 and second

by excluding them altogether.

Referring to the h-index [47], for all the publications

retrieved, citations were obtained from the Thomson Reuters

ISI Web of Science database [48], considering all the years

available (1900 to present), to cover the full period spanned

by all the professors’ publications. Because the Italian

University system’s subscription to WoS includes only the

period 1990 to present, the procedure on the WoS page

“Cited Reference Search” (instead of the page “Search”) wasused to extend the search to all available years. Using this

procedure cited reference variants and incorrect citations were

recognized, corrected and included in the counting of the

proper papers to achieve unified counting. Unfortunately, use

of this procedure does not enable self-citations to be

discovered and taken into account. However it has been

suggested that the effect of self-citations on h is much smaller

than on the total citation count; in fact they are irrelevant both

for papers with<h citations and with many more than

h citations [47]. Given the above limitations, no attempt has

been made to exclude self citations. The h-index of individual

professors was easily computed from their retrieved citationsas the number of papers with citation number ≥h, and used as

a measure of cumulative achievement of each of them up to

the present. To obtain an international comparison, following

the idea of the extension of the h-index concept to groups of

individuals [47], research groups [51], institutes [52] and

countries [53], a retrieval from the WoS database (1990 to

present) has been carried out by searching for the topic

“Analytical Chemistry” and refining results according to the

option “Countries/Territories”.

The scientometric analysis was carried out in two ways.

At the beginning all professors were considered together

and the elaboration referred to all the retrieved publications.

Subsequently, in order to evaluate whether there was a

difference in publication impact and scientific output

between senior professors and junior professors, the

scientometric analysis was also carried out separately for

the two categories of professors.

All the calculations were carried out on two datasets: the first

referring to all the publications together, the second using the

average IFs computed from the publications of each professor.

Normalization

As expected, in view of the wide and interdisciplinary and

multidisciplinary scientific interests of analytical chemists,

a substantial number of papers were published in journals

outside the JCR category of “Chemistry, Analytical”. This

situation is considered quite normal by the analytical

chemistry scientific community given the high interest of

the sector for applications in other disciplines in works

where the analytical contribution could be important or

even essential. Not to mention the cases of the always

exhorted multidisciplinary works.

18 A. Annibaldi et al.



However, it is well known that direct comparison of the

IFs of journals of different disciplines or different JCR

subject categories is inadequate [4, 13 – 22], partly because

of the size of the scientific community of each discipline,

and much more because of the diversity of citing behaviour,

i.e. the so called “citation density”, among disciplines [4,

13, 21]. For example, in the field of chemistry, an

evaluation carried out during the eighties (from 1983 to1992), showed that remarkable variability was present

between chemical disciplines, with similar IF values

(mean±SD) for “analytical” (1.2±0.1) and “general” (1.16±

0.05) chemistry, rather lower values for “applied” chemistry

(0.51±0.07), and much higher values for “ physical” (1.7±0.5),

“organic” (1.8±0.6), and “inorganic and nuclear ” (1.9±0.5)

chemistry [37].

Then, in order to make the IF data to be analyzed as

homogeneous as possible and comparable across catego-

ries, a normalization was carried out, slightly modified from

the procedure of Sen [54] to rescale all IFs to the scale of

the category “Chemistry, Analytical” as follows:

nIF cat : anal ¼ IF j

max IF cat : j

max I F cat : anal

where, nIF cat.anal is the normalized IF rescaled to the CA

category, IF j is the JCR IF for journal j, maxIF cat.j is the

maximum IF value for the JCR category to which journal j

is assigned, and maxIF cat.anal is the maximum IF value for

the JCR category CA (i.e. 5.827 in 2007). In practice, this

normalization procedure is that of a maximum scaling, with

the maximum set at the highest value of the CA category. In

this way, differently from the Sen normalization [54], which

translates data on the 0 – 10 scale, the original scale of the

CA category is here retained and easier comparisons are

possible with literature data. In cases where the JCR

assigned the journal considered to two or more categories,

the maxIF cat.j value was taken from the category with the

highest maximal value.

Concerning the number of citations and the h-index,

unlike the journal IF, these are absolute measures for which

no corrections or normalizations are possible to take

account of the differences in citation intensity and density

across disciplines and over the years. It has been observed,

for example, that h-indices in biology and life sciences are

much higher than in physics, and that further research is

needed to understand distributions and correct for differ-

ences in different fields of science [47]. It has been

proposed that a normalisation could be obtained by

dividing the h value by the average number of authors in

the h publications, on the assumption that differences

between disciplines arise from differences in the number

of authors which could produce future self-citations [55].

However this method has not so far been adopted and no

such normalisation has been carried out here, because the

procedure is not directly implemented on WoS.

