Page 100 - Becoming Metric Wise

90 Becoming Metric-Wise

Of course the obtained value is the same as above. The formula is
slightly different because the data are ranked in the opposite way.
Finally, the Lorenz curve and the Gini index are also used when one
is interested in diversity, for instance the diversity of journals shown in a
reference list. Also then the Gini index can be used, but as diversity is
seen as the opposite notion of inequality, the Gini diversity index is calcu-
lated as one minus the Gini inequality index. For the Gini diversity index
the diagonal leads to the value 1 (largest diversity).

4.11 APPLICATIONS IN INFORMETRICS

In this section we provide a list of possible applications of descriptive
statistics.
• number of publications per author, where authors are all members of
the same institute, or working in the same domain;
• number of authors per publications, again studied per domain, or over
time;
• number of coauthors per publication (per domain, over time);
• ages of references;
• time between publication and received citations; as a specific case one
has studied the time to first citation, see Subsection 9.4.2;
• number of references (per journal, per domain);
• number of collaborations between institutes or countries or sectors
(e.g., industry government collaborations);
• number of downloads per article, per journal;
• number of received citations (a specific average is called the journal
impact factor, see Chapter 6: Journal Citation Analysis);
• number or fraction of uncited articles (over time, per journal, per
author);
• fraction of citations older than n years (where n can be any natural
number).

PART B. INFERENTIAL STATISTICS
4.12 THE NORMAL DISTRIBUTION

Classical statistics is based on the normal distribution. Although most dis-
tributions observed in informetric applications do not have a normal dis-
tribution, we cannot completely do without this basic function. For this
reason we briefly recall its main definitions and features.