Because the h-index continues to increase over time

(even after the scientist stops publishing), to compare

professors of different seniority, we followed Hirsch’s

suggestion [47], and computed the variable m=h/ y (where

y=number of years of activity or scientific age).

Data analysis

This section can be found in the Electronic supplementary

material.

Results and discussion

Distribution of papers

This investigation led to a collection of 8,529 records of

papers (86% from senior professors) published in 630 journals. 8,032 papers (94% of the total) were published in

ISI journals, 4,689 of which (55% of the total) appeared in 53

of the 70 journals included in the JCR category “Chemistry,

Analytical” (journal coverage of the category approx. 76%).

2,534 papers (∼30% of the total records) were published in

analytical general scope journals, i.e. devoted to all the

aspects of analytical chemistry, whereas 5,995 papers (∼70%)

were published in well focussed, thematic journals.

The distribution of publications according to selected

categories of scientific disciplines is shown in Table 1, from

which it can be seen that the major number of publications

(1,112, 13.1%) pertains to the category of “separation

science” followed by ∼12% each for the “environmental”,

and “electroanalytical” chemistry sectors; after these come

“inorganic”, “solution equilibria ”, “agricultural and food”,

“sensors and biosensors” (between ∼7% and ∼5%), followed

by “ physical”, “industrial”, “general”, “medicinal”, “spectros-

copy” and “thermochemistry” (at ∼3%) and by “clinical and

biomedical”, “chemometrics”, “mass spectrometry”, “ bio-

chemistry and bioanalytical”, and “surface interface colloid”

chemistry (at ∼2%). Marginal contributions («1%) relate to the

subject areas of “crystallography” (no.=37), “instrumenta-

tion” (no.=16), “radioanalytical and nuclear ” (no.=11),

“supramolecular ” (no.=9), and “forensic” chemistry (no.=3),

and also within the non-chemistry areas, with “Medicine” and

“Physics” contributing ∼1% each and even less for “Biology”

(0.6%) and “Geology” (0.06%).

Note that the thematic distribution presented here is to be

treated with caution, because a substantial proportion of the

papers were assigned according to the journal name (and its

field of primary interest) because it was almost impossible

to analyze all the titles and summaries of the papers in order

to subdivide them among the various subject areas.

Scientometric analysis of national university research performance in analytical chemistry 19



Impact factors

The list of the journals ranked according to the publication

frequency, together with the IF values available for 2007

and the calculated nIFs, is reported in Table S1 (Electronic

supplementary material). All ISI journals are shown in

capital letters, according to the JCR abbreviations, and

those pertaining to the category “Chemistry, Analytical” in

boldface. The non-ISI journals are reported in lower-case

letters and abbreviated according to the Chemical Abstracts

Service (CAS) or other international abbreviations if not

present in the CAS.

Comprehensive summary statistics of normalized and

not normalized impact factors are reported in Table 2, in

terms of average, standard deviation, median, 1st – 3rd

quartiles, minimum – maximum. Results obtained from IFs

of all publications together and from average IFs calculated

for each professor are reported separately. In this table data

are given both for all professors together, and separately for

senior and junior professors. As regards non-ISI journals,results are reported of calculations for which they were

assigned IF=0 and of calculations from which they were

excluded altogether.

Because the results obtained after exclusion of non-ISI

journals were not significantly different from (only very

slightly higher than) those obtained after assigning them

IF=0, all the results reported throughout the text (with the

exception of those reported in Table 2) refer to computa-

tions carried out assigning IF=0 to non-ISI journals.

Figure 1 shows the frequency distributions of normalized

and not-normalized impact factors referred to all the

publications. A three-modal distribution, skewed towardhigh values, can be observed with modes for both nIF and

IF at approximately 0.8, 2.8, and 5.2, respectively. Among

the journals of the JCR category “Chemistry, Analytical”

the first mode is especially characterized (Table S1 in

Electronic supplementary material) first of all by the Italian

Ann Chim-Rome but also by other international journals

such as Thermochim Acta, Chromatographia, Anal Lett , Int

J Environ An Ch, J Liq Chromatogr R T , J Chemometr ,

Archaeometry, Accredit Qual Assur , Rev Anal Chem. The

second mode is characterized by the most important

European journals of analytical chemistry, for example

Anal Chim Acta, J Chromatogr A, J Electroanal Chem,

Talanta, Anal Bioanal Chem, Analyst , J Pharmaceut

Biomed , Electroanal , Rapid Commun Mass Sp, Sensor

Actuat B-Chem, J Sep Sci, and many others. The third mode

is especially characterized by Anal Chem, Biosens Bioelec-

tron, Trac-Trend Anal Chem.

An overall summary representation of the impact factor

distributions is reported using the box-plot tool in Fig. S 1

(Electronic supplementary material).

From these elaborations it can be noted (Table 2) that the

average nIF for the publications of Italian university (full)

professors of analytical chemistry is 1.969 (median 1.562,

1st – 3rd quartiles 0.700 – 3.144). It is to be noted that if we

consider not normalized IFs, we obtain values which are

significantly (even if not exceptionally) higher than the

previous ones, i.e. an average of 2.407 (median 2.580, 1st –

3rd quartiles 1.145 – 3.186). This increment (about +0.4

units for the average or +22%) shows that publications in

journals not included in the analytical category have, on

average, higher impact factors than those published in

journals of the analytical category, and this justifies our

decision to make preferential use of normalized IFs.

Table 1 Distribution of the papers according to selected branches

Scientific branch n %

Chemistry

separation science 1112 13.1

environmental 1018 11.9

electroanalytical 1009 11.8

inorganic 601 7.0

solution equilibria 565 6.6

agricultural and food 492 5.8

sensors and biosensors 444 5.2

physical 293 3.4

industrial 263 3.1

general 253 3.0

medicinal 250 2.9

spectroscopy 246 2.9

thermochemistry 242 2.8

clinical and biomedical 207 2.4

chemometrics 196 2.3

mass spectrometry 177 2.1

biochemistry and bioanalytical 161 1.9

surface interface colloid 154 1.8

materials 139 1.6

organic 135 1.6

engineering 77 0.9

archaeometry 73 0.9

organometallic 63 0.7

crystallography 37 0.4

instrumentation 16 0.2

radioanalytical and nuclear 11 0.1

supramolecular 9 0.1forensic 3 0.04

Medicine 91 1.1

Physics 72 0.8

Biology 51 0.6

Multidisciplinary 15 0.2

Geology 5 0.06

Others 49 0.6


http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-



Citations, h-index, and m

Citation summary statistics are reported in Table 2. A

unimodal distribution was observed both for total professors’

citations (mode ∼900, average 1,385, median 1,052, 1st – 3rd

quartiles 677 – 1,696, maximum 7,440), and for the number

of citations per paper (mode ∼13, average 16.9, median 14.5,

1st – 3rd quartiles 12.3 – 18.7, maximum 48.6). Note that if we

consider all the papers separately (independently of the

grouping for each professor) the average number of citations

per paper is 18.3. The top cited papers (total citations >100)

are reported in Table S2 (Electronic supplementary material).

Table 2 Comprehensive summary statistics of normalized and not normalized impact factors (with non-ISI journals assigned IF =0 or excluded),

number of citations per paper, h-index, and m* computed for all, senior, and junior professors

Database Professors Impact

factor

Non-ISI

journals

Average Standard

deviation

Median 1st – 3rd

quartiles

Min – Max

Impact factor

IF of all papers (JCR, 2008) All nIF IF=0 1.969 1.438 1.562 0.700 – 3.144 0 – 5.827

excluded 2.091 1.393 1.800 0.700 –

3.186 0.023 –

5.827IF IF=0 2.407 1.818 2.580 1.145 – 3.186 0 – 28.751

excluded 2.556 1.769 2.632 1.362 – 3.212 0.093 – 28.751

Senior nIF IF=0 1.908 1.415 1.562 0.700 – 2.971 0 – 5.827

excluded 2.026 1.374 1.589 0.700 – 3.186 0.037 – 5.827

IF IF=0 2.345 1.699 2.532 1.124 – 3.186 0 – 28.751

excluded 2.490 1.644 2.580 1.362 – 3.212 0.097 – 28.751

Junior nIF IF=0 2.370 1.517 2.632 0.869 – 3.271 0 – 5.827

excluded 2.520 1.439 2.782 1.164 – 3.279 0.023 – 5.827

IF IF=0 2.801 2.406 2.867 1.693 – 3.446 0 – 28.751

excluded 2.978 2.372 2.934 1.959 – 3.553 0.093 – 28.751

Average IF of each professor ’s

publications (JCR, 2008)

All nIF IF=0 1.955 0.601 1.924 1.556 – 2.305 0.901 – 3.600

excluded 2.066 0.604 2.074 1.658 – 2.426 0.918 – 3.671

IF IF=0 2.398 0.622 2.406 1.931 – 2.714 1.146 – 4.202

excluded 2.536 0.626 2.511 2.115 – 2.904 1.167 – 4.525

Senior nIF IF=0 1.866 0.541 1.877 1.503 – 2.175 0.901 – 3.560

excluded 1.973 0.546 2.010 1.625 – 2.288 0.918 – 3.642

IF IF=0 2.285 0.565 2.314 1.849 – 2.641 1.146 – 4.202

excluded 2.421 0.575 2.421 2.050 – 2.721 1.167 – 4.525

Junior nIF IF=0 2.345 0.707 2.456 1.806 – 2.722 1.128 – 3.600

excluded 2.465 0.695 2.486 2.046 – 2.845 1.190 – 3.671

IF IF=0 2.886 0.641 2.887 2.434 – 3.407 1.720 – 4.015

excluded 3.033 0.611 2.990 2.607 – 3.407 1.962 – 4.246

Citations per paper

Citations of papers

(WoS, all years)

All – – 16.9 8.7 14.5 12.3 – 18.7 5.1 – 48.5

Senior – – 16.6 8.6 14.2 12.2 – 18.6 5.1 – 48.5

Junior – – 18.1 9.4 15.7 12.4 – 24.1 5.7 – 43.1

h-index

All – – 19.1 7.6 18.0 14.0 – 23.0 5 – 44

Senior – – 19.2 8.0 18.0 14.0 – 23.0 5 – 44

Junior – – 18.5 6.2 19.0 14.0 – 21.0 7 – 30

m*

All – – 0.66 0.31 0.62 0.43 – 0.82 0.13 – 1.53

Senior – – 0.60 0.28 0.57 0.42 – 0.73 0.13 – 1.50

Junior – – 0.89 0.33 0.86 0.70 – 1.04 0.27 – 1.53

*m=h/ y, where y=scientific age


http://-/?-

http://-/?-



The results for the h-index and for m (concerning the

h value per year of activity) are reported in Table 2.

Unimodal distributions were also obtained in these cases.

For the h-index the average value was 19.1 (mode ∼17,

median 18.0, 1st – 3rd quartiles 14.0 – 23.0, maximum 44),

whereas for m the average was 0.66 (mode ∼0.50, median

0.62, 1st – 3rd quartiles 0.43 – 0.82, maximum 1.53). It was

observed that a value of m ≈ 1 (i.e. an h-index of 20 after

20 years of scientific activity) characterizes successful

scientists, whereas a value of m ≈ 2 is obtained only by

outstanding scientists found in top universities or major

laboratories [47].

It is worth noting that although h cannot decrease with

time (it continues to increase even after the scientist stops

publishing), m, because it is also related to productivity

(sustained research production), may decrease when a

scientist reduces his/her publishing rate or stops altogether.

The relationship between the total number of citations of

each professor, N c,tot , and his/her h-index was studied by

plotting N c,tot vs. h2 for all professors (Fig. 2). The very

good linearity obtained (r =0.972) verifies the equation N c,

tot =ah2 as predicted by Hirsch [47], with the proportionality

coefficient a=3.34. Similar results were obtained for data separated for senior and junior professors. The value a for

each professor varied between 2.38 and 5.06 (average 3.27,

SD 0.60) which compares well with the empirical interval

of 3 – 5 reported by Hirsch [47].

Comparison with literature data

Impact factor

As possible comparisons with international data we can

quote first an extensive scientometric investigation on the

world literature of analytical chemistry [40] carried out onthe papers published in the period 1978 – 80 in 22 core

journals of analytical chemistry. Here the average IF value

obtained for Italy is 1.678. More recently a study

concerning analytical chemistry in the European Union

carried out between 1993 and 1999 and based on papers

abstracted in Analytical Abstracts [35] reported an average

IF value of 1.70 for Italy compared with 1.75 for the EU. If

we consider that on average the IF of journals in the JCR

category “Chemistry, Analytical” was approximately stable

during the eighties and that it increased by about 0.4 units

from the nineties, when it stood at ∼1.45, to the year 2007,

nIF

0 5 10 15 20 25 30

F r e q u

e n c y

0

500

1000

IF 0 5 10 15 20 25 30

F r e q u e n c y

0

500

1000

b

a

Fig. 1 Frequency distribution of (a) normalized (nIF) and (b) not

normalized (IF) impact factors for all the publications of Italian

university (full) professors (February 2009). IF=0 for non-ISI journals

h2

0 500 1000 1500 2000 2500

N c , t

o t

0

2000

4000

6000

8000

Fig. 2 Relationship between the total number of citations and h2




when it had risen to 1.847 (Fig. S2 in Electronic supplemen-

tary material), the value here obtained for professors using

the 2007 IFs of journals, i.e. 1.969, seems in general

agreement with the values of 1.68 – 1.70 estimated for Italy

at the end of the seventies [40] and during the nineties [35],

if one considers the increasing trend of the past two decades;

it is, moreover, slightly better than the average of the

category “Chemistry, Analytical” in 2007, which is 1.847.For completeness of information it is to be said that a

very recent paper [36] reports higher IF values than those

observed here, but in that case only a very limited number

of journals (i.e. 18 or 4), focussed exclusively on the field

of analytical chemistry for the time period 2000 – 2007, are

considered.

Citations, h-index, and m

No other compilations have been carried out for national

scientists in the field of analytical chemistry until now. A

list of top, international, living chemists (h≥50), distributedamong the various chemistry disciplines, has been pub-

lished by Chemistry World since 2007 and was updated 11

March 2010 [57]. No Italians are present in the March 2010

list, (note that in our results for active Italian university

professors the maximum value reached for h is 44). If we

compute m, however, a few of our professors appear in the

list (Table S3 in Electronic supplementary material).

Further comparison of our m data (average 0.66, SD=

0.31, min 0.13, max 1.53) can be carried out with values

reported by Hirsch [47] for Nobel prize-winners in physics

in the last 20 years. i.e. m=1.14, SD= 0.47, mmin=0.47,

mmax=2.19.

For a broad comparison, our results of the retrieval of

citations and h-indices for major countries for the topic

“Analytical Chemistry” are reported in Table S3 (Electronic

supplementary material). Data approximately follow the

Glänzel model [53, 56] according to which h is related to

the number of papers, n, and the average citation rate per

paper, x, by the simple formula h =cn1/3 x2/3, where c is a

positive constant. A linear relationship is obtained between

h and n1/3 x2/3 with a correlation coefficient r =0.980, while

c =0.800 is to be compared with the value of 0.932 obtained

considering all fields combined for the 1996 – 2006 dataset

[53].

From these results it can be seen that the average

citation rate obtained from the papers of all professors

(18.3) is slightly higher than that retrieved for Italy

(16.56) and in the range of values obtained for major

European countries (approximately between 15 and 20).

No comparison is possible with the h-index of countries

because, as has been pointed out [47], the overall h-index

of a group will generally be larger than that of each

member of the group.

Number of papers published by professors

The frequency distribution of the number of papers

published by each professor and related comments can be

found in the supplementary material.

Comparison between senior and junior professors

If we consider the normalized impact factor distribution of

all the publications of professors distinguishing between

senior and junior professors (Table 2, and the box-plot

representations in Fig. 3), we observe a substantial

difference, especially in the centre of the distribution,

between senior professors (average nIF 1.908, median

1.562) and junior professors (average nIF 2.370, median

2.632) with a net increment of ∼0.5 units on the averages,

and much more on medians (∼1), when we pass from senior

to junior professors. Indeed it is noteworthy that while the

distribution is skewed toward high values for senior

professors, with the highest frequencies at low values(mode=0.700), the skewness is reversed for junior profes-

sors, with the highest frequencies at high values (mode=

3.186). Even the not normalized impact factors show

similar increments, with an average of 2.345 (median

2.532) for senior professors and an average of 2.801

(median 2.867) for junior professors.

Concerning the h-index, results show no great difference

between senior and junior professors (Table 2). Average

values of 19.2 and 18.5 (with medians of 18.0 and 19.0) are

observed for senior and junior professors, respectively, and

both groups showed the same mode, ∼17 (Fig. S5 in

supplementary material).

This apparently strange result (for each researcher

h increases during his/her scientific life) may be because

0

1

2

3

4

5

6

7

Senior JuniorAll

n I F

Fig. 3 Box-plot representation

of the nIF computed from all the

publications of senior, junior,

and all professors


http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-

http://-/?-



of several factors. First, during the sixties and seventies, the

productivity of researchers was somewhat reduced because

of the limited availability of instrumental methods. With the

improvement of electronics new, more efficient and rapid

instrumental techniques (including multielemental or multi-

analyte techniques) were developed and became widely

available from the seventies to eighties. In the same period

the use of computers greatly increased the potential of analytical techniques and data elaboration. More recently

the availability of on-line transfer of information increased

the rapidity of access to both scientific literature and

manuscript publications. Finally the number of missing

cited papers on the WoS database increases substantially

with the age of the papers.

Nonetheless, becauseh is a cumulative index, its absolute

value is not adequate to compare professors with different

seniority. A more correct measurement is obtained by

dividing h for the scientific age, giving the value of m [47],

to be interpreted as an h value per year of activity (results in

Table 2). From the distributions of m values (Fig. 4) it can benoted that m increases significantly passing from senior

professors (average 0.60, SD 0.28, median 0.57) to junior

professors (average 0.89, SD 0.33, median 0.86). The same

trend is found for the average citation rate per paper which

increases from 16.6 to 18.1 (Table 2).

It is possible that more careful selection of the journals

invited to publish the results of scientific research, together

with improvement in the quality of the research itself, could

explain the significant increase in both the impact factor and m

for junior professors compared with senior professors.

Average impact factors of each professor ’s publications

The statistics obtained using the average impact factor

computed from all the publications of each professor are

summarized in Table 2; details and comments are given in

supplementary material.

Conclusions

Scientometric analysis of papers produced by the 80 Italian

university (full) professors of analytical chemistry, carriedout in February 2009, led to 8,529 records, with an average

of 106.6 papers per professor. Ninety-four percent of the

papers were published in ISI journals with 55% in the JCR

category of “Chemistry, Analytical”. This datum highlights

that the scientific interests of analytical chemists also

extend to the subject areas of other chemical disciplines

(particularly environmental, general and inorganic, agricul-

tural and food, physical, industrial, medicinal, clinical/

biomedical, materials, and organic chemical branches)

underlying the interdisciplinary or even multidisciplinary

character of many studies.

The average impact factor (normalized to the scale of the“Chemistry, Analytical” JCR category) obtained from all

the publications retrieved was 1.969 (median 1.562), in

general agreement with values estimated for Italian scien-

tists in previous literature reports, with a net increase from

senior professors, which show an average of 1.908 (median

1.562), to junior professors, which have an average of

2.370 (median 2.632). The corresponding averages

obtained with not-normalized IFs increased relative to the

normalized IF, but not very much — to 2.407 for all

professors and to 2.345 and 2.801, for senior and junior

professors, respectively — signifying, however, that, on

average, publications in not-analytical (especially medical)

journals achieve higher numerical values for IF. Similar

values and a similar increment between senior and junior

professors were also obtained when averages were com-

puted from the average impact factor computed for each

professor.

The average number of citations per paper was 18.3,

irrespective of the seniority of professors and in agreement

with values obtained for major European countries

concerning the topic of analytical chemistry. The average

h-index was 19.2 for senior and 18.5 for junior professors

(19.1 for all professors together), and m, = h/ y increased

substantially from senior to junior professors, on average

from 0.66 to 0.89.

The observed improvements from senior to junior

professors could reasonably be interpreted as a result of

the selection of higher reputation and high-impact-factor

journals by junior professors, when publishing their works,

presumably corresponding to higher quality research and

results in more recent years.

Finally we do not propose use of average IF or h-index

as reference values as they stand: indeed we propose

0

1

2

Senior JuniorAll

m

Fig. 4 Box-plot representation

of m (= h/ y) for senior, junior,

and all professors




scientific (scientometric) elements to be considered by the

chemical analytical academic community in selecting such

values. These selections remain in any case the responsi-

bility of the Italian National University Council (CUN). In

particular the decision to be taken must establish which

statistics are to be considered (average, median, 1st quartile,

or other) and on a way of differentiating them with

reference to the different permanent positions for whichcandidates can compete. Moreover, inevitably, the CUN has

also to take account of the implications of discussions with

and proposals of other academic chemical disciplines and

of the necessary compromise action which will have to be

taken in order to achieve homogenous criteria of evaluation

between them, also with the contribution of the Italian

Chemical Society (SCI, Società Chimica Italiana). It is to

be stressed, however, that, according to the Ministerial note,

academic competition commissions must take account not

only of scientometric indices but also of many other

elements, in particular the consistency of the research

group and the real merits of the candidate, but also thesize of the Institution (with reference particularly to those

which are small or disadvantaged, e.g. where chemical

analytical groups are active in non-chemical Faculties), the

kind and amount of instrumentation to which he or she had

access, and, much more important, by reading the papers

presented. The IF values and/or h-indices should therefore

set no more than minimum requirements to be considered

and should help to limit the possibility of over or under-

estimating the value of candidates for permanent positions

in the discipline of analytical chemistry.

Acknowledgements The authors are grateful to Ettore Novellino, a member of the Italian National University Council (CUN), for the

initial advice that we should perform electronic retrieval of university

(full) professors’ publications in order to obtain the basic information

to answer the Ministerial request concerning a bibliometric reference

to be included in the minimum requirements in academic competition

regulations, to Luigi Campanella, the President of the Italian Chemical

Society (SCI), for his suggestion that we should publish the results of

the work, and to Maria Careri and Aldo Roda, the President and the

Past-President of the Analytical Chemistry Division of the SCI, for

stimulating and useful discussions. We are also grateful to the

Polytechnic University of Marche for its financial support.

References

1. ISI Web of Knowledge (2008) Journal Citation Reports. A

bibliometric analysis of science journals in the ISI database.

Institute for Scientific Information (ISI). Thomson Reuters, New

York

2. Herbstein FH (1993) Scientometrics 28:349 – 373

3. Swinbanks D, Nathan R (1997) Nature 389:113 – 117

4. Makino J (1998) Scientometrics 43:87 – 93

5. Braun T (1999) Scientometrics 45:425 – 432

6. Braun T, Schubert A, Schubert G (2002) Anal Chem 74:477A –

479A

7. Kyvik S (2003) Scientometrics 58:35 – 48

8. Moed HF (2005) Citation analysis in research evaluation.

Springer, Dordrecht

9. Seglen O (1997) Br Med J 314:497 – 502

10. Adam D (2002) Nature 415:726 – 729

11. Warner J (2003) Bull Am Soc Inf Sci Technol 30:26 – 27

12. EASE European Association of Science Editors statement on

impact factors (2008) http://www.ease.org.uk/artman2/uploads/

1/EASE_statement_IFs_final.pdf Apr. 20. Accessed on 2010-02-

1713. Balaban AT (1996) Scientometrics 37:495 – 498

14. Schubert A, Braun T (1996) Scientometrics 36:311 – 324

15. Kostoff RN (1997) Scientometrics 36:225 – 230

16. Ramirez AM, Garcia EO, Del Rio JA (2000) Scientometrics 47:3 –

9

17. Pudovkin AI, Garfield E (2004) Rank-normalized impact factor: a

way to compare journal performance across subject categories.

Proceedings of the 67th ASIS&T Annual Meeting 41:507 – 515

18. Dong P, Loh M, Mondry A (2005) Biomed Digit Libr 2:1 – 8

19. Sombatsompop N, Markpin T, Yochai W, Saechiew M (2005)

Scientometrics 65:293 – 305

20. Sombatsompop N, Markpin TJ (2005) Am Soc Inf Sci Technol

56:676 – 683

21. Garfield E (2006) J A Med Assoc 295:1 – 7

22. Lundberg J (2007) J Informetrics 1:145 – 154

23. Hoeffel C (1998) Allergy 53:1225

24. King AK (2004) Nature 430:311 – 316

25. OECD Organisation for Economic Co-operation and Develop-

ment (1987) Evaluation of research. A Selection of Current

Practices. OECD Paris


ment (1997) The evaluation of Scientific Research: Selected

Experiences. OECD Paris


ment (2009) Science, Technology and Industry Scoreboard 2009

(a biannual series). OECD Paris

28. Braun T, Glänzel W, Schubert A (1989) Trends Anal Chem

8:281 – 284


8:316 – 318

30. Braun T, Schubert A (1990) Trends Anal Chem 9:176 – 182

31. Valcarcel M (1993) Trends Anal Chem 12:IX – XIII

32. Thomaidis NS, Georgiou CA, Calokerinos AC (2004) Anal Chim

Acta 505:3 – 8

33. Kastelan-Macan M, Klaic B (2000) Croat Chem Acta 73:1 – 21

34. Muñoz de la Peña A, Pulgarin A (1992) Anal Proceed 29:517 – 520

35. Georgiou CA, Thomaidis NS (2001) Trends Anal Chem 20:462 –

466

36. Téllez H, Vadillo JM (2010) Anal Bioanal Chem 397:1477 – 1484;

doi:10.1007/s00216-010-3732-6

37. Valcarcel M, Rios A (1995) Trends Anal Chem 14:94 – 100

38. Rodenas-Torralba E, Morales-Rubio A, de la Guardia M (2006)

Spectr Lett 39:513 – 532

39. Braun T, Schubert A (1991) Trends Anal Chem 10:1 –

340. Braun T, Bujdosó E, Schubert A (1987) Literature of analytical

chemistry: a scientometric evaluation. CRC Press, Boca Raton


9:278 – 279

42. Burke M (2001) Anal Chem 73:595A – 597A

43. Italian law 9 January 2009 no. 1, Conversione in legge, con

modificazioni, del decreto-legge 10 novembre 2008, n. 180,

recante disposizioni urgenti per il diritto allo studio, la valor-

izzazione del merito e la qualità del sistema universitario e della

ricerca, Italian Official Bull. 9 January 2009 no. 6, art. 1. http://

www.parlamento.it/parlam/leggi/09001l.htm. Accessed February

17, 2010


http://www.ease.org.uk/artman2/uploads/1/EASE_statement_IFs_final.pdf


http://dx.doi.org/10.1007/s00216-010-3732-6

http://www.parlamento.it/parlam/leggi/09001l.htm




http://dx.doi.org/10.1007/s00216-010-3732-6





44. Italian National University Council (CUN) Documento di lavoro

del Consiglio Universitario Nazionale su: indicatori di attività

scientifica e di ricerca, Session of 16 – 18 December 2008, http://

www.cun.it/documenti/documenti-di-lavoro/2009/indicatori-

attività-scientifica-ricerca-.aspx Accessed February 17, 2010

45. DM 89/2009. Decree of the Minister of Education, Universities

and Research of Italy, no. 89, July 28th 2009, Valutazione dei

titoli e delle pubblicaizoni scientifiche, http://www.miur.it/

0006Menu_C/0012Docume/0015Atti_M/7921Valuta_cf2.htm.

Accessed February 17, 201046. SciFinder (2009) Chemical Abstract Service, American Chemical

Society. Washington, D.C., USA

47. Hirsch JE (2005) Proc Natl Acad Sci USA 102:16569 – 16572

48. ISI Web of Science (2010) Institute for Scientific Information

(ISI). Thomson Reuters, New York

49. CAplus (2009) Chemical Abstracts Service (CAS), American

Chemical Society, Washington, D.C., USA

50. MEDLINE (2009) Community of Science, U.S. National Library

of Medicine, Bethesda, Maryland, USA

51. Van Raan AFJ (2006) Scientometrics 67:491 – 502

52. Prathap G (2006) Current Science 91:1439

53. Csajbók E, Berhidi A, Vasas L, Schubert A (2007) Scientometrics

73:91 – 117

54. Sen BK (1992) J Document 48:318 – 325

55. Batista PD, Campiteli MG, Kinouchi O, Martinez AS (2006)Scientometrics 68:179 – 189

56. Glänzel W (2006) Scientometrics 67:315 – 321

57. Van Noorden R, Hirsch index ranks top chemists, Chemistry World,

23 April 2007, updated 11 March 2011. Retrieved 31 March 2011;

http://www.rsc.org/images/H-indexMarch2010_tcm18-85867.pdf

Anna Annibaldi is a research

associate in Analytical Chemistry

at the Marche Polytechnic Univer-

sity, Ancona (Italy), where she

teaches the course on “Analysis of

pollutants”. She graduated inChemistry at the University of

Bologna in 2001 and she obtained

a PhD degree in Biology and

Marine Ecology in 2005. Her

scientific interests have focussed

principally on the field of environ-

mental analytical chemistry, partic-

ularly on the voltammetric study of

heavy metals in several environmental matrices, and on the determination

of organic pollutants in food. She has contributed to the development of

chemical fractionation procedures for heavy metals in atmospheric

aerosols and in snow.

Cristina Truzzi is researcher in

Analytical Chemistry at the Marche

Polytechnic University in Ancona

(Italy) were she teaches courses on

“Instrumental analytical chemistry”

and “Food chemical analysis”. She

worked previously at the University

of Modena in the field of analytical

pharmacology and environmental

toxicology, on the determination of

heavy metals and the study of their

toxic effects in organisms. Cur-rently, her scientific interests are

focused mainly on environmental

analytical chemistry, with particular attention to the determination and

speciation of trace heavy metals in different environmental matrixes with

different amounts of anthropization (seawater, snow, aerosol. organ-

isms), using voltammetric techniques (DPASV, SWASV). She is also

interested in determination of organic pollutants in the environment and

food by gaschromatography-massspectrometry and in theapplication of

environmental chemometrics to marine science. She has taken part in

several oceanographic campaigns. She is a member of the Italian

Chemical Society (Division of Analytical Chemistry).

Silvia Illuminati is a research

associate in Analytical Chemistry

at the Marche Polytechnic Univer-

sity, Ancona (Italy). She graduated

in Biology at the same university

in 2001 and she obtained a PhDdegree in Biology and Marine

Ecology in 2005. Her scientific

interests have focussed principally

on environmental analytical chem-

istry, particularly on the voltam-

metric study of heavy metals in

different environmental and food

matrices and also of biomarkers in

marine organisms. She has taken part in two Italian Expeditions to

Antarctica and in several oceanographic campaigns (Italy, United

Kingdom), contributing to the development of a voltammetric in-situ

profiler system for real time heavy metal determination in sea water.

Giuseppe Scarponi is full profes-

sor of Analytical Chemistry of the

Marche Polytechnic University in

Ancona (Italy). He previously

worked at the Universities of

Camerino, Genoa, and Venice. He

is a member of the Analytical

Chemistry Divisions of the Ameri-

can Chemical Society (since 1983)

and the Italian Chemical Society

(since 1975) of which he has been a member of the Governing Board

since 2007. His scientific interests

have focused mainly on electroana-

lytical chemistry, environmental analytical chemistry, and chemometrics.

He has taken part in seven Italian expeditions to Antarctica and in

oceanographic cruises in the Mediterranean Sea and in the Ross Sea as

leader of the analytical chemistry group. He was the scientific coordinator

of the Italian Antarctic Station during the 1998 – 1999 austral summer

campaign. He is a member of the scientific committee of the European

ResearchCourse on Atmospheres (Grenoble) and he is a founding associate

of the Italian Aerosol Society.


http://www.cun.it/documenti/documenti-di-lavoro/2009/indicatori-attivit%E0%AD%B3cientifica-ricerca-.aspx



http://www.miur.it/0006Menu_C/0012Docume/0015Atti_M/7921Valuta_cf2.htm









analiza a a cercetarii stiintifice intr-o univ din italia - en

Documents